/usr/src/sys/dev/pci/drm/i915/i915

Bug Summary

File:	dev/pci/drm/i915/i915_vma.h
Warning:	line 147, column 22 Access to field 'obj' results in a dereference of a null pointer (loaded from variable 'vma')

Annotated Source Code

Press '?' to see keyboard shortcuts

Show analyzer invocation

clang -cc1 -cc1 -triple amd64-unknown-openbsd7.4 -analyze -disable-free -clear-ast-before-backend -disable-llvm-verifier -discard-value-names -main-file-name i915_perf.c -analyzer-checker=core -analyzer-checker=apiModeling -analyzer-checker=unix -analyzer-checker=deadcode -analyzer-checker=security.insecureAPI.UncheckedReturn -analyzer-checker=security.insecureAPI.getpw -analyzer-checker=security.insecureAPI.gets -analyzer-checker=security.insecureAPI.mktemp -analyzer-checker=security.insecureAPI.mkstemp -analyzer-checker=security.insecureAPI.vfork -analyzer-checker=nullability.NullPassedToNonnull -analyzer-checker=nullability.NullReturnedFromNonnull -analyzer-output plist -w -setup-static-analyzer -mrelocation-model static -mframe-pointer=all -relaxed-aliasing -ffp-contract=on -fno-rounding-math -mconstructor-aliases -ffreestanding -mcmodel=kernel -target-cpu x86-64 -target-feature +retpoline-indirect-calls -target-feature +retpoline-indirect-branches -target-feature -sse2 -target-feature -sse -target-feature -3dnow -target-feature -mmx -target-feature +save-args -target-feature +retpoline-external-thunk -disable-red-zone -no-implicit-float -tune-cpu generic -debugger-tuning=gdb -fcoverage-compilation-dir=/usr/src/sys/arch/amd64/compile/GENERIC.MP/obj -nostdsysteminc -nobuiltininc -resource-dir /usr/local/llvm16/lib/clang/16 -I /usr/src/sys -I /usr/src/sys/arch/amd64/compile/GENERIC.MP/obj -I /usr/src/sys/arch -I /usr/src/sys/dev/pci/drm/include -I /usr/src/sys/dev/pci/drm/include/uapi -I /usr/src/sys/dev/pci/drm/amd/include/asic_reg -I /usr/src/sys/dev/pci/drm/amd/include -I /usr/src/sys/dev/pci/drm/amd/amdgpu -I /usr/src/sys/dev/pci/drm/amd/display -I /usr/src/sys/dev/pci/drm/amd/display/include -I /usr/src/sys/dev/pci/drm/amd/display/dc -I /usr/src/sys/dev/pci/drm/amd/display/amdgpu_dm -I /usr/src/sys/dev/pci/drm/amd/pm/inc -I /usr/src/sys/dev/pci/drm/amd/pm/legacy-dpm -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/inc -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/smu11 -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/smu12 -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/smu13 -I /usr/src/sys/dev/pci/drm/amd/pm/powerplay/inc -I /usr/src/sys/dev/pci/drm/amd/pm/powerplay/hwmgr -I /usr/src/sys/dev/pci/drm/amd/pm/powerplay/smumgr -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/inc -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/inc/pmfw_if -I /usr/src/sys/dev/pci/drm/amd/display/dc/inc -I /usr/src/sys/dev/pci/drm/amd/display/dc/inc/hw -I /usr/src/sys/dev/pci/drm/amd/display/dc/clk_mgr -I /usr/src/sys/dev/pci/drm/amd/display/modules/inc -I /usr/src/sys/dev/pci/drm/amd/display/modules/hdcp -I /usr/src/sys/dev/pci/drm/amd/display/dmub/inc -I /usr/src/sys/dev/pci/drm/i915 -D DDB -D DIAGNOSTIC -D KTRACE -D ACCOUNTING -D KMEMSTATS -D PTRACE -D POOL_DEBUG -D CRYPTO -D SYSVMSG -D SYSVSEM -D SYSVSHM -D UVM_SWAP_ENCRYPT -D FFS -D FFS2 -D FFS_SOFTUPDATES -D UFS_DIRHASH -D QUOTA -D EXT2FS -D MFS -D NFSCLIENT -D NFSSERVER -D CD9660 -D UDF -D MSDOSFS -D FIFO -D FUSE -D SOCKET_SPLICE -D TCP_ECN -D TCP_SIGNATURE -D INET6 -D IPSEC -D PPP_BSDCOMP -D PPP_DEFLATE -D PIPEX -D MROUTING -D MPLS -D BOOT_CONFIG -D USER_PCICONF -D APERTURE -D MTRR -D NTFS -D SUSPEND -D HIBERNATE -D PCIVERBOSE -D USBVERBOSE -D WSDISPLAY_COMPAT_USL -D WSDISPLAY_COMPAT_RAWKBD -D WSDISPLAY_DEFAULTSCREENS=6 -D X86EMU -D ONEWIREVERBOSE -D MULTIPROCESSOR -D MAXUSERS=80 -D _KERNEL -O2 -Wno-pointer-sign -Wno-address-of-packed-member -Wno-constant-conversion -Wno-unused-but-set-variable -Wno-gnu-folding-constant -fdebug-compilation-dir=/usr/src/sys/arch/amd64/compile/GENERIC.MP/obj -ferror-limit 19 -fwrapv -D_RET_PROTECTOR -ret-protector -fcf-protection=branch -fgnuc-version=4.2.1 -vectorize-loops -vectorize-slp -fno-builtin-malloc -fno-builtin-calloc -fno-builtin-realloc -fno-builtin-valloc -fno-builtin-free -fno-builtin-strdup -fno-builtin-strndup -analyzer-output=html -faddrsig -o /home/ben/Projects/scan/2024-01-11-110808-61670-1 -x c /usr/src/sys/dev/pci/drm/i915/i915_perf.c

/usr/src/sys/dev/pci/drm/i915/i915_perf.c

→

1/*
* Copyright © 2015-2016 Intel Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*
* Authors:
*   Robert Bragg <robert@sixbynine.org>
*/


28/**
* DOC: i915 Perf Overview
*
* Gen graphics supports a large number of performance counters that can help
* driver and application developers understand and optimize their use of the
* GPU.
*
* This i915 perf interface enables userspace to configure and open a file
* descriptor representing a stream of GPU metrics which can then be read() as
* a stream of sample records.
*
* The interface is particularly suited to exposing buffered metrics that are
* captured by DMA from the GPU, unsynchronized with and unrelated to the CPU.
*
* Streams representing a single context are accessible to applications with a
* corresponding drm file descriptor, such that OpenGL can use the interface
* without special privileges. Access to system-wide metrics requires root
* privileges by default, unless changed via the dev.i915.perf_event_paranoid
* sysctl option.
*
*/

50/**
* DOC: i915 Perf History and Comparison with Core Perf
*
* The interface was initially inspired by the core Perf infrastructure but
* some notable differences are:
*
* i915 perf file descriptors represent a "stream" instead of an "event"; where
* a perf event primarily corresponds to a single 64bit value, while a stream
* might sample sets of tightly-coupled counters, depending on the
* configuration.  For example the Gen OA unit isn't designed to support
* orthogonal configurations of individual counters; it's configured for a set
* of related counters. Samples for an i915 perf stream capturing OA metrics
* will include a set of counter values packed in a compact HW specific format.
* The OA unit supports a number of different packing formats which can be
* selected by the user opening the stream. Perf has support for grouping
* events, but each event in the group is configured, validated and
* authenticated individually with separate system calls.
*
* i915 perf stream configurations are provided as an array of u64 (key,value)
* pairs, instead of a fixed struct with multiple miscellaneous config members,
* interleaved with event-type specific members.
*
* i915 perf doesn't support exposing metrics via an mmap'd circular buffer.
* The supported metrics are being written to memory by the GPU unsynchronized
* with the CPU, using HW specific packing formats for counter sets. Sometimes
* the constraints on HW configuration require reports to be filtered before it
* would be acceptable to expose them to unprivileged applications - to hide
* the metrics of other processes/contexts. For these use cases a read() based
* interface is a good fit, and provides an opportunity to filter data as it
* gets copied from the GPU mapped buffers to userspace buffers.
*
*
* Issues hit with first prototype based on Core Perf
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*
* The first prototype of this driver was based on the core perf
* infrastructure, and while we did make that mostly work, with some changes to
* perf, we found we were breaking or working around too many assumptions baked
* into perf's currently cpu centric design.
*
* In the end we didn't see a clear benefit to making perf's implementation and
* interface more complex by changing design assumptions while we knew we still
* wouldn't be able to use any existing perf based userspace tools.
*
* Also considering the Gen specific nature of the Observability hardware and
* how userspace will sometimes need to combine i915 perf OA metrics with
* side-band OA data captured via MI_REPORT_PERF_COUNT commands; we're
* expecting the interface to be used by a platform specific userspace such as
* OpenGL or tools. This is to say; we aren't inherently missing out on having
* a standard vendor/architecture agnostic interface by not using perf.
*
*
* For posterity, in case we might re-visit trying to adapt core perf to be
* better suited to exposing i915 metrics these were the main pain points we
* hit:
*
* - The perf based OA PMU driver broke some significant design assumptions:
*
*   Existing perf pmus are used for profiling work on a cpu and we were
*   introducing the idea of _IS_DEVICE pmus with different security
*   implications, the need to fake cpu-related data (such as user/kernel
*   registers) to fit with perf's current design, and adding _DEVICE records
*   as a way to forward device-specific status records.
*
*   The OA unit writes reports of counters into a circular buffer, without
*   involvement from the CPU, making our PMU driver the first of a kind.
*
*   Given the way we were periodically forward data from the GPU-mapped, OA
*   buffer to perf's buffer, those bursts of sample writes looked to perf like
*   we were sampling too fast and so we had to subvert its throttling checks.
*
*   Perf supports groups of counters and allows those to be read via
*   transactions internally but transactions currently seem designed to be
*   explicitly initiated from the cpu (say in response to a userspace read())
*   and while we could pull a report out of the OA buffer we can't
*   trigger a report from the cpu on demand.
*
*   Related to being report based; the OA counters are configured in HW as a
*   set while perf generally expects counter configurations to be orthogonal.
*   Although counters can be associated with a group leader as they are
*   opened, there's no clear precedent for being able to provide group-wide
*   configuration attributes (for example we want to let userspace choose the
*   OA unit report format used to capture all counters in a set, or specify a
*   GPU context to filter metrics on). We avoided using perf's grouping
*   feature and forwarded OA reports to userspace via perf's 'raw' sample
*   field. This suited our userspace well considering how coupled the counters
*   are when dealing with normalizing. It would be inconvenient to split
*   counters up into separate events, only to require userspace to recombine
*   them. For Mesa it's also convenient to be forwarded raw, periodic reports
*   for combining with the side-band raw reports it captures using
*   MI_REPORT_PERF_COUNT commands.
*
*   - As a side note on perf's grouping feature; there was also some concern
*     that using PERF_FORMAT_GROUP as a way to pack together counter values
*     would quite drastically inflate our sample sizes, which would likely
*     lower the effective sampling resolutions we could use when the available
*     memory bandwidth is limited.
*
*     With the OA unit's report formats, counters are packed together as 32
*     or 40bit values, with the largest report size being 256 bytes.
*
*     PERF_FORMAT_GROUP values are 64bit, but there doesn't appear to be a
*     documented ordering to the values, implying PERF_FORMAT_ID must also be
*     used to add a 64bit ID before each value; giving 16 bytes per counter.
*
*   Related to counter orthogonality; we can't time share the OA unit, while
*   event scheduling is a central design idea within perf for allowing
*   userspace to open + enable more events than can be configured in HW at any
*   one time.  The OA unit is not designed to allow re-configuration while in
*   use. We can't reconfigure the OA unit without losing internal OA unit
*   state which we can't access explicitly to save and restore. Reconfiguring
*   the OA unit is also relatively slow, involving ~100 register writes. From
*   userspace Mesa also depends on a stable OA configuration when emitting
*   MI_REPORT_PERF_COUNT commands and importantly the OA unit can't be
*   disabled while there are outstanding MI_RPC commands lest we hang the
*   command streamer.
*
*   The contents of sample records aren't extensible by device drivers (i.e.
*   the sample_type bits). As an example; Sourab Gupta had been looking to
*   attach GPU timestamps to our OA samples. We were shoehorning OA reports
*   into sample records by using the 'raw' field, but it's tricky to pack more
*   than one thing into this field because events/core.c currently only lets a
*   pmu give a single raw data pointer plus len which will be copied into the
*   ring buffer. To include more than the OA report we'd have to copy the
*   report into an intermediate larger buffer. I'd been considering allowing a
*   vector of data+len values to be specified for copying the raw data, but
*   it felt like a kludge to being using the raw field for this purpose.
*
* - It felt like our perf based PMU was making some technical compromises
*   just for the sake of using perf:
*
*   perf_event_open() requires events to either relate to a pid or a specific
*   cpu core, while our device pmu related to neither.  Events opened with a
*   pid will be automatically enabled/disabled according to the scheduling of
*   that process - so not appropriate for us. When an event is related to a
*   cpu id, perf ensures pmu methods will be invoked via an inter process
*   interrupt on that core. To avoid invasive changes our userspace opened OA
*   perf events for a specific cpu. This was workable but it meant the
*   majority of the OA driver ran in atomic context, including all OA report
*   forwarding, which wasn't really necessary in our case and seems to make
*   our locking requirements somewhat complex as we handled the interaction
*   with the rest of the i915 driver.
*/

194#include <linux/anon_inodes.h>
195#include <linux/sizes.h>
196#include <linux/uuid.h>

198#include "gem/i915_gem_context.h"
199#include "gem/i915_gem_internal.h"
200#include "gt/intel_engine_pm.h"
201#include "gt/intel_engine_regs.h"
202#include "gt/intel_engine_user.h"
203#include "gt/intel_execlists_submission.h"
204#include "gt/intel_gpu_commands.h"
205#include "gt/intel_gt.h"
206#include "gt/intel_gt_clock_utils.h"
207#include "gt/intel_gt_regs.h"
208#include "gt/intel_lrc.h"
209#include "gt/intel_lrc_reg.h"
210#include "gt/intel_ring.h"

212#include "i915_drv.h"
213#include "i915_file_private.h"
214#include "i915_perf.h"
215#include "i915_perf_oa_regs.h"

217/* HW requires this to be a power of two, between 128k and 16M, though driver
* is currently generally designed assuming the largest 16M size is used such
* that the overflow cases are unlikely in normal operation.
*/
221#define OA_BUFFER_SIZE(16 << 20)		SZ_16M(16 << 20)

223#define OA_TAKEN(tail, head)((tail - head) & ((16 << 20) - 1))	((tail - head) & (OA_BUFFER_SIZE(16 << 20) - 1))

225/**
* DOC: OA Tail Pointer Race
*
* There's a HW race condition between OA unit tail pointer register updates and
* writes to memory whereby the tail pointer can sometimes get ahead of what's
* been written out to the OA buffer so far (in terms of what's visible to the
* CPU).
*
* Although this can be observed explicitly while copying reports to userspace
* by checking for a zeroed report-id field in tail reports, we want to account
* for this earlier, as part of the oa_buffer_check_unlocked to avoid lots of
* redundant read() attempts.
*
* We workaround this issue in oa_buffer_check_unlocked() by reading the reports
* in the OA buffer, starting from the tail reported by the HW until we find a
* report with its first 2 dwords not 0 meaning its previous report is
* completely in memory and ready to be read. Those dwords are also set to 0
* once read and the whole buffer is cleared upon OA buffer initialization. The
* first dword is the reason for this report while the second is the timestamp,
* making the chances of having those 2 fields at 0 fairly unlikely. A more
* detailed explanation is available in oa_buffer_check_unlocked().
*
* Most of the implementation details for this workaround are in
* oa_buffer_check_unlocked() and _append_oa_reports()
*
* Note for posterity: previously the driver used to define an effective tail
* pointer that lagged the real pointer by a 'tail margin' measured in bytes
* derived from %OA_TAIL_MARGIN_NSEC and the configured sampling frequency.
* This was flawed considering that the OA unit may also automatically generate
* non-periodic reports (such as on context switch) or the OA unit may be
* enabled without any periodic sampling.
*/
257#define OA_TAIL_MARGIN_NSEC100000ULL	100000ULL
258#define INVALID_TAIL_PTR0xffffffff	0xffffffff

260/* The default frequency for checking whether the OA unit has written new
* reports to the circular OA buffer...
*/
263#define DEFAULT_POLL_FREQUENCY_HZ200 200
264#define DEFAULT_POLL_PERIOD_NS(1000000000L / 200) (NSEC_PER_SEC1000000000L / DEFAULT_POLL_FREQUENCY_HZ200)

266/* for sysctl proc_dointvec_minmax of dev.i915.perf_stream_paranoid */
267static u32 i915_perf_stream_paranoid = true1;

269/* The maximum exponent the hardware accepts is 63 (essentially it selects one
* of the 64bit timestamp bits to trigger reports from) but there's currently
* no known use case for sampling as infrequently as once per 47 thousand years.
*
* Since the timestamps included in OA reports are only 32bits it seems
* reasonable to limit the OA exponent where it's still possible to account for
* overflow in OA report timestamps.
*/
277#define OA_EXPONENT_MAX31 31

279#define INVALID_CTX_ID0xffffffff 0xffffffff

281/* On Gen8+ automatically triggered OA reports include a 'reason' field... */
282#define OAREPORT_REASON_MASK0x3f           0x3f
283#define OAREPORT_REASON_MASK_EXTENDED0x7f  0x7f
284#define OAREPORT_REASON_SHIFT19          19
285#define OAREPORT_REASON_TIMER(1<<0)          (1<<0)
286#define OAREPORT_REASON_CTX_SWITCH(1<<3)     (1<<3)
287#define OAREPORT_REASON_CLK_RATIO(1<<5)      (1<<5)


290/* For sysctl proc_dointvec_minmax of i915_oa_max_sample_rate
*
* The highest sampling frequency we can theoretically program the OA unit
* with is always half the timestamp frequency: E.g. 6.25Mhz for Haswell.
*
* Initialized just before we register the sysctl parameter.
*/
297static int oa_sample_rate_hard_limit;

299/* Theoretically we can program the OA unit to sample every 160ns but don't
* allow that by default unless root...
*
* The default threshold of 100000Hz is based on perf's similar
* kernel.perf_event_max_sample_rate sysctl parameter.
*/
305static u32 i915_oa_max_sample_rate = 100000;

307/* XXX: beware if future OA HW adds new report formats that the current
* code assumes all reports have a power-of-two size and ~(size - 1) can
* be used as a mask to align the OA tail pointer.
*/
311static const struct i915_oa_format oa_formats[I915_OA_FORMAT_MAX] = {
[I915_OA_FORMAT_A13]	    = { 0, 64 },
[I915_OA_FORMAT_A29]	    = { 1, 128 },
[I915_OA_FORMAT_A13_B8_C8]  = { 2, 128 },
/* A29_B8_C8 Disallowed as 192 bytes doesn't factor into buffer size */
[I915_OA_FORMAT_B4_C8]	    = { 4, 64 },
[I915_OA_FORMAT_A45_B8_C8]  = { 5, 256 },
[I915_OA_FORMAT_B4_C8_A16]  = { 6, 128 },
[I915_OA_FORMAT_C4_B8]	    = { 7, 64 },
[I915_OA_FORMAT_A12]		    = { 0, 64 },
[I915_OA_FORMAT_A12_B8_C8]	    = { 2, 128 },
[I915_OA_FORMAT_A32u40_A4u32_B8_C8] = { 5, 256 },
323};

325#define SAMPLE_OA_REPORT(1<<0)      (1<<0)

327/**
* struct perf_open_properties - for validated properties given to open a stream
* @sample_flags: `DRM_I915_PERF_PROP_SAMPLE_*` properties are tracked as flags
* @single_context: Whether a single or all gpu contexts should be monitored
* @hold_preemption: Whether the preemption is disabled for the filtered
*                   context
* @ctx_handle: A gem ctx handle for use with @single_context
* @metrics_set: An ID for an OA unit metric set advertised via sysfs
* @oa_format: An OA unit HW report format
* @oa_periodic: Whether to enable periodic OA unit sampling
* @oa_period_exponent: The OA unit sampling period is derived from this
* @engine: The engine (typically rcs0) being monitored by the OA unit
* @has_sseu: Whether @sseu was specified by userspace
* @sseu: internal SSEU configuration computed either from the userspace
*        specified configuration in the opening parameters or a default value
*        (see get_default_sseu_config())
* @poll_oa_period: The period in nanoseconds at which the CPU will check for OA
* data availability
*
* As read_properties_unlocked() enumerates and validates the properties given
* to open a stream of metrics the configuration is built up in the structure
* which starts out zero initialized.
*/
350struct perf_open_properties {
u32 sample_flags;

u64 single_context:1;
u64 hold_preemption:1;
u64 ctx_handle;

/* OA sampling state */
int metrics_set;
int oa_format;
bool_Bool oa_periodic;
int oa_period_exponent;

struct intel_engine_cs *engine;

bool_Bool has_sseu;
struct intel_sseu sseu;

u64 poll_oa_period;
369};

371struct i915_oa_config_bo {
struct llist_node node;

struct i915_oa_config *oa_config;
struct i915_vma *vma;
376};

378static struct ctl_table_header *sysctl_header;

380#ifdef notyet
381static enum hrtimer_restart oa_poll_check_timer_cb(struct hrtimer *hrtimer);
382#endif

384void i915_oa_config_release(struct kref *ref)
385{
struct i915_oa_config *oa_config =
container_of(ref, typeof(*oa_config), ref)({ const __typeof( ((typeof(*oa_config) *)0)->ref ) *__mptr
 = (ref); (typeof(*oa_config) *)( (char *)__mptr - __builtin_offsetof
(typeof(*oa_config), ref) );});

kfree(oa_config->flex_regs);
kfree(oa_config->b_counter_regs);
kfree(oa_config->mux_regs);

kfree_rcu(oa_config, rcu)do { free((void *)oa_config, 145, 0); } while(0);
394}

396struct i915_oa_config *
397i915_perf_get_oa_config(struct i915_perf *perf, int metrics_set)
398{
struct i915_oa_config *oa_config;

rcu_read_lock();
oa_config = idr_find(&perf->metrics_idr, metrics_set);
if (oa_config)
oa_config = i915_oa_config_get(oa_config);
rcu_read_unlock();

return oa_config;
408}

410#ifdef notyet

412static void free_oa_config_bo(struct i915_oa_config_bo *oa_bo)
413{
i915_oa_config_put(oa_bo->oa_config);
i915_vma_put(oa_bo->vma);
kfree(oa_bo);
417}

419#endif

421static u32 gen12_oa_hw_tail_read(struct i915_perf_stream *stream)
422{
struct intel_uncore *uncore = stream->uncore;

return intel_uncore_read(uncore, GEN12_OAG_OATAILPTR((const i915_reg_t){ .reg = (0xdb04) })) &
      GEN12_OAG_OATAILPTR_MASK0xffffffc0;
427}

429static u32 gen8_oa_hw_tail_read(struct i915_perf_stream *stream)
430{
struct intel_uncore *uncore = stream->uncore;

return intel_uncore_read(uncore, GEN8_OATAILPTR((const i915_reg_t){ .reg = (0x2B10) })) & GEN8_OATAILPTR_MASK0xffffffc0;
434}

436static u32 gen7_oa_hw_tail_read(struct i915_perf_stream *stream)
437{
struct intel_uncore *uncore = stream->uncore;
u32 oastatus1 = intel_uncore_read(uncore, GEN7_OASTATUS1((const i915_reg_t){ .reg = (0x2364) }));

return oastatus1 & GEN7_OASTATUS1_TAIL_MASK0xffffffc0;
442}

444#ifdef notyet

446/**
* oa_buffer_check_unlocked - check for data and update tail ptr state
* @stream: i915 stream instance
*
* This is either called via fops (for blocking reads in user ctx) or the poll
* check hrtimer (atomic ctx) to check the OA buffer tail pointer and check
* if there is data available for userspace to read.
*
* This function is central to providing a workaround for the OA unit tail
* pointer having a race with respect to what data is visible to the CPU.
* It is responsible for reading tail pointers from the hardware and giving
* the pointers time to 'age' before they are made available for reading.
* (See description of OA_TAIL_MARGIN_NSEC above for further details.)
*
* Besides returning true when there is data available to read() this function
* also updates the tail, aging_tail and aging_timestamp in the oa_buffer
* object.
*
* Note: It's safe to read OA config state here unlocked, assuming that this is
* only called while the stream is enabled, while the global OA configuration
* can't be modified.
*
* Returns: %true if the OA buffer contains data, else %false
*/
470static bool_Bool oa_buffer_check_unlocked(struct i915_perf_stream *stream)
471{
u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
int report_size = stream->oa_buffer.format_size;
unsigned long flags;
bool_Bool pollin;
u32 hw_tail;
u64 now;

/* We have to consider the (unlikely) possibility that read() errors
* could result in an OA buffer reset which might reset the head and
* tail state.
*/
spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags)do { flags = 0; mtx_enter(&stream->oa_buffer.ptr_lock)
; } while (0);

hw_tail = stream->perf->ops.oa_hw_tail_read(stream);

/* The tail pointer increases in 64 byte increments,
* not in report_size steps...
*/
hw_tail &= ~(report_size - 1);

now = ktime_get_mono_fast_ns();

if (hw_tail == stream->oa_buffer.aging_tail &&
   (now - stream->oa_buffer.aging_timestamp) > OA_TAIL_MARGIN_NSEC100000ULL) {
/* If the HW tail hasn't move since the last check and the HW
 * tail has been aging for long enough, declare it the new
 * tail.
 */
stream->oa_buffer.tail = stream->oa_buffer.aging_tail;
} else {
u32 head, tail, aged_tail;

/* NB: The head we observe here might effectively be a little
 * out of date. If a read() is in progress, the head could be
 * anywhere between this head and stream->oa_buffer.tail.
 */
head = stream->oa_buffer.head - gtt_offset;
aged_tail = stream->oa_buffer.tail - gtt_offset;

hw_tail -= gtt_offset;
tail = hw_tail;

/* Walk the stream backward until we find a report with dword 0
 * & 1 not at 0. Since the circular buffer pointers progress by
 * increments of 64 bytes and that reports can be up to 256
 * bytes long, we can't tell whether a report has fully landed
 * in memory before the first 2 dwords of the following report
 * have effectively landed.
 *
 * This is assuming that the writes of the OA unit land in
 * memory in the order they were written to.
 * If not : (╯°□°）╯︵ ┻━┻
 */
while (OA_TAKEN(tail, aged_tail)((tail - aged_tail) & ((16 << 20) - 1)) >= report_size) {
	u32 *report32 = (void *)(stream->oa_buffer.vaddr + tail);

	if (report32[0] != 0 || report32[1] != 0)
		break;

	tail = (tail - report_size) & (OA_BUFFER_SIZE(16 << 20) - 1);
}

if (OA_TAKEN(hw_tail, tail)((hw_tail - tail) & ((16 << 20) - 1)) > report_size &&
    __ratelimit(&stream->perf->tail_pointer_race))
	DRM_NOTE("unlanded report(s) head=0x%x "printk("\0015" "[" "drm" "] " "unlanded report(s) head=0x%x "
 "tail=0x%x hw_tail=0x%x\n", head, tail, hw_tail)
		 "tail=0x%x hw_tail=0x%x\n",printk("\0015" "[" "drm" "] " "unlanded report(s) head=0x%x "
 "tail=0x%x hw_tail=0x%x\n", head, tail, hw_tail)
		 head, tail, hw_tail)printk("\0015" "[" "drm" "] " "unlanded report(s) head=0x%x "
 "tail=0x%x hw_tail=0x%x\n", head, tail, hw_tail);

stream->oa_buffer.tail = gtt_offset + tail;
stream->oa_buffer.aging_tail = gtt_offset + hw_tail;
stream->oa_buffer.aging_timestamp = now;
}

pollin = OA_TAKEN(stream->oa_buffer.tail - gtt_offset,((stream->oa_buffer.tail - gtt_offset - stream->oa_buffer
.head - gtt_offset) & ((16 << 20) - 1))
	  stream->oa_buffer.head - gtt_offset)((stream->oa_buffer.tail - gtt_offset - stream->oa_buffer
.head - gtt_offset) & ((16 << 20) - 1)) >= report_size;

spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags)do { (void)(flags); mtx_leave(&stream->oa_buffer.ptr_lock
); } while (0);

return pollin;
551}

553#endif

555/**
* append_oa_status - Appends a status record to a userspace read() buffer.
* @stream: An i915-perf stream opened for OA metrics
* @buf: destination buffer given by userspace
* @count: the number of bytes userspace wants to read
* @offset: (inout): the current position for writing into @buf
* @type: The kind of status to report to userspace
*
* Writes a status record (such as `DRM_I915_PERF_RECORD_OA_REPORT_LOST`)
* into the userspace read() buffer.
*
* The @buf @offset will only be updated on success.
*
* Returns: 0 on success, negative error code on failure.
*/
570static int append_oa_status(struct i915_perf_stream *stream,
	    char __user *buf,
	    size_t count,
	    size_t *offset,
	    enum drm_i915_perf_record_type type)
575{
struct drm_i915_perf_record_header header = { type, 0, sizeof(header) };

if ((count - *offset) < header.size)
return -ENOSPC28;

if (copy_to_user(buf + *offset, &header, sizeof(header)))
return -EFAULT14;

(*offset) += header.size;

return 0;
587}

589/**
* append_oa_sample - Copies single OA report into userspace read() buffer.
* @stream: An i915-perf stream opened for OA metrics
* @buf: destination buffer given by userspace
* @count: the number of bytes userspace wants to read
* @offset: (inout): the current position for writing into @buf
* @report: A single OA report to (optionally) include as part of the sample
*
* The contents of a sample are configured through `DRM_I915_PERF_PROP_SAMPLE_*`
* properties when opening a stream, tracked as `stream->sample_flags`. This
* function copies the requested components of a single sample to the given
* read() @buf.
*
* The @buf @offset will only be updated on success.
*
* Returns: 0 on success, negative error code on failure.
*/
606static int append_oa_sample(struct i915_perf_stream *stream,
	    char __user *buf,
	    size_t count,
	    size_t *offset,
	    const u8 *report)
611{
int report_size = stream->oa_buffer.format_size;
struct drm_i915_perf_record_header header;

header.type = DRM_I915_PERF_RECORD_SAMPLE;
header.pad = 0;
header.size = stream->sample_size;

if ((count - *offset) < header.size)
return -ENOSPC28;

buf += *offset;
if (copy_to_user(buf, &header, sizeof(header)))
return -EFAULT14;
buf += sizeof(header);

if (copy_to_user(buf, report, report_size))
return -EFAULT14;

(*offset) += header.size;

return 0;
633}

635/**
* gen8_append_oa_reports - Copies all buffered OA reports into
*			    userspace read() buffer.
* @stream: An i915-perf stream opened for OA metrics
* @buf: destination buffer given by userspace
* @count: the number of bytes userspace wants to read
* @offset: (inout): the current position for writing into @buf
*
* Notably any error condition resulting in a short read (-%ENOSPC or
* -%EFAULT) will be returned even though one or more records may
* have been successfully copied. In this case it's up to the caller
* to decide if the error should be squashed before returning to
* userspace.
*
* Note: reports are consumed from the head, and appended to the
* tail, so the tail chases the head?... If you think that's mad
* and back-to-front you're not alone, but this follows the
* Gen PRM naming convention.
*
* Returns: 0 on success, negative error code on failure.
*/
656static int gen8_append_oa_reports(struct i915_perf_stream *stream,
		  char __user *buf,
		  size_t count,
		  size_t *offset)
660{
struct intel_uncore *uncore = stream->uncore;
int report_size = stream->oa_buffer.format_size;
u8 *oa_buf_base = stream->oa_buffer.vaddr;
u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
u32 mask = (OA_BUFFER_SIZE(16 << 20) - 1);
size_t start_offset = *offset;
unsigned long flags;
u32 head, tail;
u32 taken;
int ret = 0;

if (drm_WARN_ON(&uncore->i915->drm, !stream->enabled)({ int __ret = !!((!stream->enabled)); if (__ret) printf("%s %s: "
 "%s", dev_driver_string(((&uncore->i915->drm))->
dev), "", "drm_WARN_ON(" "!stream->enabled" ")"); __builtin_expect
(!!(__ret), 0); }))
return -EIO5;

spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags)do { flags = 0; mtx_enter(&stream->oa_buffer.ptr_lock)
; } while (0);

head = stream->oa_buffer.head;
tail = stream->oa_buffer.tail;

spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags)do { (void)(flags); mtx_leave(&stream->oa_buffer.ptr_lock
); } while (0);

/*
* NB: oa_buffer.head/tail include the gtt_offset which we don't want
* while indexing relative to oa_buf_base.
*/
head -= gtt_offset;
tail -= gtt_offset;

/*
* An out of bounds or misaligned head or tail pointer implies a driver
* bug since we validate + align the tail pointers we read from the
* hardware and we are in full control of the head pointer which should
* only be incremented by multiples of the report size (notably also
* all a power of two).
*/
if (drm_WARN_ONCE(&uncore->i915->drm,({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); })
	  head > OA_BUFFER_SIZE || head % report_size ||({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); })
	  tail > OA_BUFFER_SIZE || tail % report_size,({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); })
	  "Inconsistent OA buffer pointers: head = %u, tail = %u\n",({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); })
	  head, tail)({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); }))
return -EIO5;


for (/* none */;
    (taken = OA_TAKEN(tail, head)((tail - head) & ((16 << 20) - 1)));
    head = (head + report_size) & mask) {
u8 *report = oa_buf_base + head;
u32 *report32 = (void *)report;
u32 ctx_id;
u32 reason;

/*
 * All the report sizes factor neatly into the buffer
 * size so we never expect to see a report split
 * between the beginning and end of the buffer.
 *
 * Given the initial alignment check a misalignment
 * here would imply a driver bug that would result
 * in an overrun.
 */
if (drm_WARN_ON(&uncore->i915->drm,({ int __ret = !!((((16 << 20) - head) < report_size
)); if (__ret) printf("%s %s: " "%s", dev_driver_string(((&
uncore->i915->drm))->dev), "", "drm_WARN_ON(" "((16 << 20) - head) < report_size"
 ")"); __builtin_expect(!!(__ret), 0); })
		(OA_BUFFER_SIZE - head) < report_size)({ int __ret = !!((((16 << 20) - head) < report_size
)); if (__ret) printf("%s %s: " "%s", dev_driver_string(((&
uncore->i915->drm))->dev), "", "drm_WARN_ON(" "((16 << 20) - head) < report_size"
 ")"); __builtin_expect(!!(__ret), 0); })) {
	drm_err(&uncore->i915->drm,printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "Spurious OA head ptr: non-integral report offset\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__)
		"Spurious OA head ptr: non-integral report offset\n")printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "Spurious OA head ptr: non-integral report offset\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__);
	break;
}

/*
 * The reason field includes flags identifying what
 * triggered this specific report (mostly timer
 * triggered or e.g. due to a context switch).
 *
 * This field is never expected to be zero so we can
 * check that the report isn't invalid before copying
 * it to userspace...
 */
reason = ((report32[0] >> OAREPORT_REASON_SHIFT19) &
	  (GRAPHICS_VER(stream->perf->i915)((&(stream->perf->i915)->__runtime)->graphics
.ip.ver) == 12 ?
	   OAREPORT_REASON_MASK_EXTENDED0x7f :
	   OAREPORT_REASON_MASK0x3f));

ctx_id = report32[2] & stream->specific_ctx_id_mask;

/*
 * Squash whatever is in the CTX_ID field if it's marked as
 * invalid to be sure we avoid false-positive, single-context
 * filtering below...
 *
 * Note: that we don't clear the valid_ctx_bit so userspace can
 * understand that the ID has been squashed by the kernel.
 */
if (!(report32[0] & stream->perf->gen8_valid_ctx_bit) &&
    GRAPHICS_VER(stream->perf->i915)((&(stream->perf->i915)->__runtime)->graphics
.ip.ver) <= 11)
	ctx_id = report32[2] = INVALID_CTX_ID0xffffffff;

/*
 * NB: For Gen 8 the OA unit no longer supports clock gating
 * off for a specific context and the kernel can't securely
 * stop the counters from updating as system-wide / global
 * values.
 *
 * Automatic reports now include a context ID so reports can be
 * filtered on the cpu but it's not worth trying to
 * automatically subtract/hide counter progress for other
 * contexts while filtering since we can't stop userspace
 * issuing MI_REPORT_PERF_COUNT commands which would still
 * provide a side-band view of the real values.
 *
 * To allow userspace (such as Mesa/GL_INTEL_performance_query)
 * to normalize counters for a single filtered context then it
 * needs be forwarded bookend context-switch reports so that it
 * can track switches in between MI_REPORT_PERF_COUNT commands
 * and can itself subtract/ignore the progress of counters
 * associated with other contexts. Note that the hardware
 * automatically triggers reports when switching to a new
 * context which are tagged with the ID of the newly active
 * context. To avoid the complexity (and likely fragility) of
 * reading ahead while parsing reports to try and minimize
 * forwarding redundant context switch reports (i.e. between
 * other, unrelated contexts) we simply elect to forward them
 * all.
 *
 * We don't rely solely on the reason field to identify context
 * switches since it's not-uncommon for periodic samples to
 * identify a switch before any 'context switch' report.
 */
if (!stream->perf->exclusive_stream->ctx ||
    stream->specific_ctx_id == ctx_id ||
    stream->oa_buffer.last_ctx_id == stream->specific_ctx_id ||
    reason & OAREPORT_REASON_CTX_SWITCH(1<<3)) {

	/*
	 * While filtering for a single context we avoid
	 * leaking the IDs of other contexts.
	 */
	if (stream->perf->exclusive_stream->ctx &&
	    stream->specific_ctx_id != ctx_id) {
		report32[2] = INVALID_CTX_ID0xffffffff;
	}

	ret = append_oa_sample(stream, buf, count, offset,
			       report);
	if (ret)
		break;

	stream->oa_buffer.last_ctx_id = ctx_id;
}

/*
 * Clear out the first 2 dword as a mean to detect unlanded
 * reports.
 */
report32[0] = 0;
report32[1] = 0;
}

if (start_offset != *offset) {
i915_reg_t oaheadptr;

oaheadptr = GRAPHICS_VER(stream->perf->i915)((&(stream->perf->i915)->__runtime)->graphics
.ip.ver) == 12 ?
	    GEN12_OAG_OAHEADPTR((const i915_reg_t){ .reg = (0xdb00) }) : GEN8_OAHEADPTR((const i915_reg_t){ .reg = (0x2B0C) });

spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags)do { flags = 0; mtx_enter(&stream->oa_buffer.ptr_lock)
; } while (0);

/*
 * We removed the gtt_offset for the copy loop above, indexing
 * relative to oa_buf_base so put back here...
 */
head += gtt_offset;
intel_uncore_write(uncore, oaheadptr,
		   head & GEN12_OAG_OAHEADPTR_MASK0xffffffc0);
stream->oa_buffer.head = head;

spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags)do { (void)(flags); mtx_leave(&stream->oa_buffer.ptr_lock
); } while (0);
}

return ret;
838}

840/**
* gen8_oa_read - copy status records then buffered OA reports
* @stream: An i915-perf stream opened for OA metrics
* @buf: destination buffer given by userspace
* @count: the number of bytes userspace wants to read
* @offset: (inout): the current position for writing into @buf
*
* Checks OA unit status registers and if necessary appends corresponding
* status records for userspace (such as for a buffer full condition) and then
* initiate appending any buffered OA reports.
*
* Updates @offset according to the number of bytes successfully copied into
* the userspace buffer.
*
* NB: some data may be successfully copied to the userspace buffer
* even if an error is returned, and this is reflected in the
* updated @offset.
*
* Returns: zero on success or a negative error code
*/
860static int gen8_oa_read(struct i915_perf_stream *stream,
	char __user *buf,
	size_t count,
	size_t *offset)
864{
struct intel_uncore *uncore = stream->uncore;
u32 oastatus;
i915_reg_t oastatus_reg;
int ret;

if (drm_WARN_ON(&uncore->i915->drm, !stream->oa_buffer.vaddr)({ int __ret = !!((!stream->oa_buffer.vaddr)); if (__ret) printf
("%s %s: " "%s", dev_driver_string(((&uncore->i915->
drm))->dev), "", "drm_WARN_ON(" "!stream->oa_buffer.vaddr"
 ")"); __builtin_expect(!!(__ret), 0); }))
return -EIO5;

oastatus_reg = GRAPHICS_VER(stream->perf->i915)((&(stream->perf->i915)->__runtime)->graphics
.ip.ver) == 12 ?
       GEN12_OAG_OASTATUS((const i915_reg_t){ .reg = (0xdafc) }) : GEN8_OASTATUS((const i915_reg_t){ .reg = (0x2b08) });

oastatus = intel_uncore_read(uncore, oastatus_reg);

/*
* We treat OABUFFER_OVERFLOW as a significant error:
*
* Although theoretically we could handle this more gracefully
* sometimes, some Gens don't correctly suppress certain
* automatically triggered reports in this condition and so we
* have to assume that old reports are now being trampled
* over.
*
* Considering how we don't currently give userspace control
* over the OA buffer size and always configure a large 16MB
* buffer, then a buffer overflow does anyway likely indicate
* that something has gone quite badly wrong.
*/
if (oastatus & GEN8_OASTATUS_OABUFFER_OVERFLOW(1 << 1)) {
ret = append_oa_status(stream, buf, count, offset,
		       DRM_I915_PERF_RECORD_OA_BUFFER_LOST);
if (ret)
	return ret;

drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA buffer overflow (exponent = %d): force restart\n"
, stream->period_exponent)
	"OA buffer overflow (exponent = %d): force restart\n",__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA buffer overflow (exponent = %d): force restart\n"
, stream->period_exponent)
	stream->period_exponent)__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA buffer overflow (exponent = %d): force restart\n"
, stream->period_exponent);

stream->perf->ops.oa_disable(stream);
stream->perf->ops.oa_enable(stream);

/*
 * Note: .oa_enable() is expected to re-init the oabuffer and
 * reset GEN8_OASTATUS for us
 */
oastatus = intel_uncore_read(uncore, oastatus_reg);
}

if (oastatus & GEN8_OASTATUS_REPORT_LOST(1 << 0)) {
ret = append_oa_status(stream, buf, count, offset,
		       DRM_I915_PERF_RECORD_OA_REPORT_LOST);
if (ret)
	return ret;

intel_uncore_rmw(uncore, oastatus_reg,
		 GEN8_OASTATUS_COUNTER_OVERFLOW(1 << 2) |
		 GEN8_OASTATUS_REPORT_LOST(1 << 0),
		 IS_GRAPHICS_VER(uncore->i915, 8, 11)(((&(uncore->i915)->__runtime)->graphics.ip.ver)
 >= (8) && ((&(uncore->i915)->__runtime)
->graphics.ip.ver) <= (11)) ?
		 (GEN8_OASTATUS_HEAD_POINTER_WRAP(1 << 16) |
		  GEN8_OASTATUS_TAIL_POINTER_WRAP(1 << 17)) : 0);
}

return gen8_append_oa_reports(stream, buf, count, offset);
927}

929/**
* gen7_append_oa_reports - Copies all buffered OA reports into
*			    userspace read() buffer.
* @stream: An i915-perf stream opened for OA metrics
* @buf: destination buffer given by userspace
* @count: the number of bytes userspace wants to read
* @offset: (inout): the current position for writing into @buf
*
* Notably any error condition resulting in a short read (-%ENOSPC or
* -%EFAULT) will be returned even though one or more records may
* have been successfully copied. In this case it's up to the caller
* to decide if the error should be squashed before returning to
* userspace.
*
* Note: reports are consumed from the head, and appended to the
* tail, so the tail chases the head?... If you think that's mad
* and back-to-front you're not alone, but this follows the
* Gen PRM naming convention.
*
* Returns: 0 on success, negative error code on failure.
*/
950static int gen7_append_oa_reports(struct i915_perf_stream *stream,
		  char __user *buf,
		  size_t count,
		  size_t *offset)
954{
struct intel_uncore *uncore = stream->uncore;
int report_size = stream->oa_buffer.format_size;
u8 *oa_buf_base = stream->oa_buffer.vaddr;
u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
u32 mask = (OA_BUFFER_SIZE(16 << 20) - 1);
size_t start_offset = *offset;
unsigned long flags;
u32 head, tail;
u32 taken;
int ret = 0;

if (drm_WARN_ON(&uncore->i915->drm, !stream->enabled)({ int __ret = !!((!stream->enabled)); if (__ret) printf("%s %s: "
 "%s", dev_driver_string(((&uncore->i915->drm))->
dev), "", "drm_WARN_ON(" "!stream->enabled" ")"); __builtin_expect
(!!(__ret), 0); }))
return -EIO5;

spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags)do { flags = 0; mtx_enter(&stream->oa_buffer.ptr_lock)
; } while (0);

head = stream->oa_buffer.head;
tail = stream->oa_buffer.tail;

spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags)do { (void)(flags); mtx_leave(&stream->oa_buffer.ptr_lock
); } while (0);

/* NB: oa_buffer.head/tail include the gtt_offset which we don't want
* while indexing relative to oa_buf_base.
*/
head -= gtt_offset;
tail -= gtt_offset;

/* An out of bounds or misaligned head or tail pointer implies a driver
* bug since we validate + align the tail pointers we read from the
* hardware and we are in full control of the head pointer which should
* only be incremented by multiples of the report size (notably also
* all a power of two).
*/
if (drm_WARN_ONCE(&uncore->i915->drm,({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); })
	  head > OA_BUFFER_SIZE || head % report_size ||({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); })
	  tail > OA_BUFFER_SIZE || tail % report_size,({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); })
	  "Inconsistent OA buffer pointers: head = %u, tail = %u\n",({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); })
	  head, tail)({ static int __warned; int __ret = !!(head > (16 <<
 20) || head % report_size || tail > (16 << 20) || tail
 % report_size); if (__ret && !__warned) { printf("%s %s: "
 "Inconsistent OA buffer pointers: head = %u, tail = %u\n", dev_driver_string
((&uncore->i915->drm)->dev), "", head, tail); __warned
 = 1; } __builtin_expect(!!(__ret), 0); }))
return -EIO5;


for (/* none */;
    (taken = OA_TAKEN(tail, head)((tail - head) & ((16 << 20) - 1)));
    head = (head + report_size) & mask) {
u8 *report = oa_buf_base + head;
u32 *report32 = (void *)report;

/* All the report sizes factor neatly into the buffer
 * size so we never expect to see a report split
 * between the beginning and end of the buffer.
 *
 * Given the initial alignment check a misalignment
 * here would imply a driver bug that would result
 * in an overrun.
 */
if (drm_WARN_ON(&uncore->i915->drm,({ int __ret = !!((((16 << 20) - head) < report_size
)); if (__ret) printf("%s %s: " "%s", dev_driver_string(((&
uncore->i915->drm))->dev), "", "drm_WARN_ON(" "((16 << 20) - head) < report_size"
 ")"); __builtin_expect(!!(__ret), 0); })
		(OA_BUFFER_SIZE - head) < report_size)({ int __ret = !!((((16 << 20) - head) < report_size
)); if (__ret) printf("%s %s: " "%s", dev_driver_string(((&
uncore->i915->drm))->dev), "", "drm_WARN_ON(" "((16 << 20) - head) < report_size"
 ")"); __builtin_expect(!!(__ret), 0); })) {
	drm_err(&uncore->i915->drm,printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "Spurious OA head ptr: non-integral report offset\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__)
		"Spurious OA head ptr: non-integral report offset\n")printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "Spurious OA head ptr: non-integral report offset\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__);
	break;
}

/* The report-ID field for periodic samples includes
 * some undocumented flags related to what triggered
 * the report and is never expected to be zero so we
 * can check that the report isn't invalid before
 * copying it to userspace...
 */
if (report32[0] == 0) {
	if (__ratelimit(&stream->perf->spurious_report_rs))
		DRM_NOTE("Skipping spurious, invalid OA report\n")printk("\0015" "[" "drm" "] " "Skipping spurious, invalid OA report\n"
);
	continue;
}

ret = append_oa_sample(stream, buf, count, offset, report);
if (ret)
	break;

/* Clear out the first 2 dwords as a mean to detect unlanded
 * reports.
 */
report32[0] = 0;
report32[1] = 0;
}

if (start_offset != *offset) {
spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags)do { flags = 0; mtx_enter(&stream->oa_buffer.ptr_lock)
; } while (0);

/* We removed the gtt_offset for the copy loop above, indexing
 * relative to oa_buf_base so put back here...
 */
head += gtt_offset;

intel_uncore_write(uncore, GEN7_OASTATUS2((const i915_reg_t){ .reg = (0x2368) }),
		   (head & GEN7_OASTATUS2_HEAD_MASK0xffffffc0) |
		   GEN7_OASTATUS2_MEM_SELECT_GGTT(1 << 0));
stream->oa_buffer.head = head;

spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags)do { (void)(flags); mtx_leave(&stream->oa_buffer.ptr_lock
); } while (0);
}

return ret;
1057}

1059/**
* gen7_oa_read - copy status records then buffered OA reports
* @stream: An i915-perf stream opened for OA metrics
* @buf: destination buffer given by userspace
* @count: the number of bytes userspace wants to read
* @offset: (inout): the current position for writing into @buf
*
* Checks Gen 7 specific OA unit status registers and if necessary appends
* corresponding status records for userspace (such as for a buffer full
* condition) and then initiate appending any buffered OA reports.
*
* Updates @offset according to the number of bytes successfully copied into
* the userspace buffer.
*
* Returns: zero on success or a negative error code
*/
1075static int gen7_oa_read(struct i915_perf_stream *stream,
	char __user *buf,
	size_t count,
	size_t *offset)
1079{
struct intel_uncore *uncore = stream->uncore;
u32 oastatus1;
int ret;

if (drm_WARN_ON(&uncore->i915->drm, !stream->oa_buffer.vaddr)({ int __ret = !!((!stream->oa_buffer.vaddr)); if (__ret) printf
("%s %s: " "%s", dev_driver_string(((&uncore->i915->
drm))->dev), "", "drm_WARN_ON(" "!stream->oa_buffer.vaddr"
 ")"); __builtin_expect(!!(__ret), 0); }))
return -EIO5;

oastatus1 = intel_uncore_read(uncore, GEN7_OASTATUS1((const i915_reg_t){ .reg = (0x2364) }));

/* XXX: On Haswell we don't have a safe way to clear oastatus1
* bits while the OA unit is enabled (while the tail pointer
* may be updated asynchronously) so we ignore status bits
* that have already been reported to userspace.
*/
oastatus1 &= ~stream->perf->gen7_latched_oastatus1;

/* We treat OABUFFER_OVERFLOW as a significant error:
*
* - The status can be interpreted to mean that the buffer is
*   currently full (with a higher precedence than OA_TAKEN()
*   which will start to report a near-empty buffer after an
*   overflow) but it's awkward that we can't clear the status
*   on Haswell, so without a reset we won't be able to catch
*   the state again.
*
* - Since it also implies the HW has started overwriting old
*   reports it may also affect our sanity checks for invalid
*   reports when copying to userspace that assume new reports
*   are being written to cleared memory.
*
* - In the future we may want to introduce a flight recorder
*   mode where the driver will automatically maintain a safe
*   guard band between head/tail, avoiding this overflow
*   condition, but we avoid the added driver complexity for
*   now.
*/
if (unlikely(oastatus1 & GEN7_OASTATUS1_OABUFFER_OVERFLOW)__builtin_expect(!!(oastatus1 & (1 << 1)), 0)) {
ret = append_oa_status(stream, buf, count, offset,
		       DRM_I915_PERF_RECORD_OA_BUFFER_LOST);
if (ret)
	return ret;

drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA buffer overflow (exponent = %d): force restart\n"
, stream->period_exponent)
	"OA buffer overflow (exponent = %d): force restart\n",__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA buffer overflow (exponent = %d): force restart\n"
, stream->period_exponent)
	stream->period_exponent)__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA buffer overflow (exponent = %d): force restart\n"
, stream->period_exponent);

stream->perf->ops.oa_disable(stream);
stream->perf->ops.oa_enable(stream);

oastatus1 = intel_uncore_read(uncore, GEN7_OASTATUS1((const i915_reg_t){ .reg = (0x2364) }));
}

if (unlikely(oastatus1 & GEN7_OASTATUS1_REPORT_LOST)__builtin_expect(!!(oastatus1 & (1 << 0)), 0)) {
ret = append_oa_status(stream, buf, count, offset,
		       DRM_I915_PERF_RECORD_OA_REPORT_LOST);
if (ret)
	return ret;
stream->perf->gen7_latched_oastatus1 |=
	GEN7_OASTATUS1_REPORT_LOST(1 << 0);
}

return gen7_append_oa_reports(stream, buf, count, offset);
1142}

1144#ifdef notyet

1146/**
* i915_oa_wait_unlocked - handles blocking IO until OA data available
* @stream: An i915-perf stream opened for OA metrics
*
* Called when userspace tries to read() from a blocking stream FD opened
* for OA metrics. It waits until the hrtimer callback finds a non-empty
* OA buffer and wakes us.
*
* Note: it's acceptable to have this return with some false positives
* since any subsequent read handling will return -EAGAIN if there isn't
* really data ready for userspace yet.
*
* Returns: zero on success or a negative error code
*/
1160static int i915_oa_wait_unlocked(struct i915_perf_stream *stream)
1161{
/* We would wait indefinitely if periodic sampling is not enabled */
if (!stream->periodic)
return -EIO5;

return wait_event_interruptible(stream->poll_wq,({ int __ret = 0; if (!(oa_buffer_check_unlocked(stream))) __ret
 = ({ long __ret = 0; struct wait_queue_entry __wq_entry; init_wait_entry
(&__wq_entry, 0); do { int __error, __wait; unsigned long
 deadline; ((!cold) ? (void)0 : __assert("diagnostic ", "/usr/src/sys/dev/pci/drm/i915/i915_perf.c"
, 1167, "!cold")); prepare_to_wait(&stream->poll_wq, &
__wq_entry, 0x100); deadline = jiffies + __ret; __wait = !(oa_buffer_check_unlocked
(stream)); __error = sleep_finish(__ret, __wait); if ((0) >
 0) __ret = deadline - jiffies; if (__error == -1 || __error ==
 4) { __ret = -4; break; } if ((0) > 0 && (__ret <=
|| __error == 35)) { __ret = ((oa_buffer_check_unlocked(stream
))) ? 1 : 0; break; } } while (__ret > 0 && !(oa_buffer_check_unlocked
(stream))); finish_wait(&stream->poll_wq, &__wq_entry
); __ret; }); __ret; })
			oa_buffer_check_unlocked(stream))({ int __ret = 0; if (!(oa_buffer_check_unlocked(stream))) __ret
 = ({ long __ret = 0; struct wait_queue_entry __wq_entry; init_wait_entry
(&__wq_entry, 0); do { int __error, __wait; unsigned long
 deadline; ((!cold) ? (void)0 : __assert("diagnostic ", "/usr/src/sys/dev/pci/drm/i915/i915_perf.c"
, 1167, "!cold")); prepare_to_wait(&stream->poll_wq, &
__wq_entry, 0x100); deadline = jiffies + __ret; __wait = !(oa_buffer_check_unlocked
(stream)); __error = sleep_finish(__ret, __wait); if ((0) >
 0) __ret = deadline - jiffies; if (__error == -1 || __error ==
 4) { __ret = -4; break; } if ((0) > 0 && (__ret <=
|| __error == 35)) { __ret = ((oa_buffer_check_unlocked(stream
))) ? 1 : 0; break; } } while (__ret > 0 && !(oa_buffer_check_unlocked
(stream))); finish_wait(&stream->poll_wq, &__wq_entry
); __ret; }); __ret; });
1168}

1170/**
* i915_oa_poll_wait - call poll_wait() for an OA stream poll()
* @stream: An i915-perf stream opened for OA metrics
* @file: An i915 perf stream file
* @wait: poll() state table
*
* For handling userspace polling on an i915 perf stream opened for OA metrics,
* this starts a poll_wait with the wait queue that our hrtimer callback wakes
* when it sees data ready to read in the circular OA buffer.
*/
1180static void i915_oa_poll_wait(struct i915_perf_stream *stream,
	      struct file *file,
	      poll_table *wait)
1183{
poll_wait(file, &stream->poll_wq, wait);
1185}

1187/**
* i915_oa_read - just calls through to &i915_oa_ops->read
* @stream: An i915-perf stream opened for OA metrics
* @buf: destination buffer given by userspace
* @count: the number of bytes userspace wants to read
* @offset: (inout): the current position for writing into @buf
*
* Updates @offset according to the number of bytes successfully copied into
* the userspace buffer.
*
* Returns: zero on success or a negative error code
*/
1199static int i915_oa_read(struct i915_perf_stream *stream,
	char __user *buf,
	size_t count,
	size_t *offset)
1203{
return stream->perf->ops.read(stream, buf, count, offset);
1205}

1207static struct intel_context *oa_pin_context(struct i915_perf_stream *stream)
1208{
struct i915_gem_engines_iter it;
struct i915_gem_context *ctx = stream->ctx;
struct intel_context *ce;
struct i915_gem_ww_ctx ww;
int err = -ENODEV19;

for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it)for (i915_gem_engines_iter_init(&(it), (i915_gem_context_lock_engines
(ctx))); ((ce) = i915_gem_engines_iter_next(&(it)));) {
if (ce->engine != stream->engine) /* first match! */
	continue;

err = 0;
break;
}
i915_gem_context_unlock_engines(ctx);

if (err)
return ERR_PTR(err);

i915_gem_ww_ctx_init(&ww, true1);
1228retry:
/*
* As the ID is the gtt offset of the context's vma we
* pin the vma to ensure the ID remains fixed.
*/
err = intel_context_pin_ww(ce, &ww);
if (err == -EDEADLK11) {
err = i915_gem_ww_ctx_backoff(&ww);
if (!err)
	goto retry;
}
i915_gem_ww_ctx_fini(&ww);

if (err)
return ERR_PTR(err);

stream->pinned_ctx = ce;
return stream->pinned_ctx;
1246}

1248/**
* oa_get_render_ctx_id - determine and hold ctx hw id
* @stream: An i915-perf stream opened for OA metrics
*
* Determine the render context hw id, and ensure it remains fixed for the
* lifetime of the stream. This ensures that we don't have to worry about
* updating the context ID in OACONTROL on the fly.
*
* Returns: zero on success or a negative error code
*/
1258static int oa_get_render_ctx_id(struct i915_perf_stream *stream)
1259{
struct intel_context *ce;

ce = oa_pin_context(stream);
if (IS_ERR(ce))
return PTR_ERR(ce);

switch (GRAPHICS_VER(ce->engine->i915)((&(ce->engine->i915)->__runtime)->graphics.ip
.ver)) {
case 7: {
/*
 * On Haswell we don't do any post processing of the reports
 * and don't need to use the mask.
 */
stream->specific_ctx_id = i915_ggtt_offset(ce->state);
stream->specific_ctx_id_mask = 0;
break;
}

case 8:
case 9:
if (intel_engine_uses_guc(ce->engine)) {
	/*
	 * When using GuC, the context descriptor we write in
	 * i915 is read by GuC and rewritten before it's
	 * actually written into the hardware. The LRCA is
	 * what is put into the context id field of the
	 * context descriptor by GuC. Because it's aligned to
	 * a page, the lower 12bits are always at 0 and
	 * dropped by GuC. They won't be part of the context
	 * ID in the OA reports, so squash those lower bits.
	 */
	stream->specific_ctx_id = ce->lrc.lrca >> 12;

	/*
	 * GuC uses the top bit to signal proxy submission, so
	 * ignore that bit.
	 */
	stream->specific_ctx_id_mask =
		(1U << (GEN8_CTX_ID_WIDTH21 - 1)) - 1;
} else {
	stream->specific_ctx_id_mask =
		(1U << GEN8_CTX_ID_WIDTH21) - 1;
	stream->specific_ctx_id = stream->specific_ctx_id_mask;
}
break;

case 11:
case 12:
if (GRAPHICS_VER_FULL(ce->engine->i915)(((&(ce->engine->i915)->__runtime)->graphics.
ip.ver) << 8 | ((&(ce->engine->i915)->__runtime
)->graphics.ip.rel)) >= IP_VER(12, 50)((12) << 8 | (50))) {
	stream->specific_ctx_id_mask =
		((1U << XEHP_SW_CTX_ID_WIDTH16) - 1) <<
		(XEHP_SW_CTX_ID_SHIFT39 - 32);
	stream->specific_ctx_id =
		(XEHP_MAX_CONTEXT_HW_ID0xFFFF - 1) <<
		(XEHP_SW_CTX_ID_SHIFT39 - 32);
} else {
	stream->specific_ctx_id_mask =
		((1U << GEN11_SW_CTX_ID_WIDTH11) - 1) << (GEN11_SW_CTX_ID_SHIFT37 - 32);
	/*
	 * Pick an unused context id
	 * 0 - BITS_PER_LONG are used by other contexts
	 * GEN12_MAX_CONTEXT_HW_ID (0x7ff) is used by idle context
	 */
	stream->specific_ctx_id =
		(GEN12_MAX_CONTEXT_HW_ID((1 << 11) - 1) - 1) << (GEN11_SW_CTX_ID_SHIFT37 - 32);
}
break;

default:
MISSING_CASE(GRAPHICS_VER(ce->engine->i915))({ int __ret = !!(1); if (__ret) printf("Missing case (%s == %ld)\n"
, "((&(ce->engine->i915)->__runtime)->graphics.ip.ver)"
, (long)(((&(ce->engine->i915)->__runtime)->graphics
.ip.ver))); __builtin_expect(!!(__ret), 0); });
}

ce->tag = stream->specific_ctx_id;

drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "filtering on ctx_id=0x%x ctx_id_mask=0x%x\n"
, stream->specific_ctx_id, stream->specific_ctx_id_mask
)
"filtering on ctx_id=0x%x ctx_id_mask=0x%x\n",__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "filtering on ctx_id=0x%x ctx_id_mask=0x%x\n"
, stream->specific_ctx_id, stream->specific_ctx_id_mask
)
stream->specific_ctx_id,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "filtering on ctx_id=0x%x ctx_id_mask=0x%x\n"
, stream->specific_ctx_id, stream->specific_ctx_id_mask
)
stream->specific_ctx_id_mask)__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "filtering on ctx_id=0x%x ctx_id_mask=0x%x\n"
, stream->specific_ctx_id, stream->specific_ctx_id_mask
);

return 0;
1339}

1341/**
* oa_put_render_ctx_id - counterpart to oa_get_render_ctx_id releases hold
* @stream: An i915-perf stream opened for OA metrics
*
* In case anything needed doing to ensure the context HW ID would remain valid
* for the lifetime of the stream, then that can be undone here.
*/
1348static void oa_put_render_ctx_id(struct i915_perf_stream *stream)
1349{
struct intel_context *ce;

ce = fetch_and_zero(&stream->pinned_ctx)({ typeof(*&stream->pinned_ctx) __T = *(&stream->
pinned_ctx); *(&stream->pinned_ctx) = (typeof(*&stream
->pinned_ctx))0; __T; });
if (ce) {
ce->tag = 0; /* recomputed on next submission after parking */
intel_context_unpin(ce);
}

stream->specific_ctx_id = INVALID_CTX_ID0xffffffff;
stream->specific_ctx_id_mask = 0;
1360}

1362static void
1363free_oa_buffer(struct i915_perf_stream *stream)
1364{
i915_vma_unpin_and_release(&stream->oa_buffer.vma,
		   I915_VMA_RELEASE_MAP(1UL << (0)));

stream->oa_buffer.vaddr = NULL((void *)0);
1369}

1371static void
1372free_oa_configs(struct i915_perf_stream *stream)
1373{
struct i915_oa_config_bo *oa_bo, *tmp;

i915_oa_config_put(stream->oa_config);
llist_for_each_entry_safe(oa_bo, tmp, stream->oa_config_bos.first, node)for (oa_bo = ({ const __typeof( ((__typeof(*oa_bo) *)0)->node
 ) *__mptr = ((stream->oa_config_bos.first)); (__typeof(*oa_bo
) *)( (char *)__mptr - __builtin_offsetof(__typeof(*oa_bo), node
) );}); ((char *)(oa_bo) + __builtin_offsetof(typeof(*(oa_bo)
), node)) != ((void *)0) && (tmp = ({ const __typeof(
 ((__typeof(*oa_bo) *)0)->node ) *__mptr = (oa_bo->node
.next); (__typeof(*oa_bo) *)( (char *)__mptr - __builtin_offsetof
(__typeof(*oa_bo), node) );}), oa_bo); oa_bo = tmp)
free_oa_config_bo(oa_bo);
1379}

1381static void
1382free_noa_wait(struct i915_perf_stream *stream)
1383{
i915_vma_unpin_and_release(&stream->noa_wait, 0);
1385}

1387static void i915_oa_stream_destroy(struct i915_perf_stream *stream)
1388{
struct i915_perf *perf = stream->perf;

if (WARN_ON(stream != perf->exclusive_stream)({ int __ret = !!(stream != perf->exclusive_stream); if (__ret
) printf("WARNING %s failed at %s:%d\n", "stream != perf->exclusive_stream"
, "/usr/src/sys/dev/pci/drm/i915/i915_perf.c", 1391); __builtin_expect
(!!(__ret), 0); }))
return;

/*
* Unset exclusive_stream first, it will be checked while disabling
* the metric set on gen8+.
*
* See i915_oa_init_reg_state() and lrc_configure_all_contexts()
*/
WRITE_ONCE(perf->exclusive_stream, NULL)({ typeof(perf->exclusive_stream) __tmp = (((void *)0)); *
(volatile typeof(perf->exclusive_stream) *)&(perf->
exclusive_stream) = __tmp; __tmp; });
perf->ops.disable_metric_set(stream);

free_oa_buffer(stream);

intel_uncore_forcewake_put(stream->uncore, FORCEWAKE_ALL);
intel_engine_pm_put(stream->engine);

if (stream->ctx)
oa_put_render_ctx_id(stream);

free_oa_configs(stream);
free_noa_wait(stream);

if (perf->spurious_report_rs.missed) {
DRM_NOTE("%d spurious OA report notices suppressed due to ratelimiting\n",printk("\0015" "[" "drm" "] " "%d spurious OA report notices suppressed due to ratelimiting\n"
, perf->spurious_report_rs.missed)
	 perf->spurious_report_rs.missed)printk("\0015" "[" "drm" "] " "%d spurious OA report notices suppressed due to ratelimiting\n"
, perf->spurious_report_rs.missed);
}
1418}

1420#endif

1422static void gen7_init_oa_buffer(struct i915_perf_stream *stream)
1423{
struct intel_uncore *uncore = stream->uncore;
u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
unsigned long flags;

spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags)do { flags = 0; mtx_enter(&stream->oa_buffer.ptr_lock)
; } while (0);

/* Pre-DevBDW: OABUFFER must be set with counters off,
* before OASTATUS1, but after OASTATUS2
*/
intel_uncore_write(uncore, GEN7_OASTATUS2((const i915_reg_t){ .reg = (0x2368) }), /* head */
	   gtt_offset | GEN7_OASTATUS2_MEM_SELECT_GGTT(1 << 0));
stream->oa_buffer.head = gtt_offset;

intel_uncore_write(uncore, GEN7_OABUFFER((const i915_reg_t){ .reg = (0x23B0) }), gtt_offset);

intel_uncore_write(uncore, GEN7_OASTATUS1((const i915_reg_t){ .reg = (0x2364) }), /* tail */
	   gtt_offset | OABUFFER_SIZE_16M(7 << 3));

/* Mark that we need updated tail pointers to read from... */
stream->oa_buffer.aging_tail = INVALID_TAIL_PTR0xffffffff;
stream->oa_buffer.tail = gtt_offset;

spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags)do { (void)(flags); mtx_leave(&stream->oa_buffer.ptr_lock
); } while (0);

/* On Haswell we have to track which OASTATUS1 flags we've
* already seen since they can't be cleared while periodic
* sampling is enabled.
*/
stream->perf->gen7_latched_oastatus1 = 0;

/* NB: although the OA buffer will initially be allocated
* zeroed via shmfs (and so this memset is redundant when
* first allocating), we may re-init the OA buffer, either
* when re-enabling a stream or in error/reset paths.
*
* The reason we clear the buffer for each re-init is for the
* sanity check in gen7_append_oa_reports() that looks at the
* report-id field to make sure it's non-zero which relies on
* the assumption that new reports are being written to zeroed
* memory...
*/
memset(stream->oa_buffer.vaddr, 0, OA_BUFFER_SIZE)__builtin_memset((stream->oa_buffer.vaddr), (0), ((16 <<
 20)));
1466}

1468static void gen8_init_oa_buffer(struct i915_perf_stream *stream)
1469{
struct intel_uncore *uncore = stream->uncore;
u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
unsigned long flags;

spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags)do { flags = 0; mtx_enter(&stream->oa_buffer.ptr_lock)
; } while (0);

intel_uncore_write(uncore, GEN8_OASTATUS((const i915_reg_t){ .reg = (0x2b08) }), 0);
intel_uncore_write(uncore, GEN8_OAHEADPTR((const i915_reg_t){ .reg = (0x2B0C) }), gtt_offset);
stream->oa_buffer.head = gtt_offset;

intel_uncore_write(uncore, GEN8_OABUFFER_UDW((const i915_reg_t){ .reg = (0x23b4) }), 0);

/*
* PRM says:
*
*  "This MMIO must be set before the OATAILPTR
*  register and after the OAHEADPTR register. This is
*  to enable proper functionality of the overflow
*  bit."
*/
intel_uncore_write(uncore, GEN8_OABUFFER((const i915_reg_t){ .reg = (0x2b14) }), gtt_offset |
   OABUFFER_SIZE_16M(7 << 3) | GEN8_OABUFFER_MEM_SELECT_GGTT(1 << 0));
intel_uncore_write(uncore, GEN8_OATAILPTR((const i915_reg_t){ .reg = (0x2B10) }), gtt_offset & GEN8_OATAILPTR_MASK0xffffffc0);

/* Mark that we need updated tail pointers to read from... */
stream->oa_buffer.aging_tail = INVALID_TAIL_PTR0xffffffff;
stream->oa_buffer.tail = gtt_offset;

/*
* Reset state used to recognise context switches, affecting which
* reports we will forward to userspace while filtering for a single
* context.
*/
stream->oa_buffer.last_ctx_id = INVALID_CTX_ID0xffffffff;

spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags)do { (void)(flags); mtx_leave(&stream->oa_buffer.ptr_lock
); } while (0);

/*
* NB: although the OA buffer will initially be allocated
* zeroed via shmfs (and so this memset is redundant when
* first allocating), we may re-init the OA buffer, either
* when re-enabling a stream or in error/reset paths.
*
* The reason we clear the buffer for each re-init is for the
* sanity check in gen8_append_oa_reports() that looks at the
* reason field to make sure it's non-zero which relies on
* the assumption that new reports are being written to zeroed
* memory...
*/
memset(stream->oa_buffer.vaddr, 0, OA_BUFFER_SIZE)__builtin_memset((stream->oa_buffer.vaddr), (0), ((16 <<
 20)));
1520}

1522static void gen12_init_oa_buffer(struct i915_perf_stream *stream)
1523{
struct intel_uncore *uncore = stream->uncore;
u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
unsigned long flags;

spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags)do { flags = 0; mtx_enter(&stream->oa_buffer.ptr_lock)
; } while (0);

intel_uncore_write(uncore, GEN12_OAG_OASTATUS((const i915_reg_t){ .reg = (0xdafc) }), 0);
intel_uncore_write(uncore, GEN12_OAG_OAHEADPTR((const i915_reg_t){ .reg = (0xdb00) }),
	   gtt_offset & GEN12_OAG_OAHEADPTR_MASK0xffffffc0);
stream->oa_buffer.head = gtt_offset;

/*
* PRM says:
*
*  "This MMIO must be set before the OATAILPTR
*  register and after the OAHEADPTR register. This is
*  to enable proper functionality of the overflow
*  bit."
*/
intel_uncore_write(uncore, GEN12_OAG_OABUFFER((const i915_reg_t){ .reg = (0xdb08) }), gtt_offset |
	   OABUFFER_SIZE_16M(7 << 3) | GEN8_OABUFFER_MEM_SELECT_GGTT(1 << 0));
intel_uncore_write(uncore, GEN12_OAG_OATAILPTR((const i915_reg_t){ .reg = (0xdb04) }),
	   gtt_offset & GEN12_OAG_OATAILPTR_MASK0xffffffc0);

/* Mark that we need updated tail pointers to read from... */
stream->oa_buffer.aging_tail = INVALID_TAIL_PTR0xffffffff;
stream->oa_buffer.tail = gtt_offset;

/*
* Reset state used to recognise context switches, affecting which
* reports we will forward to userspace while filtering for a single
* context.
*/
stream->oa_buffer.last_ctx_id = INVALID_CTX_ID0xffffffff;

spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags)do { (void)(flags); mtx_leave(&stream->oa_buffer.ptr_lock
); } while (0);

/*
* NB: although the OA buffer will initially be allocated
* zeroed via shmfs (and so this memset is redundant when
* first allocating), we may re-init the OA buffer, either
* when re-enabling a stream or in error/reset paths.
*
* The reason we clear the buffer for each re-init is for the
* sanity check in gen8_append_oa_reports() that looks at the
* reason field to make sure it's non-zero which relies on
* the assumption that new reports are being written to zeroed
* memory...
*/
memset(stream->oa_buffer.vaddr, 0,__builtin_memset((stream->oa_buffer.vaddr), (0), (stream->
oa_buffer.vma->size))
      stream->oa_buffer.vma->size)__builtin_memset((stream->oa_buffer.vaddr), (0), (stream->
oa_buffer.vma->size));
1575}

1577#ifdef notyet

1579static int alloc_oa_buffer(struct i915_perf_stream *stream)
1580{
struct drm_i915_privateinteldrm_softc *i915 = stream->perf->i915;
struct drm_i915_gem_object *bo;
struct i915_vma *vma;
int ret;

if (drm_WARN_ON(&i915->drm, stream->oa_buffer.vma)({ int __ret = !!((stream->oa_buffer.vma)); if (__ret) printf
("%s %s: " "%s", dev_driver_string(((&i915->drm))->
dev), "", "drm_WARN_ON(" "stream->oa_buffer.vma" ")"); __builtin_expect
(!!(__ret), 0); }))
return -ENODEV19;

BUILD_BUG_ON_NOT_POWER_OF_2(OA_BUFFER_SIZE)0;
BUILD_BUG_ON(OA_BUFFER_SIZE < SZ_128K || OA_BUFFER_SIZE > SZ_16M)extern char _ctassert[(!((16 << 20) < (128 << 10
) || (16 << 20) > (16 << 20))) ? 1 : -1 ] __attribute__
((__unused__));

bo = i915_gem_object_create_shmem(stream->perf->i915, OA_BUFFER_SIZE(16 << 20));
if (IS_ERR(bo)) {
drm_err(&i915->drm, "Failed to allocate OA buffer\n")printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "Failed to allocate OA buffer\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__);
return PTR_ERR(bo);
}

i915_gem_object_set_cache_coherency(bo, I915_CACHE_LLC);

/* PreHSW required 512K alignment, HSW requires 16M */
vma = i915_gem_object_ggtt_pin(bo, NULL((void *)0), 0, SZ_16M(16 << 20), 0);
if (IS_ERR(vma)) {
ret = PTR_ERR(vma);
goto err_unref;
}
stream->oa_buffer.vma = vma;

stream->oa_buffer.vaddr =
i915_gem_object_pin_map_unlocked(bo, I915_MAP_WB);
if (IS_ERR(stream->oa_buffer.vaddr)) {
ret = PTR_ERR(stream->oa_buffer.vaddr);
goto err_unpin;
}

return 0;

1617err_unpin:
__i915_vma_unpin(vma);

1620err_unref:
i915_gem_object_put(bo);

stream->oa_buffer.vaddr = NULL((void *)0);
stream->oa_buffer.vma = NULL((void *)0);

return ret;
1627}

1629static u32 *save_restore_register(struct i915_perf_stream *stream, u32 *cs,
		  bool_Bool save, i915_reg_t reg, u32 offset,
		  u32 dword_count)
1632{
u32 cmd;
u32 d;

cmd = save ? MI_STORE_REGISTER_MEM(((0x0) << 29) | (0x24) << 23 | (1)) : MI_LOAD_REGISTER_MEM(((0x0) << 29) | (0x29) << 23 | (1));
cmd |= MI_SRM_LRM_GLOBAL_GTT(1<<22);
if (GRAPHICS_VER(stream->perf->i915)((&(stream->perf->i915)->__runtime)->graphics
.ip.ver) >= 8)
cmd++;

for (d = 0; d < dword_count; d++) {
*cs++ = cmd;
*cs++ = i915_mmio_reg_offset(reg) + 4 * d;
*cs++ = intel_gt_scratch_offset(stream->engine->gt,
				offset) + 4 * d;
*cs++ = 0;
}

return cs;
1650}

1652static int alloc_noa_wait(struct i915_perf_stream *stream)
1653{
struct drm_i915_privateinteldrm_softc *i915 = stream->perf->i915;
struct drm_i915_gem_object *bo;
struct i915_vma *vma;
const u64 delay_ticks = 0xffffffffffffffff -
intel_gt_ns_to_clock_interval(to_gt(stream->perf->i915),
atomic64_read(&stream->perf->noa_programming_delay)({ typeof(*(&stream->perf->noa_programming_delay)) __tmp
 = *(volatile typeof(*(&stream->perf->noa_programming_delay
)) *)&(*(&stream->perf->noa_programming_delay))
; membar_datadep_consumer(); __tmp; }));
const u32 base = stream->engine->mmio_base;
1661#define CS_GPR(x) GEN8_RING_CS_GPR(base, x)((const i915_reg_t){ .reg = ((base) + 0x600 + (x) * 8) })
u32 *batch, *ts0, *cs, *jump;
struct i915_gem_ww_ctx ww;
int ret, i;
enum {
START_TS,
NOW_TS,
DELTA_TS,
JUMP_PREDICATE,
DELTA_TARGET,
N_CS_GPR
};

bo = i915_gem_object_create_internal(i915, 4096);
if (IS_ERR(bo)) {
drm_err(&i915->drm,printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "Failed to allocate NOA wait batchbuffer\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__)
	"Failed to allocate NOA wait batchbuffer\n")printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "Failed to allocate NOA wait batchbuffer\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__);
return PTR_ERR(bo);
}

i915_gem_ww_ctx_init(&ww, true1);
1682retry:
ret = i915_gem_object_lock(bo, &ww);
if (ret)
goto out_ww;

/*
* We pin in GGTT because we jump into this buffer now because
* multiple OA config BOs will have a jump to this address and it
* needs to be fixed during the lifetime of the i915/perf stream.
*/
vma = i915_gem_object_ggtt_pin_ww(bo, &ww, NULL((void *)0), 0, 0, PIN_HIGH(1ULL << (5)));
if (IS_ERR(vma)) {
ret = PTR_ERR(vma);
goto out_ww;
}

batch = cs = i915_gem_object_pin_map(bo, I915_MAP_WB);
if (IS_ERR(batch)) {
ret = PTR_ERR(batch);
goto err_unpin;
}

/* Save registers. */
for (i = 0; i < N_CS_GPR; i++)
cs = save_restore_register(
	stream, cs, true1 /* save */, CS_GPR(i),
	INTEL_GT_SCRATCH_FIELD_PERF_CS_GPR + 8 * i, 2);
cs = save_restore_register(
stream, cs, true1 /* save */, MI_PREDICATE_RESULT_1(RENDER_RING_BASE)((const i915_reg_t){ .reg = ((0x02000) + 0x41c) }),
INTEL_GT_SCRATCH_FIELD_PERF_PREDICATE_RESULT_1, 1);

/* First timestamp snapshot location. */
ts0 = cs;

/*
* Initial snapshot of the timestamp register to implement the wait.
* We work with 32b values, so clear out the top 32b bits of the
* register because the ALU works 64bits.
*/
*cs++ = MI_LOAD_REGISTER_IMM(1)(((0x0) << 29) | (0x22) << 23 | (2*(1)-1));
*cs++ = i915_mmio_reg_offset(CS_GPR(START_TS)) + 4;
*cs++ = 0;
*cs++ = MI_LOAD_REGISTER_REG(((0x0) << 29) | (0x2A) << 23 | (1)) | (3 - 2);
*cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(base)((const i915_reg_t){ .reg = ((base) + 0x358) }));
*cs++ = i915_mmio_reg_offset(CS_GPR(START_TS));

/*
* This is the location we're going to jump back into until the
* required amount of time has passed.
*/
jump = cs;

/*
* Take another snapshot of the timestamp register. Take care to clear
* up the top 32bits of CS_GPR(1) as we're using it for other
* operations below.
*/
*cs++ = MI_LOAD_REGISTER_IMM(1)(((0x0) << 29) | (0x22) << 23 | (2*(1)-1));
*cs++ = i915_mmio_reg_offset(CS_GPR(NOW_TS)) + 4;
*cs++ = 0;
*cs++ = MI_LOAD_REGISTER_REG(((0x0) << 29) | (0x2A) << 23 | (1)) | (3 - 2);
*cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(base)((const i915_reg_t){ .reg = ((base) + 0x358) }));
*cs++ = i915_mmio_reg_offset(CS_GPR(NOW_TS));

/*
* Do a diff between the 2 timestamps and store the result back into
* CS_GPR(1).
*/
*cs++ = MI_MATH(5)(((0x0) << 29) | (0x1a) << 23 | ((5) - 1));
*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(NOW_TS))((0x080) << 20 | (0x20) << 10 | ((NOW_TS)));
*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(START_TS))((0x080) << 20 | (0x21) << 10 | ((START_TS)));
*cs++ = MI_MATH_SUB((0x101) << 20 | (0x0) << 10 | (0x0));
*cs++ = MI_MATH_STORE(MI_MATH_REG(DELTA_TS), MI_MATH_REG_ACCU)((0x180) << 20 | ((DELTA_TS)) << 10 | (0x31));
*cs++ = MI_MATH_STORE(MI_MATH_REG(JUMP_PREDICATE), MI_MATH_REG_CF)((0x180) << 20 | ((JUMP_PREDICATE)) << 10 | (0x33
));

/*
* Transfer the carry flag (set to 1 if ts1 < ts0, meaning the
* timestamp have rolled over the 32bits) into the predicate register
* to be used for the predicated jump.
*/
*cs++ = MI_LOAD_REGISTER_REG(((0x0) << 29) | (0x2A) << 23 | (1)) | (3 - 2);
*cs++ = i915_mmio_reg_offset(CS_GPR(JUMP_PREDICATE));
*cs++ = i915_mmio_reg_offset(MI_PREDICATE_RESULT_1(RENDER_RING_BASE)((const i915_reg_t){ .reg = ((0x02000) + 0x41c) }));

/* Restart from the beginning if we had timestamps roll over. */
*cs++ = (GRAPHICS_VER(i915)((&(i915)->__runtime)->graphics.ip.ver) < 8 ?
 MI_BATCH_BUFFER_START(((0x0) << 29) | (0x31) << 23 | (0)) :
 MI_BATCH_BUFFER_START_GEN8(((0x0) << 29) | (0x31) << 23 | (1))) |
MI_BATCH_PREDICATE((u32)((1UL << (15)) + 0));
*cs++ = i915_ggtt_offset(vma) + (ts0 - batch) * 4;
*cs++ = 0;

/*
* Now add the diff between to previous timestamps and add it to :
*      (((1 * << 64) - 1) - delay_ns)
*
* When the Carry Flag contains 1 this means the elapsed time is
* longer than the expected delay, and we can exit the wait loop.
*/
*cs++ = MI_LOAD_REGISTER_IMM(2)(((0x0) << 29) | (0x22) << 23 | (2*(2)-1));
*cs++ = i915_mmio_reg_offset(CS_GPR(DELTA_TARGET));
*cs++ = lower_32_bits(delay_ticks)((u32)(delay_ticks));
*cs++ = i915_mmio_reg_offset(CS_GPR(DELTA_TARGET)) + 4;
*cs++ = upper_32_bits(delay_ticks)((u32)(((delay_ticks) >> 16) >> 16));

*cs++ = MI_MATH(4)(((0x0) << 29) | (0x1a) << 23 | ((4) - 1));
*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(DELTA_TS))((0x080) << 20 | (0x20) << 10 | ((DELTA_TS)));
*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(DELTA_TARGET))((0x080) << 20 | (0x21) << 10 | ((DELTA_TARGET)));
*cs++ = MI_MATH_ADD((0x100) << 20 | (0x0) << 10 | (0x0));
*cs++ = MI_MATH_STOREINV(MI_MATH_REG(JUMP_PREDICATE), MI_MATH_REG_CF)((0x580) << 20 | ((JUMP_PREDICATE)) << 10 | (0x33
));

*cs++ = MI_ARB_CHECK(((0x0) << 29) | (0x05) << 23 | (0));

/*
* Transfer the result into the predicate register to be used for the
* predicated jump.
*/
*cs++ = MI_LOAD_REGISTER_REG(((0x0) << 29) | (0x2A) << 23 | (1)) | (3 - 2);
*cs++ = i915_mmio_reg_offset(CS_GPR(JUMP_PREDICATE));
*cs++ = i915_mmio_reg_offset(MI_PREDICATE_RESULT_1(RENDER_RING_BASE)((const i915_reg_t){ .reg = ((0x02000) + 0x41c) }));

/* Predicate the jump.  */
*cs++ = (GRAPHICS_VER(i915)((&(i915)->__runtime)->graphics.ip.ver) < 8 ?
 MI_BATCH_BUFFER_START(((0x0) << 29) | (0x31) << 23 | (0)) :
 MI_BATCH_BUFFER_START_GEN8(((0x0) << 29) | (0x31) << 23 | (1))) |
MI_BATCH_PREDICATE((u32)((1UL << (15)) + 0));
*cs++ = i915_ggtt_offset(vma) + (jump - batch) * 4;
*cs++ = 0;

/* Restore registers. */
for (i = 0; i < N_CS_GPR; i++)
cs = save_restore_register(
	stream, cs, false0 /* restore */, CS_GPR(i),
	INTEL_GT_SCRATCH_FIELD_PERF_CS_GPR + 8 * i, 2);
cs = save_restore_register(
stream, cs, false0 /* restore */, MI_PREDICATE_RESULT_1(RENDER_RING_BASE)((const i915_reg_t){ .reg = ((0x02000) + 0x41c) }),
INTEL_GT_SCRATCH_FIELD_PERF_PREDICATE_RESULT_1, 1);

/* And return to the ring. */
*cs++ = MI_BATCH_BUFFER_END(((0x0) << 29) | (0x0a) << 23 | (0));

GEM_BUG_ON(cs - batch > PAGE_SIZE / sizeof(*batch))((void)0);

i915_gem_object_flush_map(bo);
__i915_gem_object_release_map(bo);

stream->noa_wait = vma;
goto out_ww;

1831err_unpin:
i915_vma_unpin_and_release(&vma, 0);
1833out_ww:
if (ret == -EDEADLK11) {
ret = i915_gem_ww_ctx_backoff(&ww);
if (!ret)
	goto retry;
}
i915_gem_ww_ctx_fini(&ww);
if (ret)
i915_gem_object_put(bo);
return ret;
1843}

1845#endif

1847static u32 *write_cs_mi_lri(u32 *cs,
	    const struct i915_oa_reg *reg_data,
	    u32 n_regs)
1850{
u32 i;

for (i = 0; i < n_regs; i++) {
if ((i % MI_LOAD_REGISTER_IMM_MAX_REGS(126)) == 0) {
	u32 n_lri = min_t(u32,({ u32 __min_a = (n_regs - i); u32 __min_b = ((126)); __min_a
 < __min_b ? __min_a : __min_b; })
			  n_regs - i,({ u32 __min_a = (n_regs - i); u32 __min_b = ((126)); __min_a
 < __min_b ? __min_a : __min_b; })
			  MI_LOAD_REGISTER_IMM_MAX_REGS)({ u32 __min_a = (n_regs - i); u32 __min_b = ((126)); __min_a
 < __min_b ? __min_a : __min_b; });

	*cs++ = MI_LOAD_REGISTER_IMM(n_lri)(((0x0) << 29) | (0x22) << 23 | (2*(n_lri)-1));
}
*cs++ = i915_mmio_reg_offset(reg_data[i].addr);
*cs++ = reg_data[i].value;
}

return cs;
1866}

1868static int num_lri_dwords(int num_regs)
1869{
int count = 0;

if (num_regs > 0) {
count += DIV_ROUND_UP(num_regs, MI_LOAD_REGISTER_IMM_MAX_REGS)(((num_regs) + (((126)) - 1)) / ((126)));
count += num_regs * 2;
}

return count;
1878}

1880static struct i915_oa_config_bo *
1881alloc_oa_config_buffer(struct i915_perf_stream *stream,
       struct i915_oa_config *oa_config)
1883{
struct drm_i915_gem_object *obj;
struct i915_oa_config_bo *oa_bo;
struct i915_gem_ww_ctx ww;
size_t config_length = 0;
u32 *cs;
int err;

oa_bo = kzalloc(sizeof(*oa_bo), GFP_KERNEL(0x0001 | 0x0004));
14
←
Calling 'kzalloc'→
16
←
Returning from 'kzalloc'→
if (!oa_bo)
17
←
Assuming 'oa_bo' is non-null→
18
←
Taking false branch→
return ERR_PTR(-ENOMEM12);

config_length += num_lri_dwords(oa_config->mux_regs_len);
config_length += num_lri_dwords(oa_config->b_counter_regs_len);
config_length += num_lri_dwords(oa_config->flex_regs_len);
config_length += 3; /* MI_BATCH_BUFFER_START */
config_length = roundup2(sizeof(u32) * config_length, I915_GTT_PAGE_SIZE)(((sizeof(u32) * config_length) + (((1ULL << (12))) - 1
)) & (~((__typeof(sizeof(u32) * config_length))((1ULL <<
 (12))) - 1)));

obj = i915_gem_object_create_shmem(stream->perf->i915, config_length);
if (IS_ERR(obj)) {
19
←
Taking true branch→
err = PTR_ERR(obj);
goto err_free;
20
←
Control jumps to line 1961→
}

i915_gem_ww_ctx_init(&ww, true1);
1908retry:
err = i915_gem_object_lock(obj, &ww);
if (err)
goto out_ww;

cs = i915_gem_object_pin_map(obj, I915_MAP_WB);
if (IS_ERR(cs)) {
err = PTR_ERR(cs);
goto out_ww;
}

cs = write_cs_mi_lri(cs,
	     oa_config->mux_regs,
	     oa_config->mux_regs_len);
cs = write_cs_mi_lri(cs,
	     oa_config->b_counter_regs,
	     oa_config->b_counter_regs_len);
cs = write_cs_mi_lri(cs,
	     oa_config->flex_regs,
	     oa_config->flex_regs_len);

/* Jump into the active wait. */
*cs++ = (GRAPHICS_VER(stream->perf->i915)((&(stream->perf->i915)->__runtime)->graphics
.ip.ver) < 8 ?
 MI_BATCH_BUFFER_START(((0x0) << 29) | (0x31) << 23 | (0)) :
 MI_BATCH_BUFFER_START_GEN8(((0x0) << 29) | (0x31) << 23 | (1)));
*cs++ = i915_ggtt_offset(stream->noa_wait);
*cs++ = 0;

i915_gem_object_flush_map(obj);
__i915_gem_object_release_map(obj);

oa_bo->vma = i915_vma_instance(obj,
		       &stream->engine->gt->ggtt->vm,
		       NULL((void *)0));
if (IS_ERR(oa_bo->vma)) {
err = PTR_ERR(oa_bo->vma);
goto out_ww;
}

oa_bo->oa_config = i915_oa_config_get(oa_config);
llist_add(&oa_bo->node, &stream->oa_config_bos);

1950out_ww:
if (err == -EDEADLK11) {
err = i915_gem_ww_ctx_backoff(&ww);
if (!err)
	goto retry;
}
i915_gem_ww_ctx_fini(&ww);

if (err)
i915_gem_object_put(obj);
1960err_free:
if (err) {
21
←
Assuming 'err' is 0→
22
←
Taking false branch→
kfree(oa_bo);
return ERR_PTR(err);
}
return oa_bo;
1966}

1968static struct i915_vma *
1969get_oa_vma(struct i915_perf_stream *stream, struct i915_oa_config *oa_config)
1970{
struct i915_oa_config_bo *oa_bo;

/*
* Look for the buffer in the already allocated BOs attached
* to the stream.
*/
llist_for_each_entry(oa_bo, stream->oa_config_bos.first, node)for ((oa_bo) = ({ const __typeof( ((__typeof(*(oa_bo)) *)0)->
node ) *__mptr = ((stream->oa_config_bos.first)); (__typeof
(*(oa_bo)) *)( (char *)__mptr - __builtin_offsetof(__typeof(*
(oa_bo)), node) );}); ((char *)(oa_bo) + __builtin_offsetof(typeof
(*(oa_bo)), node)) != ((void *)0); (oa_bo) = ({ const __typeof
( ((__typeof(*(oa_bo)) *)0)->node ) *__mptr = ((oa_bo)->
node.next); (__typeof(*(oa_bo)) *)( (char *)__mptr - __builtin_offsetof
(__typeof(*(oa_bo)), node) );})) {
11
←
Assuming the condition is false→
12
←
Loop condition is false. Execution continues on line 1985→
if (oa_bo->oa_config == oa_config &&
    memcmp(oa_bo->oa_config->uuid,__builtin_memcmp((oa_bo->oa_config->uuid), (oa_config->
uuid), (sizeof(oa_config->uuid)))
	   oa_config->uuid,__builtin_memcmp((oa_bo->oa_config->uuid), (oa_config->
uuid), (sizeof(oa_config->uuid)))
	   sizeof(oa_config->uuid))__builtin_memcmp((oa_bo->oa_config->uuid), (oa_config->
uuid), (sizeof(oa_config->uuid))) == 0)
	goto out;
}

oa_bo = alloc_oa_config_buffer(stream, oa_config);
13
←
Calling 'alloc_oa_config_buffer'→
23
←
Returning from 'alloc_oa_config_buffer'→
if (IS_ERR(oa_bo))
24
←
Taking false branch→
return ERR_CAST(oa_bo);

1989out:
return i915_vma_get(oa_bo->vma);
25
←
Passing null pointer value via 1st parameter 'vma'→
26
←
Calling 'i915_vma_get'→
1991}

1993static int
1994emit_oa_config(struct i915_perf_stream *stream,
      struct i915_oa_config *oa_config,
      struct intel_context *ce,
      struct i915_active *active)
1998{
struct i915_request *rq;
struct i915_vma *vma;
struct i915_gem_ww_ctx ww;
int err;

vma = get_oa_vma(stream, oa_config);
10
←
Calling 'get_oa_vma'→
if (IS_ERR(vma))
return PTR_ERR(vma);

i915_gem_ww_ctx_init(&ww, true1);
2009retry:
err = i915_gem_object_lock(vma->obj, &ww);
if (err)
goto err;

err = i915_vma_pin_ww(vma, &ww, 0, 0, PIN_GLOBAL(1ULL << (10)) | PIN_HIGH(1ULL << (5)));
if (err)
goto err;

intel_engine_pm_get(ce->engine);
rq = i915_request_create(ce);
intel_engine_pm_put(ce->engine);
if (IS_ERR(rq)) {
err = PTR_ERR(rq);
goto err_vma_unpin;
}

if (!IS_ERR_OR_NULL(active)) {
/* After all individual context modifications */
err = i915_request_await_active(rq, active,
				I915_ACTIVE_AWAIT_ACTIVE(1UL << (1)));
if (err)
	goto err_add_request;

err = i915_active_add_request(active, rq);
if (err)
	goto err_add_request;
}

err = i915_request_await_object(rq, vma->obj, 0);
if (!err)
err = i915_vma_move_to_active(vma, rq, 0);
if (err)
goto err_add_request;

err = rq->engine->emit_bb_start(rq,
			vma->node.start, 0,
			I915_DISPATCH_SECURE(1UL << (0)));
if (err)
goto err_add_request;

2050err_add_request:
i915_request_add(rq);
2052err_vma_unpin:
i915_vma_unpin(vma);
2054err:
if (err == -EDEADLK11) {
err = i915_gem_ww_ctx_backoff(&ww);
if (!err)
	goto retry;
}

i915_gem_ww_ctx_fini(&ww);
i915_vma_put(vma);
return err;
2064}

2066static struct intel_context *oa_context(struct i915_perf_stream *stream)
2067{
return stream->pinned_ctx ?: stream->engine->kernel_context;
2069}

2071static int
2072hsw_enable_metric_set(struct i915_perf_stream *stream,
      struct i915_active *active)
2074{
struct intel_uncore *uncore = stream->uncore;

/*
* PRM:
*
* OA unit is using “crclk” for its functionality. When trunk
* level clock gating takes place, OA clock would be gated,
* unable to count the events from non-render clock domain.
* Render clock gating must be disabled when OA is enabled to
* count the events from non-render domain. Unit level clock
* gating for RCS should also be disabled.
*/
intel_uncore_rmw(uncore, GEN7_MISCCPCTL((const i915_reg_t){ .reg = (0x9424) }),
	 GEN7_DOP_CLOCK_GATE_ENABLE(1 << 0), 0);
intel_uncore_rmw(uncore, GEN6_UCGCTL1((const i915_reg_t){ .reg = (0x9400) }),
	 0, GEN6_CSUNIT_CLOCK_GATE_DISABLE(1 << 7));

return emit_oa_config(stream,
	      stream->oa_config, oa_context(stream),
	      active);
2095}

2097static void hsw_disable_metric_set(struct i915_perf_stream *stream)
2098{
struct intel_uncore *uncore = stream->uncore;

intel_uncore_rmw(uncore, GEN6_UCGCTL1((const i915_reg_t){ .reg = (0x9400) }),
	 GEN6_CSUNIT_CLOCK_GATE_DISABLE(1 << 7), 0);
intel_uncore_rmw(uncore, GEN7_MISCCPCTL((const i915_reg_t){ .reg = (0x9424) }),
	 0, GEN7_DOP_CLOCK_GATE_ENABLE(1 << 0));

intel_uncore_rmw(uncore, GDT_CHICKEN_BITS((const i915_reg_t){ .reg = (0x9840) }), GT_NOA_ENABLE0x00000080, 0);
2107}

2109static u32 oa_config_flex_reg(const struct i915_oa_config *oa_config,
	      i915_reg_t reg)
2111{
u32 mmio = i915_mmio_reg_offset(reg);
int i;

/*
* This arbitrary default will select the 'EU FPU0 Pipeline
* Active' event. In the future it's anticipated that there
* will be an explicit 'No Event' we can select, but not yet...
*/
if (!oa_config)
return 0;

for (i = 0; i < oa_config->flex_regs_len; i++) {
if (i915_mmio_reg_offset(oa_config->flex_regs[i].addr) == mmio)
	return oa_config->flex_regs[i].value;
}

return 0;
2129}
2130/*
* NB: It must always remain pointer safe to run this even if the OA unit
* has been disabled.
*
* It's fine to put out-of-date values into these per-context registers
* in the case that the OA unit has been disabled.
*/
2137static void
2138gen8_update_reg_state_unlocked(const struct intel_context *ce,
	       const struct i915_perf_stream *stream)
2140{
u32 ctx_oactxctrl = stream->perf->ctx_oactxctrl_offset;
u32 ctx_flexeu0 = stream->perf->ctx_flexeu0_offset;
/* The MMIO offsets for Flex EU registers aren't contiguous */
static const i915_reg_t flex_regs[] = {
EU_PERF_CNTL0((const i915_reg_t){ .reg = (0xe458) }),
EU_PERF_CNTL1((const i915_reg_t){ .reg = (0xe558) }),
EU_PERF_CNTL2((const i915_reg_t){ .reg = (0xe658) }),
EU_PERF_CNTL3((const i915_reg_t){ .reg = (0xe758) }),
EU_PERF_CNTL4((const i915_reg_t){ .reg = (0xe45c) }),
EU_PERF_CNTL5((const i915_reg_t){ .reg = (0xe55c) }),
EU_PERF_CNTL6((const i915_reg_t){ .reg = (0xe65c) }),
};
u32 *reg_state = ce->lrc_reg_state;
int i;

reg_state[ctx_oactxctrl + 1] =
(stream->period_exponent << GEN8_OA_TIMER_PERIOD_SHIFT2) |
(stream->periodic ? GEN8_OA_TIMER_ENABLE(1 << 1) : 0) |
GEN8_OA_COUNTER_RESUME(1 << 0);

for (i = 0; i < ARRAY_SIZE(flex_regs)(sizeof((flex_regs)) / sizeof((flex_regs)[0])); i++)
reg_state[ctx_flexeu0 + i * 2 + 1] =
	oa_config_flex_reg(stream->oa_config, flex_regs[i]);
2164}

2166struct flex {
i915_reg_t reg;
u32 offset;
u32 value;
2170};

2172static int
2173gen8_store_flex(struct i915_request *rq,
struct intel_context *ce,
const struct flex *flex, unsigned int count)
2176{
u32 offset;
u32 *cs;

cs = intel_ring_begin(rq, 4 * count);
if (IS_ERR(cs))
return PTR_ERR(cs);

offset = i915_ggtt_offset(ce->state) + LRC_STATE_OFFSET(((0) + (1)) * (1 << 12));
do {
*cs++ = MI_STORE_DWORD_IMM_GEN4(((0x0) << 29) | (0x20) << 23 | (2)) | MI_USE_GGTT(1 << 22);
*cs++ = offset + flex->offset * sizeof(u32);
*cs++ = 0;
*cs++ = flex->value;
} while (flex++, --count);

intel_ring_advance(rq, cs);

return 0;
2195}

2197static int
2198gen8_load_flex(struct i915_request *rq,
      struct intel_context *ce,
      const struct flex *flex, unsigned int count)
2201{
u32 *cs;

GEM_BUG_ON(!count || count > 63)((void)0);

cs = intel_ring_begin(rq, 2 * count + 2);
if (IS_ERR(cs))
return PTR_ERR(cs);

*cs++ = MI_LOAD_REGISTER_IMM(count)(((0x0) << 29) | (0x22) << 23 | (2*(count)-1));
do {
*cs++ = i915_mmio_reg_offset(flex->reg);
*cs++ = flex->value;
} while (flex++, --count);
*cs++ = MI_NOOP(((0x0) << 29) | (0) << 23 | (0));

intel_ring_advance(rq, cs);

return 0;
2220}

2222static int gen8_modify_context(struct intel_context *ce,
	       const struct flex *flex, unsigned int count)
2224{
struct i915_request *rq;
int err;

rq = intel_engine_create_kernel_request(ce->engine);
if (IS_ERR(rq))
return PTR_ERR(rq);

/* Serialise with the remote context */
err = intel_context_prepare_remote_request(ce, rq);
if (err == 0)
err = gen8_store_flex(rq, ce, flex, count);

i915_request_add(rq);
return err;
2239}

2241static int
2242gen8_modify_self(struct intel_context *ce,
 const struct flex *flex, unsigned int count,
 struct i915_active *active)
2245{
struct i915_request *rq;
int err;

intel_engine_pm_get(ce->engine);
rq = i915_request_create(ce);
intel_engine_pm_put(ce->engine);
if (IS_ERR(rq))
return PTR_ERR(rq);

if (!IS_ERR_OR_NULL(active)) {
err = i915_active_add_request(active, rq);
if (err)
	goto err_add_request;
}

err = gen8_load_flex(rq, ce, flex, count);
if (err)
goto err_add_request;

2265err_add_request:
i915_request_add(rq);
return err;
2268}

2270static int gen8_configure_context(struct i915_gem_context *ctx,
		  struct flex *flex, unsigned int count)
2272{
struct i915_gem_engines_iter it;
struct intel_context *ce;
int err = 0;

for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it)for (i915_gem_engines_iter_init(&(it), (i915_gem_context_lock_engines
(ctx))); ((ce) = i915_gem_engines_iter_next(&(it)));) {
GEM_BUG_ON(ce == ce->engine->kernel_context)((void)0);

if (ce->engine->class != RENDER_CLASS0)
	continue;

/* Otherwise OA settings will be set upon first use */
if (!intel_context_pin_if_active(ce))
	continue;

flex->value = intel_sseu_make_rpcs(ce->engine->gt, &ce->sseu);
err = gen8_modify_context(ce, flex, count);

intel_context_unpin(ce);
if (err)
	break;
}
i915_gem_context_unlock_engines(ctx);

return err;
2297}

2299static int gen12_configure_oar_context(struct i915_perf_stream *stream,
		       struct i915_active *active)
2301{
int err;
struct intel_context *ce = stream->pinned_ctx;
u32 format = stream->oa_buffer.format;
struct flex regs_context[] = {
{
	GEN8_OACTXCONTROL((const i915_reg_t){ .reg = (0x2360) }),
	stream->perf->ctx_oactxctrl_offset + 1,
	active ? GEN8_OA_COUNTER_RESUME(1 << 0) : 0,
},
};
/* Offsets in regs_lri are not used since this configuration is only
* applied using LRI. Initialize the correct offsets for posterity.
*/
2315#define GEN12_OAR_OACONTROL_OFFSET0x5B0 0x5B0
struct flex regs_lri[] = {
{
	GEN12_OAR_OACONTROL((const i915_reg_t){ .reg = (0x2960) }),
	GEN12_OAR_OACONTROL_OFFSET0x5B0 + 1,
	(format << GEN12_OAR_OACONTROL_COUNTER_FORMAT_SHIFT1) |
	(active ? GEN12_OAR_OACONTROL_COUNTER_ENABLE(1 << 0) : 0)
},
{
	RING_CONTEXT_CONTROL(ce->engine->mmio_base)((const i915_reg_t){ .reg = ((ce->engine->mmio_base) + 0x244
) }),
	CTX_CONTEXT_CONTROL(0x02 + 1),
	_MASKED_FIELD(GEN12_CTX_CTRL_OAR_CONTEXT_ENABLE,({ if (__builtin_constant_p(((u32)((1UL << (8)) + 0))))
 do { } while (0); if (__builtin_constant_p(active ? ((u32)((
1UL << (8)) + 0)) : 0)) do { } while (0); if (__builtin_constant_p
(((u32)((1UL << (8)) + 0))) && __builtin_constant_p
(active ? ((u32)((1UL << (8)) + 0)) : 0)) do { } while (
0); ((((u32)((1UL << (8)) + 0))) << 16 | (active ?
 ((u32)((1UL << (8)) + 0)) : 0)); })
		      active ?({ if (__builtin_constant_p(((u32)((1UL << (8)) + 0))))
 do { } while (0); if (__builtin_constant_p(active ? ((u32)((
1UL << (8)) + 0)) : 0)) do { } while (0); if (__builtin_constant_p
(((u32)((1UL << (8)) + 0))) && __builtin_constant_p
(active ? ((u32)((1UL << (8)) + 0)) : 0)) do { } while (
0); ((((u32)((1UL << (8)) + 0))) << 16 | (active ?
 ((u32)((1UL << (8)) + 0)) : 0)); })
		      GEN12_CTX_CTRL_OAR_CONTEXT_ENABLE :({ if (__builtin_constant_p(((u32)((1UL << (8)) + 0))))
 do { } while (0); if (__builtin_constant_p(active ? ((u32)((
1UL << (8)) + 0)) : 0)) do { } while (0); if (__builtin_constant_p
(((u32)((1UL << (8)) + 0))) && __builtin_constant_p
(active ? ((u32)((1UL << (8)) + 0)) : 0)) do { } while (
0); ((((u32)((1UL << (8)) + 0))) << 16 | (active ?
 ((u32)((1UL << (8)) + 0)) : 0)); })
		      0)({ if (__builtin_constant_p(((u32)((1UL << (8)) + 0))))
 do { } while (0); if (__builtin_constant_p(active ? ((u32)((
1UL << (8)) + 0)) : 0)) do { } while (0); if (__builtin_constant_p
(((u32)((1UL << (8)) + 0))) && __builtin_constant_p
(active ? ((u32)((1UL << (8)) + 0)) : 0)) do { } while (
0); ((((u32)((1UL << (8)) + 0))) << 16 | (active ?
 ((u32)((1UL << (8)) + 0)) : 0)); })
},
};

/* Modify the context image of pinned context with regs_context*/
err = intel_context_lock_pinned(ce);
if (err)
return err;

err = gen8_modify_context(ce, regs_context, ARRAY_SIZE(regs_context)(sizeof((regs_context)) / sizeof((regs_context)[0])));
intel_context_unlock_pinned(ce);
if (err)
return err;

/* Apply regs_lri using LRI with pinned context */
return gen8_modify_self(ce, regs_lri, ARRAY_SIZE(regs_lri)(sizeof((regs_lri)) / sizeof((regs_lri)[0])), active);
2345}

2347/*
* Manages updating the per-context aspects of the OA stream
* configuration across all contexts.
*
* The awkward consideration here is that OACTXCONTROL controls the
* exponent for periodic sampling which is primarily used for system
* wide profiling where we'd like a consistent sampling period even in
* the face of context switches.
*
* Our approach of updating the register state context (as opposed to
* say using a workaround batch buffer) ensures that the hardware
* won't automatically reload an out-of-date timer exponent even
* transiently before a WA BB could be parsed.
*
* This function needs to:
* - Ensure the currently running context's per-context OA state is
*   updated
* - Ensure that all existing contexts will have the correct per-context
*   OA state if they are scheduled for use.
* - Ensure any new contexts will be initialized with the correct
*   per-context OA state.
*
* Note: it's only the RCS/Render context that has any OA state.
* Note: the first flex register passed must always be R_PWR_CLK_STATE
*/
2372static int
2373oa_configure_all_contexts(struct i915_perf_stream *stream,
	  struct flex *regs,
	  size_t num_regs,
	  struct i915_active *active)
2377{
struct drm_i915_privateinteldrm_softc *i915 = stream->perf->i915;
struct intel_engine_cs *engine;
struct i915_gem_context *ctx, *cn;
int err;

lockdep_assert_held(&stream->perf->lock)do { (void)(&stream->perf->lock); } while(0);

/*
* The OA register config is setup through the context image. This image
* might be written to by the GPU on context switch (in particular on
* lite-restore). This means we can't safely update a context's image,
* if this context is scheduled/submitted to run on the GPU.
*
* We could emit the OA register config through the batch buffer but
* this might leave small interval of time where the OA unit is
* configured at an invalid sampling period.
*
* Note that since we emit all requests from a single ring, there
* is still an implicit global barrier here that may cause a high
* priority context to wait for an otherwise independent low priority
* context. Contexts idle at the time of reconfiguration are not
* trapped behind the barrier.
*/
spin_lock(&i915->gem.contexts.lock)mtx_enter(&i915->gem.contexts.lock);
list_for_each_entry_safe(ctx, cn, &i915->gem.contexts.list, link)for (ctx = ({ const __typeof( ((__typeof(*ctx) *)0)->link )
 *__mptr = ((&i915->gem.contexts.list)->next); (__typeof
(*ctx) *)( (char *)__mptr - __builtin_offsetof(__typeof(*ctx)
, link) );}), cn = ({ const __typeof( ((__typeof(*ctx) *)0)->
link ) *__mptr = (ctx->link.next); (__typeof(*ctx) *)( (char
 *)__mptr - __builtin_offsetof(__typeof(*ctx), link) );}); &
ctx->link != (&i915->gem.contexts.list); ctx = cn, cn
 = ({ const __typeof( ((__typeof(*cn) *)0)->link ) *__mptr
 = (cn->link.next); (__typeof(*cn) *)( (char *)__mptr - __builtin_offsetof
(__typeof(*cn), link) );})) {
if (!kref_get_unless_zero(&ctx->ref))
	continue;

spin_unlock(&i915->gem.contexts.lock)mtx_leave(&i915->gem.contexts.lock);

err = gen8_configure_context(ctx, regs, num_regs);
if (err) {
	i915_gem_context_put(ctx);
	return err;
}

spin_lock(&i915->gem.contexts.lock)mtx_enter(&i915->gem.contexts.lock);
list_safe_reset_next(ctx, cn, link)cn = ({ const __typeof( ((typeof(*(ctx)) *)0)->link ) *__mptr
 = (((ctx)->link.next)); (typeof(*(ctx)) *)( (char *)__mptr
 - __builtin_offsetof(typeof(*(ctx)), link) );});
i915_gem_context_put(ctx);
}
spin_unlock(&i915->gem.contexts.lock)mtx_leave(&i915->gem.contexts.lock);

/*
* After updating all other contexts, we need to modify ourselves.
* If we don't modify the kernel_context, we do not get events while
* idle.
*/
for_each_uabi_engine(engine, i915)for ((engine) = (linux_root_RB_MINMAX((struct linux_root *)(&
(i915)->uabi_engines), -1) ? ({ const __typeof( ((struct intel_engine_cs
 *)0)->uabi_node ) *__mptr = (linux_root_RB_MINMAX((struct
 linux_root *)(&(i915)->uabi_engines), -1)); (struct intel_engine_cs
 *)( (char *)__mptr - __builtin_offsetof(struct intel_engine_cs
, uabi_node) );}) : ((void *)0)); (engine); (engine) = (linux_root_RB_NEXT
((&(engine)->uabi_node)) ? ({ const __typeof( ((struct
 intel_engine_cs *)0)->uabi_node ) *__mptr = (linux_root_RB_NEXT
((&(engine)->uabi_node))); (struct intel_engine_cs *)(
 (char *)__mptr - __builtin_offsetof(struct intel_engine_cs, uabi_node
) );}) : ((void *)0))) {
struct intel_context *ce = engine->kernel_context;

if (engine->class != RENDER_CLASS0)
	continue;

regs[0].value = intel_sseu_make_rpcs(engine->gt, &ce->sseu);

err = gen8_modify_self(ce, regs, num_regs, active);
if (err)
	return err;
}

return 0;
2439}

2441static int
2442gen12_configure_all_contexts(struct i915_perf_stream *stream,
	     const struct i915_oa_config *oa_config,
	     struct i915_active *active)
2445{
struct flex regs[] = {
{
	GEN8_R_PWR_CLK_STATE(RENDER_RING_BASE)((const i915_reg_t){ .reg = ((0x02000) + 0xc8) }),
	CTX_R_PWR_CLK_STATE(0x42 + 1),
},
};

return oa_configure_all_contexts(stream,
			 regs, ARRAY_SIZE(regs)(sizeof((regs)) / sizeof((regs)[0])),
			 active);
2456}

2458static int
2459lrc_configure_all_contexts(struct i915_perf_stream *stream,
	   const struct i915_oa_config *oa_config,
	   struct i915_active *active)
2462{
/* The MMIO offsets for Flex EU registers aren't contiguous */
const u32 ctx_flexeu0 = stream->perf->ctx_flexeu0_offset;
2465#define ctx_flexeuN(N) (ctx_flexeu0 + 2 * (N) + 1)
struct flex regs[] = {
{
	GEN8_R_PWR_CLK_STATE(RENDER_RING_BASE)((const i915_reg_t){ .reg = ((0x02000) + 0xc8) }),
	CTX_R_PWR_CLK_STATE(0x42 + 1),
},
{
	GEN8_OACTXCONTROL((const i915_reg_t){ .reg = (0x2360) }),
	stream->perf->ctx_oactxctrl_offset + 1,
},
{ EU_PERF_CNTL0((const i915_reg_t){ .reg = (0xe458) }), ctx_flexeuN(0) },
{ EU_PERF_CNTL1((const i915_reg_t){ .reg = (0xe558) }), ctx_flexeuN(1) },
{ EU_PERF_CNTL2((const i915_reg_t){ .reg = (0xe658) }), ctx_flexeuN(2) },
{ EU_PERF_CNTL3((const i915_reg_t){ .reg = (0xe758) }), ctx_flexeuN(3) },
{ EU_PERF_CNTL4((const i915_reg_t){ .reg = (0xe45c) }), ctx_flexeuN(4) },
{ EU_PERF_CNTL5((const i915_reg_t){ .reg = (0xe55c) }), ctx_flexeuN(5) },
{ EU_PERF_CNTL6((const i915_reg_t){ .reg = (0xe65c) }), ctx_flexeuN(6) },
};
2483#undef ctx_flexeuN
int i;

regs[1].value =
(stream->period_exponent << GEN8_OA_TIMER_PERIOD_SHIFT2) |
(stream->periodic ? GEN8_OA_TIMER_ENABLE(1 << 1) : 0) |
GEN8_OA_COUNTER_RESUME(1 << 0);

for (i = 2; i < ARRAY_SIZE(regs)(sizeof((regs)) / sizeof((regs)[0])); i++)
regs[i].value = oa_config_flex_reg(oa_config, regs[i].reg);

return oa_configure_all_contexts(stream,
			 regs, ARRAY_SIZE(regs)(sizeof((regs)) / sizeof((regs)[0])),
			 active);
2497}

2499static int
2500gen8_enable_metric_set(struct i915_perf_stream *stream,
       struct i915_active *active)
2502{
struct intel_uncore *uncore = stream->uncore;
struct i915_oa_config *oa_config = stream->oa_config;
int ret;

/*
* We disable slice/unslice clock ratio change reports on SKL since
* they are too noisy. The HW generates a lot of redundant reports
* where the ratio hasn't really changed causing a lot of redundant
* work to processes and increasing the chances we'll hit buffer
* overruns.
*
* Although we don't currently use the 'disable overrun' OABUFFER
* feature it's worth noting that clock ratio reports have to be
* disabled before considering to use that feature since the HW doesn't
* correctly block these reports.
*
* Currently none of the high-level metrics we have depend on knowing
* this ratio to normalize.
*
* Note: This register is not power context saved and restored, but
* that's OK considering that we disable RC6 while the OA unit is
* enabled.
*
* The _INCLUDE_CLK_RATIO bit allows the slice/unslice frequency to
* be read back from automatically triggered reports, as part of the
* RPT_ID field.
*/
if (IS_GRAPHICS_VER(stream->perf->i915, 9, 11)(((&(stream->perf->i915)->__runtime)->graphics
.ip.ver) >= (9) && ((&(stream->perf->i915
)->__runtime)->graphics.ip.ver) <= (11))) {
intel_uncore_write(uncore, GEN8_OA_DEBUG((const i915_reg_t){ .reg = (0x2B04) }),
		   _MASKED_BIT_ENABLE(GEN9_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS |({ typeof((1 << 5) | (1 << 6)) _a = ((1 << 5
) | (1 << 6)); ({ if (__builtin_constant_p(_a)) do { } while
 (0); if (__builtin_constant_p(_a)) do { } while (0); if (__builtin_constant_p
(_a) && __builtin_constant_p(_a)) do { } while (0); (
(_a) << 16 | (_a)); }); })
				      GEN9_OA_DEBUG_INCLUDE_CLK_RATIO)({ typeof((1 << 5) | (1 << 6)) _a = ((1 << 5
) | (1 << 6)); ({ if (__builtin_constant_p(_a)) do { } while
 (0); if (__builtin_constant_p(_a)) do { } while (0); if (__builtin_constant_p
(_a) && __builtin_constant_p(_a)) do { } while (0); (
(_a) << 16 | (_a)); }); }));
}

/*
* Update all contexts prior writing the mux configurations as we need
* to make sure all slices/subslices are ON before writing to NOA
* registers.
*/
ret = lrc_configure_all_contexts(stream, oa_config, active);
if (ret)
return ret;

return emit_oa_config(stream,
	      stream->oa_config, oa_context(stream),
	      active);
2548}

2550static u32 oag_report_ctx_switches(const struct i915_perf_stream *stream)
2551{
return _MASKED_FIELD(GEN12_OAG_OA_DEBUG_DISABLE_CTX_SWITCH_REPORTS,({ if (__builtin_constant_p((1 << 1))) do { } while (0)
; if (__builtin_constant_p((stream->sample_flags & (1<<
0)) ? 0 : (1 << 1))) do { } while (0); if (__builtin_constant_p
((1 << 1)) && __builtin_constant_p((stream->
sample_flags & (1<<0)) ? 0 : (1 << 1))) do { }
 while (0); (((1 << 1)) << 16 | ((stream->sample_flags
 & (1<<0)) ? 0 : (1 << 1))); })
	     (stream->sample_flags & SAMPLE_OA_REPORT) ?({ if (__builtin_constant_p((1 << 1))) do { } while (0)
; if (__builtin_constant_p((stream->sample_flags & (1<<
0)) ? 0 : (1 << 1))) do { } while (0); if (__builtin_constant_p
((1 << 1)) && __builtin_constant_p((stream->
sample_flags & (1<<0)) ? 0 : (1 << 1))) do { }
 while (0); (((1 << 1)) << 16 | ((stream->sample_flags
 & (1<<0)) ? 0 : (1 << 1))); })
	     0 : GEN12_OAG_OA_DEBUG_DISABLE_CTX_SWITCH_REPORTS)({ if (__builtin_constant_p((1 << 1))) do { } while (0)
; if (__builtin_constant_p((stream->sample_flags & (1<<
0)) ? 0 : (1 << 1))) do { } while (0); if (__builtin_constant_p
((1 << 1)) && __builtin_constant_p((stream->
sample_flags & (1<<0)) ? 0 : (1 << 1))) do { }
 while (0); (((1 << 1)) << 16 | ((stream->sample_flags
 & (1<<0)) ? 0 : (1 << 1))); });
2555}

2557static int
2558gen12_enable_metric_set(struct i915_perf_stream *stream,
	struct i915_active *active)
2560{
struct intel_uncore *uncore = stream->uncore;
struct i915_oa_config *oa_config = stream->oa_config;
bool_Bool periodic = stream->periodic;
u32 period_exponent = stream->period_exponent;
int ret;

intel_uncore_write(uncore, GEN12_OAG_OA_DEBUG((const i915_reg_t){ .reg = (0xdaf8) }),
	   /* Disable clk ratio reports, like previous Gens. */
	   _MASKED_BIT_ENABLE(GEN12_OAG_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS |({ typeof((1 << 5) | (1 << 6)) _a = ((1 << 5
) | (1 << 6)); ({ if (__builtin_constant_p(_a)) do { } while
 (0); if (__builtin_constant_p(_a)) do { } while (0); if (__builtin_constant_p
(_a) && __builtin_constant_p(_a)) do { } while (0); (
(_a) << 16 | (_a)); }); })
1
Taking false branch→
2
←
Taking false branch→
			      GEN12_OAG_OA_DEBUG_INCLUDE_CLK_RATIO)({ typeof((1 << 5) | (1 << 6)) _a = ((1 << 5
) | (1 << 6)); ({ if (__builtin_constant_p(_a)) do { } while
 (0); if (__builtin_constant_p(_a)) do { } while (0); if (__builtin_constant_p
(_a) && __builtin_constant_p(_a)) do { } while (0); (
(_a) << 16 | (_a)); }); }) |
	   /*
	    * If the user didn't require OA reports, instruct
	    * the hardware not to emit ctx switch reports.
	    */
	   oag_report_ctx_switches(stream));

intel_uncore_write(uncore, GEN12_OAG_OAGLBCTXCTRL((const i915_reg_t){ .reg = (0x2b28) }), periodic ?
3
←
Assuming 'periodic' is false→
4
←
'?' condition is false→
	   (GEN12_OAG_OAGLBCTXCTRL_COUNTER_RESUME(1 << 0) |
	    GEN12_OAG_OAGLBCTXCTRL_TIMER_ENABLE(1 << 1) |
	    (period_exponent << GEN12_OAG_OAGLBCTXCTRL_TIMER_PERIOD_SHIFT2))
	    : 0);

/*
* Update all contexts prior writing the mux configurations as we need
* to make sure all slices/subslices are ON before writing to NOA
* registers.
*/
ret = gen12_configure_all_contexts(stream, oa_config, active);
if (ret)
5
←
Assuming 'ret' is 0→
6
←
Taking false branch→
return ret;

/*
* For Gen12, performance counters are context
* saved/restored. Only enable it for the context that
* requested this.
*/
if (stream->ctx) {
7
←
Assuming field 'ctx' is null→
8
←
Taking false branch→
ret = gen12_configure_oar_context(stream, active);
if (ret)
	return ret;
}

return emit_oa_config(stream,
9
←
Calling 'emit_oa_config'→
	      stream->oa_config, oa_context(stream),
	      active);
2606}

2608static void gen8_disable_metric_set(struct i915_perf_stream *stream)
2609{
struct intel_uncore *uncore = stream->uncore;

/* Reset all contexts' slices/subslices configurations. */
lrc_configure_all_contexts(stream, NULL((void *)0), NULL((void *)0));

intel_uncore_rmw(uncore, GDT_CHICKEN_BITS((const i915_reg_t){ .reg = (0x9840) }), GT_NOA_ENABLE0x00000080, 0);
2616}

2618static void gen11_disable_metric_set(struct i915_perf_stream *stream)
2619{
struct intel_uncore *uncore = stream->uncore;

/* Reset all contexts' slices/subslices configurations. */
lrc_configure_all_contexts(stream, NULL((void *)0), NULL((void *)0));

/* Make sure we disable noa to save power. */
intel_uncore_rmw(uncore, RPM_CONFIG1((const i915_reg_t){ .reg = (0xd04) }), GEN10_GT_NOA_ENABLE(1 << 9), 0);
2627}

2629static void gen12_disable_metric_set(struct i915_perf_stream *stream)
2630{
struct intel_uncore *uncore = stream->uncore;

/* Reset all contexts' slices/subslices configurations. */
gen12_configure_all_contexts(stream, NULL((void *)0), NULL((void *)0));

/* disable the context save/restore or OAR counters */
if (stream->ctx)
gen12_configure_oar_context(stream, NULL((void *)0));

/* Make sure we disable noa to save power. */
intel_uncore_rmw(uncore, RPM_CONFIG1((const i915_reg_t){ .reg = (0xd04) }), GEN10_GT_NOA_ENABLE(1 << 9), 0);
2642}

2644static void gen7_oa_enable(struct i915_perf_stream *stream)
2645{
struct intel_uncore *uncore = stream->uncore;
struct i915_gem_context *ctx = stream->ctx;
u32 ctx_id = stream->specific_ctx_id;
bool_Bool periodic = stream->periodic;
u32 period_exponent = stream->period_exponent;
u32 report_format = stream->oa_buffer.format;

/*
* Reset buf pointers so we don't forward reports from before now.
*
* Think carefully if considering trying to avoid this, since it
* also ensures status flags and the buffer itself are cleared
* in error paths, and we have checks for invalid reports based
* on the assumption that certain fields are written to zeroed
* memory which this helps maintains.
*/
gen7_init_oa_buffer(stream);

intel_uncore_write(uncore, GEN7_OACONTROL((const i915_reg_t){ .reg = (0x2360) }),
	   (ctx_id & GEN7_OACONTROL_CTX_MASK0xFFFFF000) |
	   (period_exponent <<
	    GEN7_OACONTROL_TIMER_PERIOD_SHIFT6) |
	   (periodic ? GEN7_OACONTROL_TIMER_ENABLE(1 << 5) : 0) |
	   (report_format << GEN7_OACONTROL_FORMAT_SHIFT2) |
	   (ctx ? GEN7_OACONTROL_PER_CTX_ENABLE(1 << 1) : 0) |
	   GEN7_OACONTROL_ENABLE(1 << 0));
2672}

2674static void gen8_oa_enable(struct i915_perf_stream *stream)
2675{
struct intel_uncore *uncore = stream->uncore;
u32 report_format = stream->oa_buffer.format;

/*
* Reset buf pointers so we don't forward reports from before now.
*
* Think carefully if considering trying to avoid this, since it
* also ensures status flags and the buffer itself are cleared
* in error paths, and we have checks for invalid reports based
* on the assumption that certain fields are written to zeroed
* memory which this helps maintains.
*/
gen8_init_oa_buffer(stream);

/*
* Note: we don't rely on the hardware to perform single context
* filtering and instead filter on the cpu based on the context-id
* field of reports
*/
intel_uncore_write(uncore, GEN8_OACONTROL((const i915_reg_t){ .reg = (0x2B00) }),
	   (report_format << GEN8_OA_REPORT_FORMAT_SHIFT2) |
	   GEN8_OA_COUNTER_ENABLE(1 << 0));
2698}

2700static void gen12_oa_enable(struct i915_perf_stream *stream)
2701{
struct intel_uncore *uncore = stream->uncore;
u32 report_format = stream->oa_buffer.format;

/*
* If we don't want OA reports from the OA buffer, then we don't even
* need to program the OAG unit.
*/
if (!(stream->sample_flags & SAMPLE_OA_REPORT(1<<0)))
return;

gen12_init_oa_buffer(stream);

intel_uncore_write(uncore, GEN12_OAG_OACONTROL((const i915_reg_t){ .reg = (0xdaf4) }),
	   (report_format << GEN12_OAG_OACONTROL_OA_COUNTER_FORMAT_SHIFT2) |
	   GEN12_OAG_OACONTROL_OA_COUNTER_ENABLE(1 << 0));
2717}

2719#ifdef notyet

2721/**
* i915_oa_stream_enable - handle `I915_PERF_IOCTL_ENABLE` for OA stream
* @stream: An i915 perf stream opened for OA metrics
*
* [Re]enables hardware periodic sampling according to the period configured
* when opening the stream. This also starts a hrtimer that will periodically
* check for data in the circular OA buffer for notifying userspace (e.g.
* during a read() or poll()).
*/
2730static void i915_oa_stream_enable(struct i915_perf_stream *stream)
2731{
stream->pollin = false0;

stream->perf->ops.oa_enable(stream);

if (stream->sample_flags & SAMPLE_OA_REPORT(1<<0))
hrtimer_start(&stream->poll_check_timer,
	      ns_to_ktime(stream->poll_oa_period),
	      HRTIMER_MODE_REL_PINNED);
2740}

2742#endif

2744static void gen7_oa_disable(struct i915_perf_stream *stream)
2745{
struct intel_uncore *uncore = stream->uncore;

intel_uncore_write(uncore, GEN7_OACONTROL((const i915_reg_t){ .reg = (0x2360) }), 0);
if (intel_wait_for_register(uncore,
		    GEN7_OACONTROL((const i915_reg_t){ .reg = (0x2360) }), GEN7_OACONTROL_ENABLE(1 << 0), 0,
		    50))
drm_err(&stream->perf->i915->drm,printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "wait for OA to be disabled timed out\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__)
	"wait for OA to be disabled timed out\n")printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "wait for OA to be disabled timed out\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__);
2754}

2756static void gen8_oa_disable(struct i915_perf_stream *stream)
2757{
struct intel_uncore *uncore = stream->uncore;

intel_uncore_write(uncore, GEN8_OACONTROL((const i915_reg_t){ .reg = (0x2B00) }), 0);
if (intel_wait_for_register(uncore,
		    GEN8_OACONTROL((const i915_reg_t){ .reg = (0x2B00) }), GEN8_OA_COUNTER_ENABLE(1 << 0), 0,
		    50))
drm_err(&stream->perf->i915->drm,printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "wait for OA to be disabled timed out\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__)
	"wait for OA to be disabled timed out\n")printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "wait for OA to be disabled timed out\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__);
2766}

2768static void gen12_oa_disable(struct i915_perf_stream *stream)
2769{
struct intel_uncore *uncore = stream->uncore;

intel_uncore_write(uncore, GEN12_OAG_OACONTROL((const i915_reg_t){ .reg = (0xdaf4) }), 0);
if (intel_wait_for_register(uncore,
		    GEN12_OAG_OACONTROL((const i915_reg_t){ .reg = (0xdaf4) }),
		    GEN12_OAG_OACONTROL_OA_COUNTER_ENABLE(1 << 0), 0,
		    50))
drm_err(&stream->perf->i915->drm,printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "wait for OA to be disabled timed out\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__)
	"wait for OA to be disabled timed out\n")printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "wait for OA to be disabled timed out\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__);

intel_uncore_write(uncore, GEN12_OA_TLB_INV_CR((const i915_reg_t){ .reg = (0xceec) }), 1);
if (intel_wait_for_register(uncore,
		    GEN12_OA_TLB_INV_CR((const i915_reg_t){ .reg = (0xceec) }),
		    1, 0,
		    50))
drm_err(&stream->perf->i915->drm,printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "wait for OA tlb invalidate timed out\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__)
	"wait for OA tlb invalidate timed out\n")printf("drm:pid%d:%s *ERROR* " "[drm] " "*ERROR* " "wait for OA tlb invalidate timed out\n"
, ({struct cpu_info *__ci; asm volatile("movq %%gs:%P1,%0" : "=r"
 (__ci) :"n" (__builtin_offsetof(struct cpu_info, ci_self)));
 __ci;})->ci_curproc->p_p->ps_pid, __func__);
2787}

2789#ifdef notyet

2791/**
* i915_oa_stream_disable - handle `I915_PERF_IOCTL_DISABLE` for OA stream
* @stream: An i915 perf stream opened for OA metrics
*
* Stops the OA unit from periodically writing counter reports into the
* circular OA buffer. This also stops the hrtimer that periodically checks for
* data in the circular OA buffer, for notifying userspace.
*/
2799static void i915_oa_stream_disable(struct i915_perf_stream *stream)
2800{
stream->perf->ops.oa_disable(stream);

if (stream->sample_flags & SAMPLE_OA_REPORT(1<<0))
hrtimer_cancel(&stream->poll_check_timer)timeout_del_barrier(&stream->poll_check_timer);
2805}

2807static const struct i915_perf_stream_ops i915_oa_stream_ops = {
.destroy = i915_oa_stream_destroy,
.enable = i915_oa_stream_enable,
.disable = i915_oa_stream_disable,
.wait_unlocked = i915_oa_wait_unlocked,
.poll_wait = i915_oa_poll_wait,
.read = i915_oa_read,
2814};

2816static int i915_perf_stream_enable_sync(struct i915_perf_stream *stream)
2817{
struct i915_active *active;
int err;

active = i915_active_create();
if (!active)
return -ENOMEM12;

err = stream->perf->ops.enable_metric_set(stream, active);
if (err == 0)
__i915_active_wait(active, TASK_UNINTERRUPTIBLE0);

i915_active_put(active);
return err;
2831}

2833static void
2834get_default_sseu_config(struct intel_sseu *out_sseu,
	struct intel_engine_cs *engine)
2836{
const struct sseu_dev_info *devinfo_sseu = &engine->gt->info.sseu;

*out_sseu = intel_sseu_from_device_info(devinfo_sseu);

if (GRAPHICS_VER(engine->i915)((&(engine->i915)->__runtime)->graphics.ip.ver) == 11) {
/*
 * We only need subslice count so it doesn't matter which ones
 * we select - just turn off low bits in the amount of half of
 * all available subslices per slice.
 */
out_sseu->subslice_mask =
	~(~0 << (hweight8(out_sseu->subslice_mask) / 2));
out_sseu->slice_mask = 0x1;
}
2851}

2853#endif

2855static int
2856get_sseu_config(struct intel_sseu *out_sseu,
struct intel_engine_cs *engine,
const struct drm_i915_gem_context_param_sseu *drm_sseu)
2859{
if (drm_sseu->engine.engine_class != engine->uabi_class ||
   drm_sseu->engine.engine_instance != engine->uabi_instance)
return -EINVAL22;

return i915_gem_user_to_context_sseu(engine->gt, drm_sseu, out_sseu);
2865}

2867#ifdef notyet

2869/**
* i915_oa_stream_init - validate combined props for OA stream and init
* @stream: An i915 perf stream
* @param: The open parameters passed to `DRM_I915_PERF_OPEN`
* @props: The property state that configures stream (individually validated)
*
* While read_properties_unlocked() validates properties in isolation it
* doesn't ensure that the combination necessarily makes sense.
*
* At this point it has been determined that userspace wants a stream of
* OA metrics, but still we need to further validate the combined
* properties are OK.
*
* If the configuration makes sense then we can allocate memory for
* a circular OA buffer and apply the requested metric set configuration.
*
* Returns: zero on success or a negative error code.
*/
2887static int i915_oa_stream_init(struct i915_perf_stream *stream,
	       struct drm_i915_perf_open_param *param,
	       struct perf_open_properties *props)
2890{
struct drm_i915_privateinteldrm_softc *i915 = stream->perf->i915;
struct i915_perf *perf = stream->perf;
int format_size;
int ret;

if (!props->engine) {
drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA engine not specified\n")
	"OA engine not specified\n")__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA engine not specified\n");
return -EINVAL22;
}

/*
* If the sysfs metrics/ directory wasn't registered for some
* reason then don't let userspace try their luck with config
* IDs
*/
if (!perf->metrics_kobj) {
drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA metrics weren't advertised via sysfs\n"
)
	"OA metrics weren't advertised via sysfs\n")__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA metrics weren't advertised via sysfs\n"
);
return -EINVAL22;
}

if (!(props->sample_flags & SAMPLE_OA_REPORT(1<<0)) &&
   (GRAPHICS_VER(perf->i915)((&(perf->i915)->__runtime)->graphics.ip.ver) < 12 || !stream->ctx)) {
drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Only OA report sampling supported\n")
	"Only OA report sampling supported\n")__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Only OA report sampling supported\n");
return -EINVAL22;
}

if (!perf->ops.enable_metric_set) {
drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA unit not supported\n")
	"OA unit not supported\n")__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA unit not supported\n");
return -ENODEV19;
}

/*
* To avoid the complexity of having to accurately filter
* counter reports and marshal to the appropriate client
* we currently only allow exclusive access
*/
if (perf->exclusive_stream) {
drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA unit already in use\n")
	"OA unit already in use\n")__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA unit already in use\n");
return -EBUSY16;
}

if (!props->oa_format) {
drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA report format not specified\n")
	"OA report format not specified\n")__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "OA report format not specified\n");
return -EINVAL22;
}

stream->engine = props->engine;
stream->uncore = stream->engine->gt->uncore;

stream->sample_size = sizeof(struct drm_i915_perf_record_header);

format_size = perf->oa_formats[props->oa_format].size;

stream->sample_flags = props->sample_flags;
stream->sample_size += format_size;

stream->oa_buffer.format_size = format_size;
if (drm_WARN_ON(&i915->drm, stream->oa_buffer.format_size == 0)({ int __ret = !!((stream->oa_buffer.format_size == 0)); if
 (__ret) printf("%s %s: " "%s", dev_driver_string(((&i915
->drm))->dev), "", "drm_WARN_ON(" "stream->oa_buffer.format_size == 0"
 ")"); __builtin_expect(!!(__ret), 0); }))
return -EINVAL22;

stream->hold_preemption = props->hold_preemption;

stream->oa_buffer.format =
perf->oa_formats[props->oa_format].format;

stream->periodic = props->oa_periodic;
if (stream->periodic)
stream->period_exponent = props->oa_period_exponent;

if (stream->ctx) {
ret = oa_get_render_ctx_id(stream);
if (ret) {
	drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Invalid context id to filter with\n")
		"Invalid context id to filter with\n")__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Invalid context id to filter with\n");
	return ret;
}
}

ret = alloc_noa_wait(stream);
if (ret) {
drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Unable to allocate NOA wait batch buffer\n"
)
	"Unable to allocate NOA wait batch buffer\n")__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Unable to allocate NOA wait batch buffer\n"
);
goto err_noa_wait_alloc;
}

stream->oa_config = i915_perf_get_oa_config(perf, props->metrics_set);
if (!stream->oa_config) {
drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Invalid OA config id=%i\n", props->
metrics_set)
	"Invalid OA config id=%i\n", props->metrics_set)__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Invalid OA config id=%i\n", props->
metrics_set);
ret = -EINVAL22;
goto err_config;
}

/* PRM - observability performance counters:
*
*   OACONTROL, performance counter enable, note:
*
*   "When this bit is set, in order to have coherent counts,
*   RC6 power state and trunk clock gating must be disabled.
*   This can be achieved by programming MMIO registers as
*   0xA094=0 and 0xA090[31]=1"
*
*   In our case we are expecting that taking pm + FORCEWAKE
*   references will effectively disable RC6.
*/
intel_engine_pm_get(stream->engine);
intel_uncore_forcewake_get(stream->uncore, FORCEWAKE_ALL);

ret = alloc_oa_buffer(stream);
if (ret)
goto err_oa_buf_alloc;

stream->ops = &i915_oa_stream_ops;

perf->sseu = props->sseu;
WRITE_ONCE(perf->exclusive_stream, stream)({ typeof(perf->exclusive_stream) __tmp = (stream); *(volatile
 typeof(perf->exclusive_stream) *)&(perf->exclusive_stream
) = __tmp; __tmp; });

ret = i915_perf_stream_enable_sync(stream);
if (ret) {
drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Unable to enable metric set\n")
	"Unable to enable metric set\n")__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "Unable to enable metric set\n");
goto err_enable;
}

drm_dbg(&stream->perf->i915->drm,__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "opening stream oa config uuid=%s\n", stream
->oa_config->uuid)
"opening stream oa config uuid=%s\n",__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "opening stream oa config uuid=%s\n", stream
->oa_config->uuid)
  stream->oa_config->uuid)__drm_dev_dbg(((void *)0), (&stream->perf->i915->
drm) ? (&stream->perf->i915->drm)->dev : ((void
 *)0), DRM_UT_DRIVER, "opening stream oa config uuid=%s\n", stream
->oa_config->uuid);

hrtimer_init(&stream->poll_check_timer,
     CLOCK_MONOTONIC3, HRTIMER_MODE_REL1);
stream->poll_check_timer.function = oa_poll_check_timer_cb;
init_waitqueue_head(&stream->poll_wq);
mtx_init(&stream->oa_buffer.ptr_lock, IPL_TTY)do { (void)(((void *)0)); (void)(0); __mtx_init((&stream->
oa_buffer.ptr_lock), ((((0x9)) > 0x0 && ((0x9)) <
 0x9) ? 0x9 : ((0x9)))); } while (0);

return 0;

3033err_enable:
WRITE_ONCE(perf->exclusive_stream, NULL)({ typeof(perf->exclusive_stream) __tmp = (((void *)0)); *
(volatile typeof(perf->exclusive_stream) *)&(perf->
exclusive_stream) = __tmp; __tmp; });
perf->ops.disable_metric_set(stream);

free_oa_buffer(stream);

3039err_oa_buf_alloc:
free_oa_configs(stream);

intel_uncore_forcewake_put(stream->uncore, FORCEWAKE_ALL);
intel_engine_pm_put(stream->engine);

3045err_config:
free_noa_wait(stream);

3048err_noa_wait_alloc:
if (stream->ctx)
oa_put_render_ctx_id(stream);

return ret;
3053}

3055#endif

3057void i915_oa_init_reg_state(const struct intel_context *ce,
	    const struct intel_engine_cs *engine)
3059{
struct i915_perf_stream *stream;

if (engine->class != RENDER_CLASS0)
return;

/* perf.exclusive_stream serialised by lrc_configure_all_contexts() */
stream = READ_ONCE(engine->i915->perf.exclusive_stream)({ typeof(engine->i915->perf.exclusive_stream) __tmp = *
(volatile typeof(engine->i915->perf.exclusive_stream) *
)&(engine->i915->perf.exclusive_stream); membar_datadep_consumer
(); __tmp; });
if (stream && GRAPHICS_VER(stream->perf->i915)((&(stream->perf->i915)->__runtime)->graphics
.ip.ver) < 12)
gen8_update_reg_state_unlocked(ce, stream);
3069}

3071#ifdef notyet

3073/**
* i915_perf_read - handles read() FOP for i915 perf stream FDs
* @file: An i915 perf stream file
* @buf: destination buffer given by userspace
* @count: the number of bytes userspace wants to read
* @ppos: (inout) file seek position (unused)
*
* The entry point for handling a read() on a stream file descriptor from
* userspace. Most of the work is left to the i915_perf_read_locked() and
* &i915_perf_stream_ops->read but to save having stream implementations (of
* which we might have multiple later) we handle blocking read here.
*
* We can also consistently treat trying to read from a disabled stream
* as an IO error so implementations can assume the stream is enabled
* while reading.
*
* Returns: The number of bytes copied or a negative error code on failure.
*/
3091static ssize_t i915_perf_read(struct file *file,
	      char __user *buf,
	      size_t count,
	      loff_t *ppos)
3095{
struct i915_perf_stream *stream = file->private_data;
struct i915_perf *perf = stream->perf;
size_t offset = 0;
int ret;

/* To ensure it's handled consistently we simply treat all reads of a
* disabled stream as an error. In particular it might otherwise lead
* to a deadlock for blocking file descriptors...
*/
if (!stream->enabled || !(stream->sample_flags & SAMPLE_OA_REPORT(1<<0)))
return -EIO5;

if (!(file->f_flags & O_NONBLOCK0x0004)) {
/* There's the small chance of false positives from
 * stream->ops->wait_unlocked.
 *
 * E.g. with single context filtering since we only wait until
 * oabuffer has >= 1 report we don't immediately know whether
 * any reports really belong to the current context
 */
do {
	ret = stream->ops->wait_unlocked(stream);
	if (ret)
		return ret;

	mutex_lock(&perf->lock)rw_enter_write(&perf->lock);
	ret = stream->ops->read(stream, buf, count, &offset);
	mutex_unlock(&perf->lock)rw_exit_write(&perf->lock);
} while (!offset && !ret);
} else {
mutex_lock(&perf->lock)rw_enter_write(&perf->lock);
ret = stream->ops->read(stream, buf, count, &offset);
mutex_unlock(&perf->lock)rw_exit_write(&perf->lock);
}

/* We allow the poll checking to sometimes report false positive EPOLLIN
* events where we might actually report EAGAIN on read() if there's
* not really any data available. In this situation though we don't
* want to enter a busy loop between poll() reporting a EPOLLIN event
* and read() returning -EAGAIN. Clearing the oa.pollin state here
* effectively ensures we back off until the next hrtimer callback
* before reporting another EPOLLIN event.
* The exception to this is if ops->read() returned -ENOSPC which means
* that more OA data is available than could fit in the user provided
* buffer. In this case we want the next poll() call to not block.
*/
if (ret != -ENOSPC28)
stream->pollin = false0;

/* Possible values for ret are 0, -EFAULT, -ENOSPC, -EIO, ... */
return offset ?: (ret ?: -EAGAIN35);
3147}

3149static enum hrtimer_restart oa_poll_check_timer_cb(struct hrtimer *hrtimer)
3150{
struct i915_perf_stream *stream =
container_of(hrtimer, typeof(*stream), poll_check_timer)({ const __typeof( ((typeof(*stream) *)0)->poll_check_timer
 ) *__mptr = (hrtimer); (typeof(*stream) *)( (char *)__mptr -
 __builtin_offsetof(typeof(*stream), poll_check_timer) );});

if (oa_buffer_check_unlocked(stream)) {
stream->pollin = true1;
wake_up(&stream->poll_wq);
}

hrtimer_forward_now(hrtimer,
	    ns_to_ktime(stream->poll_oa_period));

return HRTIMER_RESTART;
3163}

3165/**
* i915_perf_poll_locked - poll_wait() with a suitable wait queue for stream
* @stream: An i915 perf stream
* @file: An i915 perf stream file
* @wait: poll() state table
*
* For handling userspace polling on an i915 perf stream, this calls through to
* &i915_perf_stream_ops->poll_wait to call poll_wait() with a wait queue that
* will be woken for new stream data.
*
* Note: The &perf->lock mutex has been taken to serialize
* with any non-file-operation driver hooks.
*
* Returns: any poll events that are ready without sleeping
*/
3180static __poll_t i915_perf_poll_locked(struct i915_perf_stream *stream,
		      struct file *file,
		      poll_table *wait)
3183{
__poll_t events = 0;

stream->ops->poll_wait(stream, file, wait);

/* Note: we don't explicitly check whether there's something to read
* here since this path may be very hot depending on what else
* userspace is polling, or on the timeout in use. We rely solely on
* the hrtimer/oa_poll_check_timer_cb to notify us when there are
* samples to read.
*/
if (stream->pollin)
events |= EPOLLIN;

return events;
3198}

3200/**
* i915_perf_poll - call poll_wait() with a suitable wait queue for stream
* @file: An i915 perf stream file
* @wait: poll() state table
*
* For handling userspace polling on an i915 perf stream, this ensures
* poll_wait() gets called with a wait queue that will be woken for new stream
* data.
*
* Note: Implementation deferred to i915_perf_poll_locked()
*
* Returns: any poll events that are ready without sleeping
*/
3213static __poll_t i915_perf_poll(struct file *file, poll_table *wait)
3214{
struct i915_perf_stream *stream = file->private_data;
struct i915_perf *perf = stream->perf;
__poll_t ret;

mutex_lock(&perf->lock)rw_enter_write(&perf->lock);
ret = i915_perf_poll_locked(stream, file, wait);
mutex_unlock(&perf->lock)rw_exit_write(&perf->lock);

return ret;
3224}

3226/**
* i915_perf_enable_locked - handle `I915_PERF_IOCTL_ENABLE` ioctl
* @stream: A disabled i915 perf stream
*
* [Re]enables the associated capture of data for this stream.
*
* If a stream was previously enabled then there's currently no intention
* to provide userspace any guarantee about the preservation of previously
* buffered data.
*/
3236static void i915_perf_enable_locked(struct i915_perf_stream *stream)
3237{
if (stream->enabled)
return;

/* Allow stream->ops->enable() to refer to this */
stream->enabled = true1;

if (stream->ops->enable)
stream->ops->enable(stream);

if (stream->hold_preemption)
intel_context_set_nopreempt(stream->pinned_ctx);
3249}

3251/**
* i915_perf_disable_locked - handle `I915_PERF_IOCTL_DISABLE` ioctl
* @stream: An enabled i915 perf stream
*
* Disables the associated capture of data for this stream.
*
* The intention is that disabling an re-enabling a stream will ideally be
* cheaper than destroying and re-opening a stream with the same configuration,
* though there are no formal guarantees about what state or buffered data
* must be retained between disabling and re-enabling a stream.
*
* Note: while a stream is disabled it's considered an error for userspace
* to attempt to read from the stream (-EIO).
*/
3265static void i915_perf_disable_locked(struct i915_perf_stream *stream)
3266{
if (!stream->enabled)
return;

/* Allow stream->ops->disable() to refer to this */
stream->enabled = false0;

if (stream->hold_preemption)
intel_context_clear_nopreempt(stream->pinned_ctx);

if (stream->ops->disable)
stream->ops->disable(stream);
3278}

3280static long i915_perf_config_locked(struct i915_perf_stream *stream,
		    unsigned long metrics_set)
3282{
struct i915_oa_config *config;
long ret = stream->oa_config->id;

config = i915_perf_get_oa_config(stream->perf, metrics_set);
if (!config)
return -EINVAL22;

if (config != stream->oa_config) {
int err;

/*
 * If OA is bound to a specific context, emit the
 * reconfiguration inline from that context. The update
 * will then be ordered with respect to submission on that
 * context.
 *
 * When set globally, we use a low priority kernel context,
 * so it will effectively take effect when idle.
 */
err = emit_oa_config(stream, config, oa_context(stream), NULL((void *)0));
if (!err)
	config = xchg(&stream->oa_config, config)__sync_lock_test_and_set(&stream->oa_config, config);
else
	ret = err;
}

i915_oa_config_put(config);

return ret;
3312}

3314/**
* i915_perf_ioctl_locked - support ioctl() usage with i915 perf stream FDs
* @stream: An i915 perf stream
* @cmd: the ioctl request
* @arg: the ioctl data
*
* Note: The &perf->lock mutex has been taken to serialize
* with any non-file-operation driver hooks.
*
* Returns: zero on success or a negative error code. Returns -EINVAL for
* an unknown ioctl request.
*/
3326static long i915_perf_ioctl_locked(struct i915_perf_stream *stream,
		   unsigned int cmd,
		   unsigned long arg)
3329{
switch (cmd) {
case I915_PERF_IOCTL_ENABLE((unsigned long)0x20000000 | ((0 & 0x1fff) << 16) |
 ((('i')) << 8) | ((0x0))):
i915_perf_enable_locked(stream);
return 0;
case I915_PERF_IOCTL_DISABLE((unsigned long)0x20000000 | ((0 & 0x1fff) << 16) |
 ((('i')) << 8) | ((0x1))):
i915_perf_disable_locked(stream);
return 0;
case I915_PERF_IOCTL_CONFIG((unsigned long)0x20000000 | ((0 & 0x1fff) << 16) |
 ((('i')) << 8) | ((0x2))):
return i915_perf_config_locked(stream, arg);
}

return -EINVAL22;
3342}

3344/**
* i915_perf_ioctl - support ioctl() usage with i915 perf stream FDs
* @file: An i915 perf stream file
* @cmd: the ioctl request
* @arg: the ioctl data
*
* Implementation deferred to i915_perf_ioctl_locked().
*
* Returns: zero on success or a negative error code. Returns -EINVAL for
* an unknown ioctl request.
*/
3355static long i915_perf_ioctl(struct file *file,
	    unsigned int cmd,
	    unsigned long arg)
3358{
struct i915_perf_stream *stream = file->private_data;
struct i915_perf *perf = stream->perf;
long ret;

mutex_lock(&perf->lock)rw_enter_write(&perf->lock);
ret = i915_perf_ioctl_locked(stream, cmd, arg);
mutex_unlock(&perf->lock)rw_exit_write(&perf->lock);

return ret;
3368}

3370/**
* i915_perf_destroy_locked - destroy an i915 perf stream
* @stream: An i915 perf stream
*
* Frees all resources associated with the given i915 perf @stream, disabling
* any associated data capture in the process.
*
* Note: The &perf->lock mutex has been taken to serialize
* with any non-file-operation driver hooks.
*/
3380static void i915_perf_destroy_locked(struct i915_perf_stream *stream)
3381{
if (stream->enabled)
i915_perf_disable_locked(stream);

if (stream->ops->destroy)
stream->ops->destroy(stream);

if (stream->ctx)
i915_gem_context_put(stream->ctx);

kfree(stream);
3392}

3394/**
* i915_perf_release - handles userspace close() of a stream file
* @inode: anonymous inode associated with file
* @file: An i915 perf stream file
*
* Cleans up any resources associated with an open i915 perf stream file.
*
* NB: close() can't really fail from the userspace point of view.
*
* Returns: zero on success or a negative error code.
*/
3405static int i915_perf_release(struct inode *inode, struct file *file)
3406{
struct i915_perf_stream *stream = file->private_data;
struct i915_perf *perf = stream->perf;

mutex_lock(&perf->lock)rw_enter_write(&perf->lock);
i915_perf_destroy_locked(stream);
mutex_unlock(&perf->lock)rw_exit_write(&perf->lock);

/* Release the reference the perf stream kept on the driver. */
drm_dev_put(&perf->i915->drm);

return 0;
3418}


3421static const struct file_operations fops = {
.owner		= THIS_MODULE((void *)0),
.llseek		= no_llseek,
.release	= i915_perf_release,
.poll		= i915_perf_poll,
.read		= i915_perf_read,
.unlocked_ioctl	= i915_perf_ioctl,
/* Our ioctl have no arguments, so it's safe to use the same function
* to handle 32bits compatibility.
*/
.compat_ioctl   = i915_perf_ioctl,
3432};

3434#endif /* notyet */

3436/**
* i915_perf_open_ioctl_locked - DRM ioctl() for userspace to open a stream FD
* @perf: i915 perf instance
* @param: The open parameters passed to 'DRM_I915_PERF_OPEN`
* @props: individually validated u64 property value pairs
* @file: drm file
*
* See i915_perf_ioctl_open() for interface details.
*
* Implements further stream config validation and stream initialization on
* behalf of i915_perf_open_ioctl() with the &perf->lock mutex
* taken to serialize with any non-file-operation driver hooks.
*
* Note: at this point the @props have only been validated in isolation and
* it's still necessary to validate that the combination of properties makes
* sense.
*
* In the case where userspace is interested in OA unit metrics then further
* config validation and stream initialization details will be handled by
* i915_oa_stream_init(). The code here should only validate config state that
* will be relevant to all stream types / backends.
*
* Returns: zero on success or a negative error code.
*/
3460static int
3461i915_perf_open_ioctl_locked(struct i915_perf *perf,
	    struct drm_i915_perf_open_param *param,
	    struct perf_open_properties *props,
	    struct drm_file *file)
3465{
STUB()do { printf("%s: stub\n", __func__); } while(0);
return -ENOSYS78;
3468#ifdef notyet
struct i915_gem_context *specific_ctx = NULL((void *)0);
struct i915_perf_stream *stream = NULL((void *)0);
unsigned long f_flags = 0;
bool_Bool privileged_op = true1;
int stream_fd;
int ret;

if (props->single_context) {
u32 ctx_handle = props->ctx_handle;
struct drm_i915_file_private *file_priv = file->driver_priv;

specific_ctx = i915_gem_context_lookup(file_priv, ctx_handle);
if (IS_ERR(specific_ctx)) {
	drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to look up context with ID %u for opening perf stream\n"
, ctx_handle)
		"Failed to look up context with ID %u for opening perf stream\n",__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to look up context with ID %u for opening perf stream\n"
, ctx_handle)
		  ctx_handle)__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to look up context with ID %u for opening perf stream\n"
, ctx_handle);
	ret = PTR_ERR(specific_ctx);
	goto err;
}
}

/*
* On Haswell the OA unit supports clock gating off for a specific
* context and in this mode there's no visibility of metrics for the
* rest of the system, which we consider acceptable for a
* non-privileged client.
*
* For Gen8->11 the OA unit no longer supports clock gating off for a
* specific context and the kernel can't securely stop the counters
* from updating as system-wide / global values. Even though we can
* filter reports based on the included context ID we can't block
* clients from seeing the raw / global counter values via
* MI_REPORT_PERF_COUNT commands and so consider it a privileged op to
* enable the OA unit by default.
*
* For Gen12+ we gain a new OAR unit that only monitors the RCS on a
* per context basis. So we can relax requirements there if the user
* doesn't request global stream access (i.e. query based sampling
* using MI_RECORD_PERF_COUNT.
*/
if (IS_HASWELL(perf->i915)IS_PLATFORM(perf->i915, INTEL_HASWELL) && specific_ctx)
privileged_op = false0;
else if (GRAPHICS_VER(perf->i915)((&(perf->i915)->__runtime)->graphics.ip.ver) == 12 && specific_ctx &&
 (props->sample_flags & SAMPLE_OA_REPORT(1<<0)) == 0)
privileged_op = false0;

if (props->hold_preemption) {
if (!props->single_context) {
	drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "preemption disable with no context\n"
)
		"preemption disable with no context\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "preemption disable with no context\n"
);
	ret = -EINVAL22;
	goto err;
}
privileged_op = true1;
}

/*
* Asking for SSEU configuration is a priviliged operation.
*/
if (props->has_sseu)
privileged_op = true1;
else
get_default_sseu_config(&props->sseu, props->engine);

/* Similar to perf's kernel.perf_paranoid_cpu sysctl option
* we check a dev.i915.perf_stream_paranoid sysctl option
* to determine if it's ok to access system wide OA counters
* without CAP_PERFMON or CAP_SYS_ADMIN privileges.
*/
if (privileged_op &&
   i915_perf_stream_paranoid && !perfmon_capable()) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Insufficient privileges to open i915 perf stream\n"
)
	"Insufficient privileges to open i915 perf stream\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Insufficient privileges to open i915 perf stream\n"
);
ret = -EACCES13;
goto err_ctx;
}

stream = kzalloc(sizeof(*stream), GFP_KERNEL(0x0001 | 0x0004));
if (!stream) {
ret = -ENOMEM12;
goto err_ctx;
}

stream->perf = perf;
stream->ctx = specific_ctx;
stream->poll_oa_period = props->poll_oa_period;

ret = i915_oa_stream_init(stream, param, props);
if (ret)
goto err_alloc;

/* we avoid simply assigning stream->sample_flags = props->sample_flags
* to have _stream_init check the combination of sample flags more
* thoroughly, but still this is the expected result at this point.
*/
if (WARN_ON(stream->sample_flags != props->sample_flags)({ int __ret = !!(stream->sample_flags != props->sample_flags
); if (__ret) printf("WARNING %s failed at %s:%d\n", "stream->sample_flags != props->sample_flags"
, "/usr/src/sys/dev/pci/drm/i915/i915_perf.c", 3564); __builtin_expect
(!!(__ret), 0); })) {
ret = -ENODEV19;
goto err_flags;
}

if (param->flags & I915_PERF_FLAG_FD_CLOEXEC(1<<0))
f_flags |= O_CLOEXEC0x10000;
if (param->flags & I915_PERF_FLAG_FD_NONBLOCK(1<<1))
f_flags |= O_NONBLOCK0x0004;

stream_fd = anon_inode_getfd("[i915_perf]", &fops, stream, f_flags);
if (stream_fd < 0) {
ret = stream_fd;
goto err_flags;
}

if (!(param->flags & I915_PERF_FLAG_DISABLED(1<<2)))
i915_perf_enable_locked(stream);

/* Take a reference on the driver that will be kept with stream_fd
* until its release.
*/
drm_dev_get(&perf->i915->drm);

return stream_fd;

3590err_flags:
if (stream->ops->destroy)
stream->ops->destroy(stream);
3593err_alloc:
kfree(stream);
3595err_ctx:
if (specific_ctx)
i915_gem_context_put(specific_ctx);
3598err:
return ret;
3600#endif
3601}

3603static u64 oa_exponent_to_ns(struct i915_perf *perf, int exponent)
3604{
return intel_gt_clock_interval_to_ns(to_gt(perf->i915),
			     2ULL << exponent);
3607}

3609static __always_inlineinline __attribute__((__always_inline__)) bool_Bool
3610oa_format_valid(struct i915_perf *perf, enum drm_i915_oa_format format)
3611{
return test_bit(format, perf->format_mask);
3613}

3615static __always_inlineinline __attribute__((__always_inline__)) void
3616oa_format_add(struct i915_perf *perf, enum drm_i915_oa_format format)
3617{
__set_bit(format, perf->format_mask);
3619}

3621/**
* read_properties_unlocked - validate + copy userspace stream open properties
* @perf: i915 perf instance
* @uprops: The array of u64 key value pairs given by userspace
* @n_props: The number of key value pairs expected in @uprops
* @props: The stream configuration built up while validating properties
*
* Note this function only validates properties in isolation it doesn't
* validate that the combination of properties makes sense or that all
* properties necessary for a particular kind of stream have been set.
*
* Note that there currently aren't any ordering requirements for properties so
* we shouldn't validate or assume anything about ordering here. This doesn't
* rule out defining new properties with ordering requirements in the future.
*/
3636static int read_properties_unlocked(struct i915_perf *perf,
		    u64 __user *uprops,
		    u32 n_props,
		    struct perf_open_properties *props)
3640{
u64 __user *uprop = uprops;
u32 i;
int ret;

memset(props, 0, sizeof(struct perf_open_properties))__builtin_memset((props), (0), (sizeof(struct perf_open_properties
)));
props->poll_oa_period = DEFAULT_POLL_PERIOD_NS(1000000000L / 200);

if (!n_props) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "No i915 perf properties given\n"
)
	"No i915 perf properties given\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "No i915 perf properties given\n"
);
return -EINVAL22;
}

/* At the moment we only support using i915-perf on the RCS. */
props->engine = intel_engine_lookup_user(perf->i915,
				 I915_ENGINE_CLASS_RENDER,
				 0);
if (!props->engine) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "No RENDER-capable engines\n"
)
	"No RENDER-capable engines\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "No RENDER-capable engines\n"
);
return -EINVAL22;
}

/* Considering that ID = 0 is reserved and assuming that we don't
* (currently) expect any configurations to ever specify duplicate
* values for a particular property ID then the last _PROP_MAX value is
* one greater than the maximum number of properties we expect to get
* from userspace.
*/
if (n_props >= DRM_I915_PERF_PROP_MAX) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "More i915 perf properties specified than exist\n"
)
	"More i915 perf properties specified than exist\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "More i915 perf properties specified than exist\n"
);
return -EINVAL22;
}

for (i = 0; i < n_props; i++) {
u64 oa_period, oa_freq_hz;
u64 id, value;

ret = get_user(id, uprop)-copyin(uprop, &(id), sizeof(id));
if (ret)
	return ret;

ret = get_user(value, uprop + 1)-copyin(uprop + 1, &(value), sizeof(value));
if (ret)
	return ret;

if (id == 0 || id >= DRM_I915_PERF_PROP_MAX) {
	drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unknown i915 perf property ID\n"
)
		"Unknown i915 perf property ID\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unknown i915 perf property ID\n"
);
	return -EINVAL22;
}

switch ((enum drm_i915_perf_property_id)id) {
case DRM_I915_PERF_PROP_CTX_HANDLE:
	props->single_context = 1;
	props->ctx_handle = value;
	break;
case DRM_I915_PERF_PROP_SAMPLE_OA:
	if (value)
		props->sample_flags |= SAMPLE_OA_REPORT(1<<0);
	break;
case DRM_I915_PERF_PROP_OA_METRICS_SET:
	if (value == 0) {
		drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unknown OA metric set ID\n"
)
			"Unknown OA metric set ID\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unknown OA metric set ID\n"
);
		return -EINVAL22;
	}
	props->metrics_set = value;
	break;
case DRM_I915_PERF_PROP_OA_FORMAT:
	if (value == 0 || value >= I915_OA_FORMAT_MAX) {
		drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Out-of-range OA report format %llu\n"
, value)
			"Out-of-range OA report format %llu\n",__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Out-of-range OA report format %llu\n"
, value)
			  value)__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Out-of-range OA report format %llu\n"
, value);
		return -EINVAL22;
	}
	if (!oa_format_valid(perf, value)) {
		drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unsupported OA report format %llu\n"
, value)
			"Unsupported OA report format %llu\n",__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unsupported OA report format %llu\n"
, value)
			  value)__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unsupported OA report format %llu\n"
, value);
		return -EINVAL22;
	}
	props->oa_format = value;
	break;
case DRM_I915_PERF_PROP_OA_EXPONENT:
	if (value > OA_EXPONENT_MAX31) {
		drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA timer exponent too high (> %u)\n"
, 31)
			"OA timer exponent too high (> %u)\n",__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA timer exponent too high (> %u)\n"
, 31)
			 OA_EXPONENT_MAX)__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA timer exponent too high (> %u)\n"
, 31);
		return -EINVAL22;
	}

	/* Theoretically we can program the OA unit to sample
	 * e.g. every 160ns for HSW, 167ns for BDW/SKL or 104ns
	 * for BXT. We don't allow such high sampling
	 * frequencies by default unless root.
	 */

	BUILD_BUG_ON(sizeof(oa_period) != 8)extern char _ctassert[(!(sizeof(oa_period) != 8)) ? 1 : -1 ] __attribute__
((__unused__));
	oa_period = oa_exponent_to_ns(perf, value);

	/* This check is primarily to ensure that oa_period <=
	 * UINT32_MAX (before passing to do_div which only
	 * accepts a u32 denominator), but we can also skip
	 * checking anything < 1Hz which implicitly can't be
	 * limited via an integer oa_max_sample_rate.
	 */
	if (oa_period <= NSEC_PER_SEC1000000000L) {
		u64 tmp = NSEC_PER_SEC1000000000L;
		do_div(tmp, oa_period)({ uint32_t __base = (oa_period); uint32_t __rem = ((uint64_t
)(tmp)) % __base; (tmp) = ((uint64_t)(tmp)) / __base; __rem; }
);
		oa_freq_hz = tmp;
	} else
		oa_freq_hz = 0;

	if (oa_freq_hz > i915_oa_max_sample_rate && !perfmon_capable()) {
		drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA exponent would exceed the max sampling frequency (sysctl dev.i915.oa_max_sample_rate) %uHz without CAP_PERFMON or CAP_SYS_ADMIN privileges\n"
, i915_oa_max_sample_rate)
			"OA exponent would exceed the max sampling frequency (sysctl dev.i915.oa_max_sample_rate) %uHz without CAP_PERFMON or CAP_SYS_ADMIN privileges\n",__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA exponent would exceed the max sampling frequency (sysctl dev.i915.oa_max_sample_rate) %uHz without CAP_PERFMON or CAP_SYS_ADMIN privileges\n"
, i915_oa_max_sample_rate)
			  i915_oa_max_sample_rate)__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA exponent would exceed the max sampling frequency (sysctl dev.i915.oa_max_sample_rate) %uHz without CAP_PERFMON or CAP_SYS_ADMIN privileges\n"
, i915_oa_max_sample_rate);
		return -EACCES13;
	}

	props->oa_periodic = true1;
	props->oa_period_exponent = value;
	break;
case DRM_I915_PERF_PROP_HOLD_PREEMPTION:
	props->hold_preemption = !!value;
	break;
case DRM_I915_PERF_PROP_GLOBAL_SSEU: {
	struct drm_i915_gem_context_param_sseu user_sseu;

	if (GRAPHICS_VER_FULL(perf->i915)(((&(perf->i915)->__runtime)->graphics.ip.ver) <<
| ((&(perf->i915)->__runtime)->graphics.ip.rel
)) >= IP_VER(12, 50)((12) << 8 | (50))) {
		drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "SSEU config not supported on gfx %x\n"
, (((&(perf->i915)->__runtime)->graphics.ip.ver)
 << 8 | ((&(perf->i915)->__runtime)->graphics
.ip.rel)))
			"SSEU config not supported on gfx %x\n",__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "SSEU config not supported on gfx %x\n"
, (((&(perf->i915)->__runtime)->graphics.ip.ver)
 << 8 | ((&(perf->i915)->__runtime)->graphics
.ip.rel)))
			GRAPHICS_VER_FULL(perf->i915))__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "SSEU config not supported on gfx %x\n"
, (((&(perf->i915)->__runtime)->graphics.ip.ver)
 << 8 | ((&(perf->i915)->__runtime)->graphics
.ip.rel)));
		return -ENODEV19;
	}

	if (copy_from_user(&user_sseu,
			   u64_to_user_ptr(value)((void *)(uintptr_t)(value)),
			   sizeof(user_sseu))) {
		drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unable to copy global sseu parameter\n"
)
			"Unable to copy global sseu parameter\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unable to copy global sseu parameter\n"
);
		return -EFAULT14;
	}

	ret = get_sseu_config(&props->sseu, props->engine, &user_sseu);
	if (ret) {
		drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Invalid SSEU configuration\n"
)
			"Invalid SSEU configuration\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Invalid SSEU configuration\n"
);
		return ret;
	}
	props->has_sseu = true1;
	break;
}
case DRM_I915_PERF_PROP_POLL_OA_PERIOD:
	if (value < 100000 /* 100us */) {
		drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA availability timer too small (%lluns < 100us)\n"
, value)
			"OA availability timer too small (%lluns < 100us)\n",__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA availability timer too small (%lluns < 100us)\n"
, value)
			  value)__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA availability timer too small (%lluns < 100us)\n"
, value);
		return -EINVAL22;
	}
	props->poll_oa_period = value;
	break;
case DRM_I915_PERF_PROP_MAX:
	MISSING_CASE(id)({ int __ret = !!(1); if (__ret) printf("Missing case (%s == %ld)\n"
, "id", (long)(id)); __builtin_expect(!!(__ret), 0); });
	return -EINVAL22;
}

uprop += 2;
}

return 0;
3814}

3816/**
* i915_perf_open_ioctl - DRM ioctl() for userspace to open a stream FD
* @dev: drm device
* @data: ioctl data copied from userspace (unvalidated)
* @file: drm file
*
* Validates the stream open parameters given by userspace including flags
* and an array of u64 key, value pair properties.
*
* Very little is assumed up front about the nature of the stream being
* opened (for instance we don't assume it's for periodic OA unit metrics). An
* i915-perf stream is expected to be a suitable interface for other forms of
* buffered data written by the GPU besides periodic OA metrics.
*
* Note we copy the properties from userspace outside of the i915 perf
* mutex to avoid an awkward lockdep with mmap_lock.
*
* Most of the implementation details are handled by
* i915_perf_open_ioctl_locked() after taking the &perf->lock
* mutex for serializing with any non-file-operation driver hooks.
*
* Return: A newly opened i915 Perf stream file descriptor or negative
* error code on failure.
*/
3840int i915_perf_open_ioctl(struct drm_device *dev, void *data,
	 struct drm_file *file)
3842{
struct i915_perf *perf = &to_i915(dev)->perf;
struct drm_i915_perf_open_param *param = data;
struct perf_open_properties props;
u32 known_open_flags;
int ret;

if (!perf->i915)
return -ENOTSUPP91;

known_open_flags = I915_PERF_FLAG_FD_CLOEXEC(1<<0) |
	   I915_PERF_FLAG_FD_NONBLOCK(1<<1) |
	   I915_PERF_FLAG_DISABLED(1<<2);
if (param->flags & ~known_open_flags) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unknown drm_i915_perf_open_param flag\n"
)
	"Unknown drm_i915_perf_open_param flag\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Unknown drm_i915_perf_open_param flag\n"
);
return -EINVAL22;
}

ret = read_properties_unlocked(perf,
		       u64_to_user_ptr(param->properties_ptr)((void *)(uintptr_t)(param->properties_ptr)),
		       param->num_properties,
		       &props);
if (ret)
return ret;

mutex_lock(&perf->lock)rw_enter_write(&perf->lock);
ret = i915_perf_open_ioctl_locked(perf, param, &props, file);
mutex_unlock(&perf->lock)rw_exit_write(&perf->lock);

return ret;
3873}

3875/**
* i915_perf_register - exposes i915-perf to userspace
* @i915: i915 device instance
*
* In particular OA metric sets are advertised under a sysfs metrics/
* directory allowing userspace to enumerate valid IDs that can be
* used to open an i915-perf stream.
*/
3883void i915_perf_register(struct drm_i915_privateinteldrm_softc *i915)
3884{
3885#ifdef __linux__
struct i915_perf *perf = &i915->perf;

if (!perf->i915)
return;

/* To be sure we're synchronized with an attempted
* i915_perf_open_ioctl(); considering that we register after
* being exposed to userspace.
*/
mutex_lock(&perf->lock)rw_enter_write(&perf->lock);

perf->metrics_kobj =
kobject_create_and_add("metrics",
		       &i915->drm.primary->kdev->kobj);

mutex_unlock(&perf->lock)rw_exit_write(&perf->lock);
3902#endif
3903}

3905/**
* i915_perf_unregister - hide i915-perf from userspace
* @i915: i915 device instance
*
* i915-perf state cleanup is split up into an 'unregister' and
* 'deinit' phase where the interface is first hidden from
* userspace by i915_perf_unregister() before cleaning up
* remaining state in i915_perf_fini().
*/
3914void i915_perf_unregister(struct drm_i915_privateinteldrm_softc *i915)
3915{
struct i915_perf *perf = &i915->perf;

if (!perf->metrics_kobj)
return;

kobject_put(perf->metrics_kobj);
perf->metrics_kobj = NULL((void *)0);
3923}

3925static bool_Bool gen8_is_valid_flex_addr(struct i915_perf *perf, u32 addr)
3926{
static const i915_reg_t flex_eu_regs[] = {
EU_PERF_CNTL0((const i915_reg_t){ .reg = (0xe458) }),
EU_PERF_CNTL1((const i915_reg_t){ .reg = (0xe558) }),
EU_PERF_CNTL2((const i915_reg_t){ .reg = (0xe658) }),
EU_PERF_CNTL3((const i915_reg_t){ .reg = (0xe758) }),
EU_PERF_CNTL4((const i915_reg_t){ .reg = (0xe45c) }),
EU_PERF_CNTL5((const i915_reg_t){ .reg = (0xe55c) }),
EU_PERF_CNTL6((const i915_reg_t){ .reg = (0xe65c) }),
};
int i;

for (i = 0; i < ARRAY_SIZE(flex_eu_regs)(sizeof((flex_eu_regs)) / sizeof((flex_eu_regs)[0])); i++) {
if (i915_mmio_reg_offset(flex_eu_regs[i]) == addr)
	return true1;
}
return false0;
3943}

3945static bool_Bool reg_in_range_table(u32 addr, const struct i915_range *table)
3946{
while (table->start || table->end) {
if (addr >= table->start && addr <= table->end)
	return true1;

table++;
}

return false0;
3955}

3957#define REG_EQUAL(addr, mmio)((addr) == i915_mmio_reg_offset(mmio)) \
((addr) == i915_mmio_reg_offset(mmio))

3960static const struct i915_range gen7_oa_b_counters[] = {
{ .start = 0x2710, .end = 0x272c },	/* OASTARTTRIG[1-8] */
{ .start = 0x2740, .end = 0x275c },	/* OAREPORTTRIG[1-8] */
{ .start = 0x2770, .end = 0x27ac },	/* OACEC[0-7][0-1] */
{}
3965};

3967static const struct i915_range gen12_oa_b_counters[] = {
{ .start = 0x2b2c, .end = 0x2b2c },	/* GEN12_OAG_OA_PESS */
{ .start = 0xd900, .end = 0xd91c },	/* GEN12_OAG_OASTARTTRIG[1-8] */
{ .start = 0xd920, .end = 0xd93c },	/* GEN12_OAG_OAREPORTTRIG1[1-8] */
{ .start = 0xd940, .end = 0xd97c },	/* GEN12_OAG_CEC[0-7][0-1] */
{ .start = 0xdc00, .end = 0xdc3c },	/* GEN12_OAG_SCEC[0-7][0-1] */
{ .start = 0xdc40, .end = 0xdc40 },	/* GEN12_OAG_SPCTR_CNF */
{ .start = 0xdc44, .end = 0xdc44 },	/* GEN12_OAA_DBG_REG */
{}
3976};

3978static const struct i915_range gen7_oa_mux_regs[] = {
{ .start = 0x91b8, .end = 0x91cc },	/* OA_PERFCNT[1-2], OA_PERFMATRIX */
{ .start = 0x9800, .end = 0x9888 },	/* MICRO_BP0_0 - NOA_WRITE */
{ .start = 0xe180, .end = 0xe180 },	/* HALF_SLICE_CHICKEN2 */
{}
3983};

3985static const struct i915_range hsw_oa_mux_regs[] = {
{ .start = 0x09e80, .end = 0x09ea4 }, /* HSW_MBVID2_NOA[0-9] */
{ .start = 0x09ec0, .end = 0x09ec0 }, /* HSW_MBVID2_MISR0 */
{ .start = 0x25100, .end = 0x2ff90 },
{}
3990};

3992static const struct i915_range chv_oa_mux_regs[] = {
{ .start = 0x182300, .end = 0x1823a4 },
{}
3995};

3997static const struct i915_range gen8_oa_mux_regs[] = {
{ .start = 0x0d00, .end = 0x0d2c },	/* RPM_CONFIG[0-1], NOA_CONFIG[0-8] */
{ .start = 0x20cc, .end = 0x20cc },	/* WAIT_FOR_RC6_EXIT */
{}
4001};

4003static const struct i915_range gen11_oa_mux_regs[] = {
{ .start = 0x91c8, .end = 0x91dc },	/* OA_PERFCNT[3-4] */
{}
4006};

4008static const struct i915_range gen12_oa_mux_regs[] = {
{ .start = 0x0d00, .end = 0x0d04 },     /* RPM_CONFIG[0-1] */
{ .start = 0x0d0c, .end = 0x0d2c },     /* NOA_CONFIG[0-8] */
{ .start = 0x9840, .end = 0x9840 },	/* GDT_CHICKEN_BITS */
{ .start = 0x9884, .end = 0x9888 },	/* NOA_WRITE */
{ .start = 0x20cc, .end = 0x20cc },	/* WAIT_FOR_RC6_EXIT */
{}
4015};

4017static bool_Bool gen7_is_valid_b_counter_addr(struct i915_perf *perf, u32 addr)
4018{
return reg_in_range_table(addr, gen7_oa_b_counters);
4020}

4022static bool_Bool gen8_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
4023{
return reg_in_range_table(addr, gen7_oa_mux_regs) ||
reg_in_range_table(addr, gen8_oa_mux_regs);
4026}

4028static bool_Bool gen11_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
4029{
return reg_in_range_table(addr, gen7_oa_mux_regs) ||
reg_in_range_table(addr, gen8_oa_mux_regs) ||
reg_in_range_table(addr, gen11_oa_mux_regs);
4033}

4035static bool_Bool hsw_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
4036{
return reg_in_range_table(addr, gen7_oa_mux_regs) ||
reg_in_range_table(addr, hsw_oa_mux_regs);
4039}

4041static bool_Bool chv_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
4042{
return reg_in_range_table(addr, gen7_oa_mux_regs) ||
reg_in_range_table(addr, chv_oa_mux_regs);
4045}

4047static bool_Bool gen12_is_valid_b_counter_addr(struct i915_perf *perf, u32 addr)
4048{
return reg_in_range_table(addr, gen12_oa_b_counters);
4050}

4052static bool_Bool gen12_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
4053{
return reg_in_range_table(addr, gen12_oa_mux_regs);
4055}

4057#ifdef notyet

4059static u32 mask_reg_value(u32 reg, u32 val)
4060{
/* HALF_SLICE_CHICKEN2 is programmed with a the
* WaDisableSTUnitPowerOptimization workaround. Make sure the value
* programmed by userspace doesn't change this.
*/
if (REG_EQUAL(reg, HALF_SLICE_CHICKEN2)((reg) == i915_mmio_reg_offset(((const i915_reg_t){ .reg = (0xe180
) }))))
val = val & ~_MASKED_BIT_ENABLE(GEN8_ST_PO_DISABLE)({ typeof((1 << 13)) _a = ((1 << 13)); ({ if (__builtin_constant_p
(_a)) do { } while (0); if (__builtin_constant_p(_a)) do { } while
 (0); if (__builtin_constant_p(_a) && __builtin_constant_p
(_a)) do { } while (0); ((_a) << 16 | (_a)); }); });

/* WAIT_FOR_RC6_EXIT has only one bit fullfilling the function
* indicated by its name and a bunch of selection fields used by OA
* configs.
*/
if (REG_EQUAL(reg, WAIT_FOR_RC6_EXIT)((reg) == i915_mmio_reg_offset(((const i915_reg_t){ .reg = (0x20cc
) }))))
val = val & ~_MASKED_BIT_ENABLE(HSW_WAIT_FOR_RC6_EXIT_ENABLE)({ typeof((1 << 0)) _a = ((1 << 0)); ({ if (__builtin_constant_p
(_a)) do { } while (0); if (__builtin_constant_p(_a)) do { } while
 (0); if (__builtin_constant_p(_a) && __builtin_constant_p
(_a)) do { } while (0); ((_a) << 16 | (_a)); }); });

return val;
4076}

4078static struct i915_oa_reg *alloc_oa_regs(struct i915_perf *perf,
			 bool_Bool (*is_valid)(struct i915_perf *perf, u32 addr),
			 u32 __user *regs,
			 u32 n_regs)
4082{
struct i915_oa_reg *oa_regs;
int err;
u32 i;

if (!n_regs)
return NULL((void *)0);

/* No is_valid function means we're not allowing any register to be programmed. */
GEM_BUG_ON(!is_valid)((void)0);
if (!is_valid)
return ERR_PTR(-EINVAL22);

oa_regs = kmalloc_array(n_regs, sizeof(*oa_regs), GFP_KERNEL(0x0001 | 0x0004));
if (!oa_regs)
return ERR_PTR(-ENOMEM12);

for (i = 0; i < n_regs; i++) {
u32 addr, value;

err = get_user(addr, regs)-copyin(regs, &(addr), sizeof(addr));
if (err)
	goto addr_err;

if (!is_valid(perf, addr)) {
	drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Invalid oa_reg address: %X\n"
, addr)
		"Invalid oa_reg address: %X\n", addr)__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Invalid oa_reg address: %X\n"
, addr);
	err = -EINVAL22;
	goto addr_err;
}

err = get_user(value, regs + 1)-copyin(regs + 1, &(value), sizeof(value));
if (err)
	goto addr_err;

oa_regs[i].addr = _MMIO(addr)((const i915_reg_t){ .reg = (addr) });
oa_regs[i].value = mask_reg_value(addr, value);

regs += 2;
}

return oa_regs;

4125addr_err:
kfree(oa_regs);
return ERR_PTR(err);
4128}

4130static ssize_t show_dynamic_id(struct kobject *kobj,
	       struct kobj_attribute *attr,
	       char *buf)
4133{
struct i915_oa_config *oa_config =
container_of(attr, typeof(*oa_config), sysfs_metric_id)({ const __typeof( ((typeof(*oa_config) *)0)->sysfs_metric_id
 ) *__mptr = (attr); (typeof(*oa_config) *)( (char *)__mptr -
 __builtin_offsetof(typeof(*oa_config), sysfs_metric_id) );});

return sprintf(buf, "%d\n", oa_config->id);
4138}

4140static int create_dynamic_oa_sysfs_entry(struct i915_perf *perf,
			 struct i915_oa_config *oa_config)
4142{
sysfs_attr_init(&oa_config->sysfs_metric_id.attr);
oa_config->sysfs_metric_id.attr.name = "id";
oa_config->sysfs_metric_id.attr.mode = S_IRUGO;
oa_config->sysfs_metric_id.show = show_dynamic_id;
oa_config->sysfs_metric_id.store = NULL((void *)0);

oa_config->attrs[0] = &oa_config->sysfs_metric_id.attr;
oa_config->attrs[1] = NULL((void *)0);

oa_config->sysfs_metric.name = oa_config->uuid;
oa_config->sysfs_metric.attrs = oa_config->attrs;

return sysfs_create_group(perf->metrics_kobj,0
		  &oa_config->sysfs_metric)0;
4157}

4159#endif

4161/**
* i915_perf_add_config_ioctl - DRM ioctl() for userspace to add a new OA config
* @dev: drm device
* @data: ioctl data (pointer to struct drm_i915_perf_oa_config) copied from
*        userspace (unvalidated)
* @file: drm file
*
* Validates the submitted OA register to be saved into a new OA config that
* can then be used for programming the OA unit and its NOA network.
*
* Returns: A new allocated config number to be used with the perf open ioctl
* or a negative error code on failure.
*/
4174int i915_perf_add_config_ioctl(struct drm_device *dev, void *data,
	       struct drm_file *file)
4176{
STUB()do { printf("%s: stub\n", __func__); } while(0);
return -ENOSYS78;
4179#ifdef notyet
struct i915_perf *perf = &to_i915(dev)->perf;
struct drm_i915_perf_oa_config *args = data;
struct i915_oa_config *oa_config, *tmp;
struct i915_oa_reg *regs;
int err, id;

if (!perf->i915)
return -ENOTSUPP91;

if (!perf->metrics_kobj) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA metrics weren't advertised via sysfs\n"
)
	"OA metrics weren't advertised via sysfs\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA metrics weren't advertised via sysfs\n"
);
return -EINVAL22;
}

if (i915_perf_stream_paranoid && !perfmon_capable()) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Insufficient privileges to add i915 OA config\n"
)
	"Insufficient privileges to add i915 OA config\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Insufficient privileges to add i915 OA config\n"
);
return -EACCES13;
}

if ((!args->mux_regs_ptr || !args->n_mux_regs) &&
   (!args->boolean_regs_ptr || !args->n_boolean_regs) &&
   (!args->flex_regs_ptr || !args->n_flex_regs)) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "No OA registers given\n"
)
	"No OA registers given\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "No OA registers given\n"
);
return -EINVAL22;
}

oa_config = kzalloc(sizeof(*oa_config), GFP_KERNEL(0x0001 | 0x0004));
if (!oa_config) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to allocate memory for the OA config\n"
)
	"Failed to allocate memory for the OA config\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to allocate memory for the OA config\n"
);
return -ENOMEM12;
}

oa_config->perf = perf;
kref_init(&oa_config->ref);

if (!uuid_is_valid(args->uuid)) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Invalid uuid format for OA config\n"
)
	"Invalid uuid format for OA config\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Invalid uuid format for OA config\n"
);
err = -EINVAL22;
goto reg_err;
}

/* Last character in oa_config->uuid will be 0 because oa_config is
* kzalloc.
*/
memcpy(oa_config->uuid, args->uuid, sizeof(args->uuid))__builtin_memcpy((oa_config->uuid), (args->uuid), (sizeof
(args->uuid)));

oa_config->mux_regs_len = args->n_mux_regs;
regs = alloc_oa_regs(perf,
	     perf->ops.is_valid_mux_reg,
	     u64_to_user_ptr(args->mux_regs_ptr)((void *)(uintptr_t)(args->mux_regs_ptr)),
	     args->n_mux_regs);

if (IS_ERR(regs)) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create OA config for mux_regs\n"
)
	"Failed to create OA config for mux_regs\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create OA config for mux_regs\n"
);
err = PTR_ERR(regs);
goto reg_err;
}
oa_config->mux_regs = regs;

oa_config->b_counter_regs_len = args->n_boolean_regs;
regs = alloc_oa_regs(perf,
	     perf->ops.is_valid_b_counter_reg,
	     u64_to_user_ptr(args->boolean_regs_ptr)((void *)(uintptr_t)(args->boolean_regs_ptr)),
	     args->n_boolean_regs);

if (IS_ERR(regs)) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create OA config for b_counter_regs\n"
)
	"Failed to create OA config for b_counter_regs\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create OA config for b_counter_regs\n"
);
err = PTR_ERR(regs);
goto reg_err;
}
oa_config->b_counter_regs = regs;

if (GRAPHICS_VER(perf->i915)((&(perf->i915)->__runtime)->graphics.ip.ver) < 8) {
if (args->n_flex_regs != 0) {
	err = -EINVAL22;
	goto reg_err;
}
} else {
oa_config->flex_regs_len = args->n_flex_regs;
regs = alloc_oa_regs(perf,
		     perf->ops.is_valid_flex_reg,
		     u64_to_user_ptr(args->flex_regs_ptr)((void *)(uintptr_t)(args->flex_regs_ptr)),
		     args->n_flex_regs);

if (IS_ERR(regs)) {
	drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create OA config for flex_regs\n"
)
		"Failed to create OA config for flex_regs\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create OA config for flex_regs\n"
);
	err = PTR_ERR(regs);
	goto reg_err;
}
oa_config->flex_regs = regs;
}

err = mutex_lock_interruptible(&perf->metrics_lock);
if (err)
goto reg_err;

/* We shouldn't have too many configs, so this iteration shouldn't be
* too costly.
*/
idr_for_each_entry(&perf->metrics_idr, tmp, id)for (id = 0; ((tmp) = idr_get_next(&perf->metrics_idr,
 &(id))) != ((void *)0); id++) {
if (!strcmp(tmp->uuid, oa_config->uuid)) {
	drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA config already exists with this uuid\n"
)
		"OA config already exists with this uuid\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "OA config already exists with this uuid\n"
);
	err = -EADDRINUSE48;
	goto sysfs_err;
}
}

err = create_dynamic_oa_sysfs_entry(perf, oa_config);
if (err) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create sysfs entry for OA config\n"
)
	"Failed to create sysfs entry for OA config\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create sysfs entry for OA config\n"
);
goto sysfs_err;
}

/* Config id 0 is invalid, id 1 for kernel stored test config. */
oa_config->id = idr_alloc(&perf->metrics_idr,
		  oa_config, 2,
		  0, GFP_KERNEL(0x0001 | 0x0004));
if (oa_config->id < 0) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create sysfs entry for OA config\n"
)
	"Failed to create sysfs entry for OA config\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to create sysfs entry for OA config\n"
);
err = oa_config->id;
goto sysfs_err;
}
id = oa_config->id;

drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Added config %s id=%i\n"
, oa_config->uuid, oa_config->id)
"Added config %s id=%i\n", oa_config->uuid, oa_config->id)__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Added config %s id=%i\n"
, oa_config->uuid, oa_config->id);
mutex_unlock(&perf->metrics_lock)rw_exit_write(&perf->metrics_lock);

return id;

4321sysfs_err:
mutex_unlock(&perf->metrics_lock)rw_exit_write(&perf->metrics_lock);
4323reg_err:
i915_oa_config_put(oa_config);
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to add new OA config\n"
)
"Failed to add new OA config\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to add new OA config\n"
);
return err;
4328#endif
4329}

4331/**
* i915_perf_remove_config_ioctl - DRM ioctl() for userspace to remove an OA config
* @dev: drm device
* @data: ioctl data (pointer to u64 integer) copied from userspace
* @file: drm file
*
* Configs can be removed while being used, the will stop appearing in sysfs
* and their content will be freed when the stream using the config is closed.
*
* Returns: 0 on success or a negative error code on failure.
*/
4342int i915_perf_remove_config_ioctl(struct drm_device *dev, void *data,
		  struct drm_file *file)
4344{
struct i915_perf *perf = &to_i915(dev)->perf;
u64 *arg = data;
struct i915_oa_config *oa_config;
int ret;

if (!perf->i915)
return -ENOTSUPP91;

if (i915_perf_stream_paranoid && !perfmon_capable()) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Insufficient privileges to remove i915 OA config\n"
)
	"Insufficient privileges to remove i915 OA config\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Insufficient privileges to remove i915 OA config\n"
);
return -EACCES13;
}

ret = mutex_lock_interruptible(&perf->metrics_lock);
if (ret)
return ret;

oa_config = idr_find(&perf->metrics_idr, *arg);
if (!oa_config) {
drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to remove unknown OA config\n"
)
	"Failed to remove unknown OA config\n")__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Failed to remove unknown OA config\n"
);
ret = -ENOENT2;
goto err_unlock;
}

GEM_BUG_ON(*arg != oa_config->id)((void)0);

sysfs_remove_group(perf->metrics_kobj, &oa_config->sysfs_metric);

idr_remove(&perf->metrics_idr, *arg);

mutex_unlock(&perf->metrics_lock)rw_exit_write(&perf->metrics_lock);

drm_dbg(&perf->i915->drm,__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Removed config %s id=%i\n"
, oa_config->uuid, oa_config->id)
"Removed config %s id=%i\n", oa_config->uuid, oa_config->id)__drm_dev_dbg(((void *)0), (&perf->i915->drm) ? (&
perf->i915->drm)->dev : ((void *)0), DRM_UT_DRIVER, "Removed config %s id=%i\n"
, oa_config->uuid, oa_config->id);

i915_oa_config_put(oa_config);

return 0;

4386err_unlock:
mutex_unlock(&perf->metrics_lock)rw_exit_write(&perf->metrics_lock);
return ret;
4389}

4391#ifdef notyet
4392static struct ctl_table oa_table[] = {
{
.procname = "perf_stream_paranoid",
.data = &i915_perf_stream_paranoid,
.maxlen = sizeof(i915_perf_stream_paranoid),
.mode = 0644,
.proc_handler = proc_dointvec_minmax,
.extra1 = SYSCTL_ZERO,
.extra2 = SYSCTL_ONE,
},
{
.procname = "oa_max_sample_rate",
.data = &i915_oa_max_sample_rate,
.maxlen = sizeof(i915_oa_max_sample_rate),
.mode = 0644,
.proc_handler = proc_dointvec_minmax,
.extra1 = SYSCTL_ZERO,
.extra2 = &oa_sample_rate_hard_limit,
},
{}
4412};
4413#endif

4415static void oa_init_supported_formats(struct i915_perf *perf)
4416{
struct drm_i915_privateinteldrm_softc *i915 = perf->i915;
enum intel_platform platform = INTEL_INFO(i915)(&(i915)->__info)->platform;

switch (platform) {
case INTEL_HASWELL:
oa_format_add(perf, I915_OA_FORMAT_A13);
oa_format_add(perf, I915_OA_FORMAT_A13);
oa_format_add(perf, I915_OA_FORMAT_A29);
oa_format_add(perf, I915_OA_FORMAT_A13_B8_C8);
oa_format_add(perf, I915_OA_FORMAT_B4_C8);
oa_format_add(perf, I915_OA_FORMAT_A45_B8_C8);
oa_format_add(perf, I915_OA_FORMAT_B4_C8_A16);
oa_format_add(perf, I915_OA_FORMAT_C4_B8);
break;

case INTEL_BROADWELL:
case INTEL_CHERRYVIEW:
case INTEL_SKYLAKE:
case INTEL_BROXTON:
case INTEL_KABYLAKE:
case INTEL_GEMINILAKE:
case INTEL_COFFEELAKE:
case INTEL_COMETLAKE:
case INTEL_ICELAKE:
case INTEL_ELKHARTLAKE:
case INTEL_JASPERLAKE:
case INTEL_TIGERLAKE:
case INTEL_ROCKETLAKE:
case INTEL_DG1:
case INTEL_ALDERLAKE_S:
case INTEL_ALDERLAKE_P:
oa_format_add(perf, I915_OA_FORMAT_A12);
oa_format_add(perf, I915_OA_FORMAT_A12_B8_C8);
oa_format_add(perf, I915_OA_FORMAT_A32u40_A4u32_B8_C8);
oa_format_add(perf, I915_OA_FORMAT_C4_B8);
break;

default:
MISSING_CASE(platform)({ int __ret = !!(1); if (__ret) printf("Missing case (%s == %ld)\n"
, "platform", (long)(platform)); __builtin_expect(!!(__ret), 0
); });
}
4457}

4459/**
* i915_perf_init - initialize i915-perf state on module bind
* @i915: i915 device instance
*
* Initializes i915-perf state without exposing anything to userspace.
*
* Note: i915-perf initialization is split into an 'init' and 'register'
* phase with the i915_perf_register() exposing state to userspace.
*/
4468void i915_perf_init(struct drm_i915_privateinteldrm_softc *i915)
4469{
struct i915_perf *perf = &i915->perf;

/* XXX const struct i915_perf_ops! */

/* i915_perf is not enabled for DG2 yet */
if (IS_DG2(i915)IS_PLATFORM(i915, INTEL_DG2))
return;

perf->oa_formats = oa_formats;
if (IS_HASWELL(i915)IS_PLATFORM(i915, INTEL_HASWELL)) {
perf->ops.is_valid_b_counter_reg = gen7_is_valid_b_counter_addr;
perf->ops.is_valid_mux_reg = hsw_is_valid_mux_addr;
perf->ops.is_valid_flex_reg = NULL((void *)0);
perf->ops.enable_metric_set = hsw_enable_metric_set;
perf->ops.disable_metric_set = hsw_disable_metric_set;
perf->ops.oa_enable = gen7_oa_enable;
perf->ops.oa_disable = gen7_oa_disable;
perf->ops.read = gen7_oa_read;
perf->ops.oa_hw_tail_read = gen7_oa_hw_tail_read;
} else if (HAS_LOGICAL_RING_CONTEXTS(i915)((&(i915)->__info)->has_logical_ring_contexts)) {
/* Note: that although we could theoretically also support the
 * legacy ringbuffer mode on BDW (and earlier iterations of
 * this driver, before upstreaming did this) it didn't seem
 * worth the complexity to maintain now that BDW+ enable
 * execlist mode by default.
 */
perf->ops.read = gen8_oa_read;

if (IS_GRAPHICS_VER(i915, 8, 9)(((&(i915)->__runtime)->graphics.ip.ver) >= (8) &&
 ((&(i915)->__runtime)->graphics.ip.ver) <= (9))) {
	perf->ops.is_valid_b_counter_reg =
		gen7_is_valid_b_counter_addr;
	perf->ops.is_valid_mux_reg =
		gen8_is_valid_mux_addr;
	perf->ops.is_valid_flex_reg =
		gen8_is_valid_flex_addr;

	if (IS_CHERRYVIEW(i915)IS_PLATFORM(i915, INTEL_CHERRYVIEW)) {
		perf->ops.is_valid_mux_reg =
			chv_is_valid_mux_addr;
	}

	perf->ops.oa_enable = gen8_oa_enable;
	perf->ops.oa_disable = gen8_oa_disable;
	perf->ops.enable_metric_set = gen8_enable_metric_set;
	perf->ops.disable_metric_set = gen8_disable_metric_set;
	perf->ops.oa_hw_tail_read = gen8_oa_hw_tail_read;

	if (GRAPHICS_VER(i915)((&(i915)->__runtime)->graphics.ip.ver) == 8) {
		perf->ctx_oactxctrl_offset = 0x120;
		perf->ctx_flexeu0_offset = 0x2ce;

		perf->gen8_valid_ctx_bit = BIT(25)(1UL << (25));
	} else {
		perf->ctx_oactxctrl_offset = 0x128;
		perf->ctx_flexeu0_offset = 0x3de;

		perf->gen8_valid_ctx_bit = BIT(16)(1UL << (16));
	}
} else if (GRAPHICS_VER(i915)((&(i915)->__runtime)->graphics.ip.ver) == 11) {
	perf->ops.is_valid_b_counter_reg =
		gen7_is_valid_b_counter_addr;
	perf->ops.is_valid_mux_reg =
		gen11_is_valid_mux_addr;
	perf->ops.is_valid_flex_reg =
		gen8_is_valid_flex_addr;

	perf->ops.oa_enable = gen8_oa_enable;
	perf->ops.oa_disable = gen8_oa_disable;
	perf->ops.enable_metric_set = gen8_enable_metric_set;
	perf->ops.disable_metric_set = gen11_disable_metric_set;
	perf->ops.oa_hw_tail_read = gen8_oa_hw_tail_read;

	perf->ctx_oactxctrl_offset = 0x124;
	perf->ctx_flexeu0_offset = 0x78e;

	perf->gen8_valid_ctx_bit = BIT(16)(1UL << (16));
} else if (GRAPHICS_VER(i915)((&(i915)->__runtime)->graphics.ip.ver) == 12) {
	perf->ops.is_valid_b_counter_reg =
		gen12_is_valid_b_counter_addr;
	perf->ops.is_valid_mux_reg =
		gen12_is_valid_mux_addr;
	perf->ops.is_valid_flex_reg =
		gen8_is_valid_flex_addr;

	perf->ops.oa_enable = gen12_oa_enable;
	perf->ops.oa_disable = gen12_oa_disable;
	perf->ops.enable_metric_set = gen12_enable_metric_set;
	perf->ops.disable_metric_set = gen12_disable_metric_set;
	perf->ops.oa_hw_tail_read = gen12_oa_hw_tail_read;

	perf->ctx_flexeu0_offset = 0;
	perf->ctx_oactxctrl_offset = 0x144;
}
}

if (perf->ops.enable_metric_set) {
rw_init(&perf->lock, "perflk")_rw_init_flags(&perf->lock, "perflk", 0, ((void *)0));

/* Choose a representative limit */
oa_sample_rate_hard_limit = to_gt(i915)->clock_frequency / 2;

rw_init(&perf->metrics_lock, "metricslk")_rw_init_flags(&perf->metrics_lock, "metricslk", 0, ((
void *)0));
idr_init_base(&perf->metrics_idr, 1);

/* We set up some ratelimit state to potentially throttle any
 * _NOTES about spurious, invalid OA reports which we don't
 * forward to userspace.
 *
 * We print a _NOTE about any throttling when closing the
 * stream instead of waiting until driver _fini which no one
 * would ever see.
 *
 * Using the same limiting factors as printk_ratelimit()
 */
ratelimit_state_init(&perf->spurious_report_rs, 5 * HZhz, 10);
/* Since we use a DRM_NOTE for spurious reports it would be
 * inconsistent to let __ratelimit() automatically print a
 * warning for throttling.
 */
ratelimit_set_flags(&perf->spurious_report_rs,
		    RATELIMIT_MSG_ON_RELEASE(1 << 0));

ratelimit_state_init(&perf->tail_pointer_race,
		     5 * HZhz, 10);
ratelimit_set_flags(&perf->tail_pointer_race,
		    RATELIMIT_MSG_ON_RELEASE(1 << 0));

atomic64_set(&perf->noa_programming_delay,({ typeof(*(&perf->noa_programming_delay)) __tmp = ((500
 * 1000)); *(volatile typeof(*(&perf->noa_programming_delay
)) *)&(*(&perf->noa_programming_delay)) = __tmp; __tmp
; })
	     500 * 1000 /* 500us */)({ typeof(*(&perf->noa_programming_delay)) __tmp = ((500
 * 1000)); *(volatile typeof(*(&perf->noa_programming_delay
)) *)&(*(&perf->noa_programming_delay)) = __tmp; __tmp
; });

perf->i915 = i915;

oa_init_supported_formats(perf);
}
4604}

4606static int destroy_config(int id, void *p, void *data)
4607{
i915_oa_config_put(p);
return 0;
4610}

4612int i915_perf_sysctl_register(void)
4613{
4614#ifdef notyet
sysctl_header = register_sysctl("dev/i915", oa_table);
4616#endif
return 0;
4618}

4620void i915_perf_sysctl_unregister(void)
4621{
4622#ifdef notyet
unregister_sysctl_table(sysctl_header);
4624#endif
4625}

4627/**
* i915_perf_fini - Counter part to i915_perf_init()
* @i915: i915 device instance
*/
4631void i915_perf_fini(struct drm_i915_privateinteldrm_softc *i915)
4632{
struct i915_perf *perf = &i915->perf;

if (!perf->i915)
return;

idr_for_each(&perf->metrics_idr, destroy_config, perf);
idr_destroy(&perf->metrics_idr);

memset(&perf->ops, 0, sizeof(perf->ops))__builtin_memset((&perf->ops), (0), (sizeof(perf->ops
)));
perf->i915 = NULL((void *)0);
4643}

4645/**
* i915_perf_ioctl_version - Version of the i915-perf subsystem
*
* This version number is used by userspace to detect available features.
*/
4650int i915_perf_ioctl_version(void)
4651{
/*
* 1: Initial version
*   I915_PERF_IOCTL_ENABLE
*   I915_PERF_IOCTL_DISABLE
*
* 2: Added runtime modification of OA config.
*   I915_PERF_IOCTL_CONFIG
*
* 3: Add DRM_I915_PERF_PROP_HOLD_PREEMPTION parameter to hold
*    preemption on a particular context so that performance data is
*    accessible from a delta of MI_RPC reports without looking at the
*    OA buffer.
*
* 4: Add DRM_I915_PERF_PROP_ALLOWED_SSEU to limit what contexts can
*    be run for the duration of the performance recording based on
*    their SSEU configuration.
*
* 5: Add DRM_I915_PERF_PROP_POLL_OA_PERIOD parameter that controls the
*    interval for the hrtimer used to check for OA data.
*/
return 5;
4673}

4675#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)0
4676#include "selftests/i915_perf.c"
4677#endif

←

/usr/src/sys/dev/pci/drm/include/linux/slab.h

→

1/* Public domain. */
2 
3#ifndef _LINUX_SLAB_H
4#define _LINUX_SLAB_H
5 
6#include <sys/types.h>
7#include <sys/malloc.h>
8 
9#include <linux/types.h>
10#include <linux/workqueue.h>
11#include <linux/gfp.h>
12 
13#include <linux/processor.h>	/* for CACHELINESIZE */
14 
15#define ARCH_KMALLOC_MINALIGN64 CACHELINESIZE64
16 
17#define ZERO_SIZE_PTR((void *)0) NULL((void *)0)
18 
19static inline void *
20kmalloc(size_t size, int flags)
21{
22	return malloc(size, M_DRM145, flags);
23}
24 
25static inline void *
26kmalloc_array(size_t n, size_t size, int flags)
27{
28	if (n != 0 && SIZE_MAX0xffffffffffffffffUL / n < size)
29		return NULL((void *)0);
30	return malloc(n * size, M_DRM145, flags);
31}
32 
33static inline void *
34kcalloc(size_t n, size_t size, int flags)
35{
36	if (n != 0 && SIZE_MAX0xffffffffffffffffUL / n < size)
37		return NULL((void *)0);
38	return malloc(n * size, M_DRM145, flags | M_ZERO0x0008);
39}
40 
41static inline void *
42kzalloc(size_t size, int flags)
43{
44	return malloc(size, M_DRM145, flags | M_ZERO0x0008);
15
←
Null pointer value stored to field 'vma'→
45}
46 
47static inline void
48kfree(const void *objp)
49{
50	free((void *)objp, M_DRM145, 0);
51}
52 
53#endif

←

/usr/src/sys/dev/pci/drm/i915/i915_vma.h

1/*
2 * Copyright © 2016 Intel Corporation
3 *
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
10 *
11 * The above copyright notice and this permission notice (including the next
12 * paragraph) shall be included in all copies or substantial portions of the
13 * Software.
14 *
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
18 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
20 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
21 * IN THE SOFTWARE.
22 *
23 */
24 
25#ifndef __I915_VMA_H__
26#define __I915_VMA_H__
27 
28#include <linux/io-mapping.h>
29#include <linux/rbtree.h>
30 
31#include <drm/drm_mm.h>
32 
33#include "gt/intel_ggtt_fencing.h"
34#include "gem/i915_gem_object.h"
35 
36#include "i915_gem_gtt.h"
37 
38#include "i915_active.h"
39#include "i915_request.h"
40#include "i915_vma_resource.h"
41#include "i915_vma_types.h"
42 
43struct i915_vma *
44i915_vma_instance(struct drm_i915_gem_object *obj,
45		  struct i915_address_space *vm,
46		  const struct i915_gtt_view *view);
47 
48void i915_vma_unpin_and_release(struct i915_vma **p_vma, unsigned int flags);
49#define I915_VMA_RELEASE_MAP(1UL << (0)) BIT(0)(1UL << (0))
50 
51static inline bool_Bool i915_vma_is_active(const struct i915_vma *vma)
52{
53	return !i915_active_is_idle(&vma->active);
54}
55 
56/* do not reserve memory to prevent deadlocks */
57#define __EXEC_OBJECT_NO_RESERVE(1UL << (31)) BIT(31)(1UL << (31))
58 
59int __must_check _i915_vma_move_to_active(struct i915_vma *vma,
60					  struct i915_request *rq,
61					  struct dma_fence *fence,
62					  unsigned int flags);
63static inline int __must_check
64i915_vma_move_to_active(struct i915_vma *vma, struct i915_request *rq,
65			unsigned int flags)
66{
67	return _i915_vma_move_to_active(vma, rq, &rq->fence, flags);
68}
69 
70#ifdef __linux__
71#define __i915_vma_flags(v)((unsigned long *)&(v)->flags) ((unsigned long *)&(v)->flags.counter)
72#else
73#define __i915_vma_flags(v)((unsigned long *)&(v)->flags) ((unsigned long *)&(v)->flags)
74#endif
75 
76static inline bool_Bool i915_vma_is_ggtt(const struct i915_vma *vma)
77{
78	return test_bit(I915_VMA_GGTT_BIT13, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
79}
80 
81static inline bool_Bool i915_vma_is_dpt(const struct i915_vma *vma)
82{
83	return i915_is_dpt(vma->vm)((vma->vm)->is_dpt);
84}
85 
86static inline bool_Bool i915_vma_has_ggtt_write(const struct i915_vma *vma)
87{
88	return test_bit(I915_VMA_GGTT_WRITE_BIT16, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
89}
90 
91static inline void i915_vma_set_ggtt_write(struct i915_vma *vma)
92{
93	GEM_BUG_ON(!i915_vma_is_ggtt(vma))((void)0);
94	set_bit(I915_VMA_GGTT_WRITE_BIT16, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
95}
96 
97static inline bool_Bool i915_vma_unset_ggtt_write(struct i915_vma *vma)
98{
99	return test_and_clear_bit(I915_VMA_GGTT_WRITE_BIT16,
100				  __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
101}
102 
103void i915_vma_flush_writes(struct i915_vma *vma);
104 
105static inline bool_Bool i915_vma_is_map_and_fenceable(const struct i915_vma *vma)
106{
107	return test_bit(I915_VMA_CAN_FENCE_BIT14, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
108}
109 
110static inline bool_Bool i915_vma_set_userfault(struct i915_vma *vma)
111{
112	GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma))((void)0);
113	return test_and_set_bit(I915_VMA_USERFAULT_BIT15, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
114}
115 
116static inline void i915_vma_unset_userfault(struct i915_vma *vma)
117{
118	return clear_bit(I915_VMA_USERFAULT_BIT15, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
119}
120 
121static inline bool_Bool i915_vma_has_userfault(const struct i915_vma *vma)
122{
123	return test_bit(I915_VMA_USERFAULT_BIT15, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
124}
125 
126static inline bool_Bool i915_vma_is_closed(const struct i915_vma *vma)
127{
128	return !list_empty(&vma->closed_link);
129}
130 
131static inline u32 i915_ggtt_offset(const struct i915_vma *vma)
132{
133	GEM_BUG_ON(!i915_vma_is_ggtt(vma))((void)0);
134	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node))((void)0);
135	GEM_BUG_ON(upper_32_bits(vma->node.start))((void)0);
136	GEM_BUG_ON(upper_32_bits(vma->node.start + vma->node.size - 1))((void)0);
137	return lower_32_bits(vma->node.start)((u32)(vma->node.start));
138}
139 
140static inline u32 i915_ggtt_pin_bias(struct i915_vma *vma)
141{
142	return i915_vm_to_ggtt(vma->vm)->pin_bias;
143}
144 
145static inline struct i915_vma *i915_vma_get(struct i915_vma *vma)
146{
147	i915_gem_object_get(vma->obj);
27
←
Access to field 'obj' results in a dereference of a null pointer (loaded from variable 'vma')
148	return vma;
149}
150 
151static inline struct i915_vma *i915_vma_tryget(struct i915_vma *vma)
152{
153	if (likely(kref_get_unless_zero(&vma->obj->base.refcount))__builtin_expect(!!(kref_get_unless_zero(&vma->obj->
base.refcount)), 1))
154		return vma;
155 
156	return NULL((void *)0);
157}
158 
159static inline void i915_vma_put(struct i915_vma *vma)
160{
161	i915_gem_object_put(vma->obj);
162}
163 
164static inline long
165i915_vma_compare(struct i915_vma *vma,
166		 struct i915_address_space *vm,
167		 const struct i915_gtt_view *view)
168{
169	ptrdiff_t cmp;
170 
171	GEM_BUG_ON(view && !i915_is_ggtt_or_dpt(vm))((void)0);
172 
173	cmp = ptrdiff(vma->vm, vm);
174	if (cmp)
175		return cmp;
176 
177	BUILD_BUG_ON(I915_GTT_VIEW_NORMAL != 0)extern char _ctassert[(!(I915_GTT_VIEW_NORMAL != 0)) ? 1 : -1
 ] __attribute__((__unused__));
178	cmp = vma->gtt_view.type;
179	if (!view)
180		return cmp;
181 
182	cmp -= view->type;
183	if (cmp)
184		return cmp;
185 
186	assert_i915_gem_gtt_types();
187 
188	/* gtt_view.type also encodes its size so that we both distinguish
189	 * different views using it as a "type" and also use a compact (no
190	 * accessing of uninitialised padding bytes) memcmp without storing
191	 * an extra parameter or adding more code.
192	 *
193	 * To ensure that the memcmp is valid for all branches of the union,
194	 * even though the code looks like it is just comparing one branch,
195	 * we assert above that all branches have the same address, and that
196	 * each branch has a unique type/size.
197	 */
198	BUILD_BUG_ON(I915_GTT_VIEW_NORMAL >= I915_GTT_VIEW_PARTIAL)extern char _ctassert[(!(I915_GTT_VIEW_NORMAL >= I915_GTT_VIEW_PARTIAL
)) ? 1 : -1 ] __attribute__((__unused__));
199	BUILD_BUG_ON(I915_GTT_VIEW_PARTIAL >= I915_GTT_VIEW_ROTATED)extern char _ctassert[(!(I915_GTT_VIEW_PARTIAL >= I915_GTT_VIEW_ROTATED
)) ? 1 : -1 ] __attribute__((__unused__));
200	BUILD_BUG_ON(I915_GTT_VIEW_ROTATED >= I915_GTT_VIEW_REMAPPED)extern char _ctassert[(!(I915_GTT_VIEW_ROTATED >= I915_GTT_VIEW_REMAPPED
)) ? 1 : -1 ] __attribute__((__unused__));
201	BUILD_BUG_ON(offsetof(typeof(*view), rotated) !=extern char _ctassert[(!(__builtin_offsetof(typeof(*view), rotated
) != __builtin_offsetof(typeof(*view), partial))) ? 1 : -1 ] __attribute__
((__unused__))
202		     offsetof(typeof(*view), partial))extern char _ctassert[(!(__builtin_offsetof(typeof(*view), rotated
) != __builtin_offsetof(typeof(*view), partial))) ? 1 : -1 ] __attribute__
((__unused__));
203	BUILD_BUG_ON(offsetof(typeof(*view), rotated) !=extern char _ctassert[(!(__builtin_offsetof(typeof(*view), rotated
) != __builtin_offsetof(typeof(*view), remapped))) ? 1 : -1 ]
 __attribute__((__unused__))
204		     offsetof(typeof(*view), remapped))extern char _ctassert[(!(__builtin_offsetof(typeof(*view), rotated
) != __builtin_offsetof(typeof(*view), remapped))) ? 1 : -1 ]
 __attribute__((__unused__));
205	return memcmp(&vma->gtt_view.partial, &view->partial, view->type)__builtin_memcmp((&vma->gtt_view.partial), (&view->
partial), (view->type));
206}
207 
208struct i915_vma_work *i915_vma_work(void);
209int i915_vma_bind(struct i915_vma *vma,
210		  enum i915_cache_level cache_level,
211		  u32 flags,
212		  struct i915_vma_work *work,
213		  struct i915_vma_resource *vma_res);
214 
215bool_Bool i915_gem_valid_gtt_space(struct i915_vma *vma, unsigned long color);
216bool_Bool i915_vma_misplaced(const struct i915_vma *vma,
217			u64 size, u64 alignment, u64 flags);
218void __i915_vma_set_map_and_fenceable(struct i915_vma *vma);
219void i915_vma_revoke_mmap(struct i915_vma *vma);
220void vma_invalidate_tlb(struct i915_address_space *vm, u32 *tlb);
221struct dma_fence *__i915_vma_evict(struct i915_vma *vma, bool_Bool async);
222int __i915_vma_unbind(struct i915_vma *vma);
223int __must_check i915_vma_unbind(struct i915_vma *vma);
224int __must_check i915_vma_unbind_async(struct i915_vma *vma, bool_Bool trylock_vm);
225int __must_check i915_vma_unbind_unlocked(struct i915_vma *vma);
226void i915_vma_unlink_ctx(struct i915_vma *vma);
227void i915_vma_close(struct i915_vma *vma);
228void i915_vma_reopen(struct i915_vma *vma);
229 
230void i915_vma_destroy_locked(struct i915_vma *vma);
231void i915_vma_destroy(struct i915_vma *vma);
232 
233#define assert_vma_held(vma)do { (void)(&((vma)->obj->base.resv)->lock.base)
; } while(0) dma_resv_assert_held((vma)->obj->base.resv)do { (void)(&((vma)->obj->base.resv)->lock.base)
; } while(0)
234 
235static inline void i915_vma_lock(struct i915_vma *vma)
236{
237	dma_resv_lock(vma->obj->base.resv, NULL((void *)0));
238}
239 
240static inline void i915_vma_unlock(struct i915_vma *vma)
241{
242	dma_resv_unlock(vma->obj->base.resv);
243}
244 
245int __must_check
246i915_vma_pin_ww(struct i915_vma *vma, struct i915_gem_ww_ctx *ww,
247		u64 size, u64 alignment, u64 flags);
248 
249static inline int __must_check
250i915_vma_pin(struct i915_vma *vma, u64 size, u64 alignment, u64 flags)
251{
252	struct i915_gem_ww_ctx ww;
253	int err;
254 
255	i915_gem_ww_ctx_init(&ww, true1);
256retry:
257	err = i915_gem_object_lock(vma->obj, &ww);
258	if (!err)
259		err = i915_vma_pin_ww(vma, &ww, size, alignment, flags);
260	if (err == -EDEADLK11) {
261		err = i915_gem_ww_ctx_backoff(&ww);
262		if (!err)
263			goto retry;
264	}
265	i915_gem_ww_ctx_fini(&ww);
266 
267	return err;
268}
269 
270int i915_ggtt_pin(struct i915_vma *vma, struct i915_gem_ww_ctx *ww,
271		  u32 align, unsigned int flags);
272 
273static inline int i915_vma_pin_count(const struct i915_vma *vma)
274{
275	return atomic_read(&vma->flags)({ typeof(*(&vma->flags)) __tmp = *(volatile typeof(*(
&vma->flags)) *)&(*(&vma->flags)); membar_datadep_consumer
(); __tmp; }) & I915_VMA_PIN_MASK0x3ff;
276}
277 
278static inline bool_Bool i915_vma_is_pinned(const struct i915_vma *vma)
279{
280	return i915_vma_pin_count(vma);
281}
282 
283static inline void __i915_vma_pin(struct i915_vma *vma)
284{
285	atomic_inc(&vma->flags)__sync_fetch_and_add(&vma->flags, 1);
286	GEM_BUG_ON(!i915_vma_is_pinned(vma))((void)0);
287}
288 
289static inline void __i915_vma_unpin(struct i915_vma *vma)
290{
291	GEM_BUG_ON(!i915_vma_is_pinned(vma))((void)0);
292	atomic_dec(&vma->flags)__sync_fetch_and_sub(&vma->flags, 1);
293}
294 
295static inline void i915_vma_unpin(struct i915_vma *vma)
296{
297	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node))((void)0);
298	__i915_vma_unpin(vma);
299}
300 
301static inline bool_Bool i915_vma_is_bound(const struct i915_vma *vma,
302				     unsigned int where)
303{
304	return atomic_read(&vma->flags)({ typeof(*(&vma->flags)) __tmp = *(volatile typeof(*(
&vma->flags)) *)&(*(&vma->flags)); membar_datadep_consumer
(); __tmp; }) & where;
305}
306 
307static inline bool_Bool i915_node_color_differs(const struct drm_mm_node *node,
308					   unsigned long color)
309{
310	return drm_mm_node_allocated(node) && node->color != color;
311}
312 
313/**
314 * i915_vma_pin_iomap - calls ioremap_wc to map the GGTT VMA via the aperture
315 * @vma: VMA to iomap
316 *
317 * The passed in VMA has to be pinned in the global GTT mappable region.
318 * An extra pinning of the VMA is acquired for the return iomapping,
319 * the caller must call i915_vma_unpin_iomap to relinquish the pinning
320 * after the iomapping is no longer required.
321 *
322 * Returns a valid iomapped pointer or ERR_PTR.
323 */
324void __iomem *i915_vma_pin_iomap(struct i915_vma *vma);
325 
326/**
327 * i915_vma_unpin_iomap - unpins the mapping returned from i915_vma_iomap
328 * @vma: VMA to unpin
329 *
330 * Unpins the previously iomapped VMA from i915_vma_pin_iomap().
331 *
332 * This function is only valid to be called on a VMA previously
333 * iomapped by the caller with i915_vma_pin_iomap().
334 */
335void i915_vma_unpin_iomap(struct i915_vma *vma);
336 
337/**
338 * i915_vma_pin_fence - pin fencing state
339 * @vma: vma to pin fencing for
340 *
341 * This pins the fencing state (whether tiled or untiled) to make sure the
342 * vma (and its object) is ready to be used as a scanout target. Fencing
343 * status must be synchronize first by calling i915_vma_get_fence():
344 *
345 * The resulting fence pin reference must be released again with
346 * i915_vma_unpin_fence().
347 *
348 * Returns:
349 *
350 * True if the vma has a fence, false otherwise.
351 */
352int __must_check i915_vma_pin_fence(struct i915_vma *vma);
353void i915_vma_revoke_fence(struct i915_vma *vma);
354 
355int __i915_vma_pin_fence(struct i915_vma *vma);
356 
357static inline void __i915_vma_unpin_fence(struct i915_vma *vma)
358{
359	GEM_BUG_ON(atomic_read(&vma->fence->pin_count) <= 0)((void)0);
360	atomic_dec(&vma->fence->pin_count)__sync_fetch_and_sub(&vma->fence->pin_count, 1);
361}
362 
363/**
364 * i915_vma_unpin_fence - unpin fencing state
365 * @vma: vma to unpin fencing for
366 *
367 * This releases the fence pin reference acquired through
368 * i915_vma_pin_fence. It will handle both objects with and without an
369 * attached fence correctly, callers do not need to distinguish this.
370 */
371static inline void
372i915_vma_unpin_fence(struct i915_vma *vma)
373{
374	if (vma->fence)
375		__i915_vma_unpin_fence(vma);
376}
377 
378void i915_vma_parked(struct intel_gt *gt);
379 
380static inline bool_Bool i915_vma_is_scanout(const struct i915_vma *vma)
381{
382	return test_bit(I915_VMA_SCANOUT_BIT17, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
383}
384 
385static inline void i915_vma_mark_scanout(struct i915_vma *vma)
386{
387	set_bit(I915_VMA_SCANOUT_BIT17, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
388}
389 
390static inline void i915_vma_clear_scanout(struct i915_vma *vma)
391{
392	clear_bit(I915_VMA_SCANOUT_BIT17, __i915_vma_flags(vma)((unsigned long *)&(vma)->flags));
393}
394 
395#define for_each_until(cond)if (cond) break; else if (cond) break; else
396 
397/**
398 * for_each_ggtt_vma - Iterate over the GGTT VMA belonging to an object.
399 * @V: the #i915_vma iterator
400 * @OBJ: the #drm_i915_gem_object
401 *
402 * GGTT VMA are placed at the being of the object's vma_list, see
403 * vma_create(), so we can stop our walk as soon as we see a ppgtt VMA,
404 * or the list is empty ofc.
405 */
406#define for_each_ggtt_vma(V, OBJ)for (V = ({ const __typeof( ((__typeof(*V) *)0)->obj_link )
 *__mptr = ((&(OBJ)->vma.list)->next); (__typeof(*V
) *)( (char *)__mptr - __builtin_offsetof(__typeof(*V), obj_link
) );}); &V->obj_link != (&(OBJ)->vma.list); V =
 ({ const __typeof( ((__typeof(*V) *)0)->obj_link ) *__mptr
 = (V->obj_link.next); (__typeof(*V) *)( (char *)__mptr - __builtin_offsetof
(__typeof(*V), obj_link) );})) if (!i915_vma_is_ggtt(V)) break
; else \
407	list_for_each_entry(V, &(OBJ)->vma.list, obj_link)for (V = ({ const __typeof( ((__typeof(*V) *)0)->obj_link )
 *__mptr = ((&(OBJ)->vma.list)->next); (__typeof(*V
) *)( (char *)__mptr - __builtin_offsetof(__typeof(*V), obj_link
) );}); &V->obj_link != (&(OBJ)->vma.list); V =
 ({ const __typeof( ((__typeof(*V) *)0)->obj_link ) *__mptr
 = (V->obj_link.next); (__typeof(*V) *)( (char *)__mptr - __builtin_offsetof
(__typeof(*V), obj_link) );}))		\
408		for_each_until(!i915_vma_is_ggtt(V))if (!i915_vma_is_ggtt(V)) break; else
409 
410struct i915_vma *i915_vma_make_unshrinkable(struct i915_vma *vma);
411void i915_vma_make_shrinkable(struct i915_vma *vma);
412void i915_vma_make_purgeable(struct i915_vma *vma);
413 
414int i915_vma_wait_for_bind(struct i915_vma *vma);
415 
416static inline int i915_vma_sync(struct i915_vma *vma)
417{
418	/* Wait for the asynchronous bindings and pending GPU reads */
419	return i915_active_wait(&vma->active);
420}
421 
422/**
423 * i915_vma_get_current_resource - Get the current resource of the vma
424 * @vma: The vma to get the current resource from.
425 *
426 * It's illegal to call this function if the vma is not bound.
427 *
428 * Return: A refcounted pointer to the current vma resource
429 * of the vma, assuming the vma is bound.
430 */
431static inline struct i915_vma_resource *
432i915_vma_get_current_resource(struct i915_vma *vma)
433{
434	return i915_vma_resource_get(vma->resource);
435}
436 
437#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)0
438void i915_vma_resource_init_from_vma(struct i915_vma_resource *vma_res,
439				     struct i915_vma *vma);
440#endif
441 
442void i915_vma_module_exit(void);
443int i915_vma_module_init(void);
444 
445I915_SELFTEST_DECLARE(int i915_vma_get_pages(struct i915_vma *vma));
446I915_SELFTEST_DECLARE(void i915_vma_put_pages(struct i915_vma *vma));
447 
448#endif