- 15 Dec, 2023 4 commits
-
-
Zhao Liu authored
The use of kmap_atomic() is being deprecated in favor of kmap_local_page()[1], and this patch converts the call from kmap_atomic() to kmap_local_page(). The main difference between atomic and local mappings is that local mappings doesn't disable page faults or preemption (the preemption is disabled for !PREEMPT_RT case, otherwise it only disables migration). With kmap_local_page(), we can avoid the often unwanted side effect of unnecessary page faults or preemption disables. In drm/i915/gem/selftests/huge_pages.c, function __cpu_check_shmem() mainly uses mapping to flush cache and check the value. There're 2 reasons why __cpu_check_shmem() doesn't need to disable pagefaults and preemption for mapping: 1. The flush operation is safe. Function __cpu_check_shmem() calls drm_clflush_virt_range() to use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86 and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush operation is global. 2. Any context switch caused by preemption or page faults (page fault may cause sleep) doesn't affect the validity of local mapping. Therefore, __cpu_check_shmem() is a function where the use of kmap_local_page() in place of kmap_atomic() is correctly suited. Convert the calls of kmap_atomic() / kunmap_atomic() to kmap_local_page() / kunmap_local(). [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.comSuggested-by:
Dave Hansen <dave.hansen@intel.com> Suggested-by:
Ira Weiny <ira.weiny@intel.com> Suggested-by:
Fabio M. De Francesco <fmdefrancesco@gmail.com> Signed-off-by:
Zhao Liu <zhao1.liu@intel.com> Reviewed-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20231203132947.2328805-5-zhao1.liu@linux.intel.com
-
Zhao Liu authored
The use of kmap_atomic() is being deprecated in favor of kmap_local_page()[1]. The main difference between atomic and local mappings is that local mappings doesn't disable page faults or preemption (the preemption is disabled for !PREEMPT_RT case, otherwise it only disables migration). With kmap_local_page(), we can avoid the often unwanted side effect of unnecessary page faults or preemption disables. In drm/i915/gem/i915_gem_shmem.c, the function shmem_pwrite() need to disable pagefault to eliminate the potential recursion fault[2]. But here __copy_from_user_inatomic() doesn't need to disable preemption and local mapping is valid for sched in/out. So it can use kmap_local_page() / kunmap_local() with pagefault_disable() / pagefault_enable() to replace atomic mapping. [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com [2]: https://patchwork.freedesktop.org/patch/295840/Suggested-by:
-
Zhao Liu authored
The use of kmap_atomic() is being deprecated in favor of kmap_local_page()[1], and this patch converts the call from kmap_atomic() + memcpy() to memcpy_[from/to]_page(), which use kmap_local_page() to build local mapping and then do memcpy(). The main difference between atomic and local mappings is that local mappings doesn't disable page faults or preemption (the preemption is disabled for !PREEMPT_RT case, otherwise it only disables migration). With kmap_local_page(), we can avoid the often unwanted side effect of unnecessary page faults and preemption disables. In drm/i915/gem/i915_gem_phys.c, the functions i915_gem_object_get_pages_phys() and i915_gem_object_put_pages_phys() don't need to disable pagefaults and preemption for mapping because of 2 reasons: 1. The flush operation is safe. In drm/i915/gem/i915_gem_object.c, i915_gem_object_get_pages_phys() and i915_gem_object_put_pages_phys() calls drm_clflush_virt_range() to use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86 and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush operation is global. 2. Any context switch caused by preemption or page faults (page fault may cause sleep) doesn't affect the validity of local mapping. Therefore, i915_gem_object_get_pages_phys() and i915_gem_object_put_pages_phys() are two functions where the uses of local mappings in place of atomic mappings are correctly suited. Convert the calls of kmap_atomic() / kunmap_atomic() + memcpy() to memcpy_from_page() and memcpy_to_page(). [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.comSuggested-by:
-
Zhao Liu authored
The use of kmap_atomic() is being deprecated in favor of kmap_local_page()[1], and this patch converts the call from kmap_atomic() to kmap_local_page(). The main difference between atomic and local mappings is that local mappings doesn't disable page faults or preemption (the preemption is disabled for !PREEMPT_RT case, otherwise it only disables migration). With kmap_local_page(), we can avoid the often unwanted side effect of unnecessary page faults and preemption disables. There're 2 reasons why i915_gem_object_read_from_page_kmap() doesn't need to disable pagefaults and preemption for mapping: 1. The flush operation is safe. In drm/i915/gem/i915_gem_object.c, i915_gem_object_read_from_page_kmap() calls drm_clflush_virt_range() to use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86 and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush operation is global. 2. Any context switch caused by preemption or page faults (page fault may cause sleep) doesn't affect the validity of local mapping. Therefore, i915_gem_object_read_from_page_kmap() is a function where the use of kmap_local_page() in place of kmap_atomic() is correctly suited. Convert the calls of kmap_atomic() / kunmap_atomic() to kmap_local_page() / kunmap_local(). And remove the redundant variable that stores the address of the mapped page since kunmap_local() can accept any pointer within the page. [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.comSuggested-by:
-
- 14 Dec, 2023 2 commits
-
-
Haridhar Kalvala authored
Enable Force Dispatch Ends Collection for DG2. BSpec: 46001 Signed-off-by:
Haridhar Kalvala <haridhar.kalvala@intel.com> Reviewed-by:
Matt Roper <matthew.d.roper@intel.com> Signed-off-by:
-
Randy Dunlap authored
Use "its" for possessive form instead of "it's". Hyphenate multi-word adjectives. Correct some spelling. End one line of code with ';' instead of ','. The before and after object files are identical. Signed-off-by:
Randy Dunlap <rdunlap@infradead.org> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: intel-gfx@lists.freedesktop.org Reviewed-by:
-
- 13 Dec, 2023 1 commit
-
-
Karthik Poosa authored
Updated i915 hwmon with fixes for issues reported by static analysis tool. Fixed integer overflow with upcasting. v2: - Added Fixes tag (Badal). - Updated commit message as per review comments (Anshuman). Fixes: 4c2572fe ("drm/i915/hwmon: Expose power1_max_interval") Reviewed-by:
Badal Nilawar <badal.nilawar@intel.com> Reviewed-by:
Anshuman Gupta <anshuman.gupta@intel.com> Signed-off-by:
Karthik Poosa <karthik.poosa@intel.com> Signed-off-by:
-
- 11 Dec, 2023 1 commit
-
-
Colin Ian King authored
There is a spelling mistake in a pr_err error message. Fix it. Signed-off-by:
Colin Ian King <colin.i.king@gmail.com> Signed-off-by:
Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20231209230541.4055786-1-colin.i.king@gmail.com
-
- 08 Dec, 2023 2 commits
-
-
Andi Shyti authored
Given a reference to "guc", the guc_to_i915() returns the pointer to "i915" private data. Signed-off-by:
Andi Shyti <andi.shyti@linux.intel.com> Reviewed-by:
Nirmoy Das <nirmoy.das@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20231206184322.57111-1-andi.shyti@linux.intel.com
-
Andrzej Hajda authored
After i915_active_unlock_wait i915_active can be still non-idle due to barrier async handling in signal_irq_work. As a result one can observe following errors: bcs0: heartbeat pulse did not flush idle tasks *ERROR* pulse active pulse_active [i915]:pulse_retire [i915] *ERROR* pulse count: 0 *ERROR* pulse preallocated barriers? no To prevent it let's wait explicitly for idleness. v2: wait only in live_idle tests Signed-off-by:
Andrzej Hajda <andrzej.hajda@intel.com> Reviewed-by:
-
- 07 Dec, 2023 2 commits
-
-
Tvrtko Ursulin authored
Commit 50357944 ("drm/i915/gsc: Mark internal GSC engine with reserved uabi class") made the GSC0 engine not have a valid uabi class and so broke the engine reset counting, which in turn was made class based in cb823ed9 ("drm/i915/gt: Use intel_gt as the primary object for handling resets"). Despite the title and commit text of the latter is not mentioning it (and has left the storage array incorrectly sized), tracking by class, despite it adding aliasing in hypthotetical multi-tile systems, is handy for virtual engines which for instance do not have a valid engine->id. Therefore we keep that but just change it to use the internal class which is always valid. We also add a helper to increment the count, which aligns with the existing getter. What was broken without this fix were out of bounds reads every time a reset would happen on the GSC0 engine, or during selftests when storing and cross-checking the counts in igt_live_test_begin and igt_live_test_end. Signed-off-by:
Alan Previn Teres Alexis <alan.previn.teres.alexis@intel.com> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Reviewed-by:
Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20231201122109.729006-2-tvrtko.ursulin@linux.intel.com
-
Tvrtko Ursulin authored
Engine->id namespace is per-tile so struct igt_live_test->reset_engine[] needs to be two-dimensional so engine reset counts from all tiles can be stored with no aliasing. With aliasing, if we had a real multi-tile platform, the reset counts would be incorrect for same engine instance on different tiles. Signed-off-by:
-
- 30 Nov, 2023 3 commits
-
-
John Harrison authored
There is a mechanism for reporting errors from fire and forget H2G messages. This is the only way to find out about almost any error in the GuC backend submission path. So it would be useful to know that it is working. v2: Fix some dumb over-complications and a couple of typos - review feedback from Daniele. Signed-off-by:
John Harrison <John.C.Harrison@Intel.com> Reviewed-by:
-
John Harrison authored
Noticed that the hangcheck selftest is submitting a non-preemptoble spinner. That means that even if the GuC does not die, the heartbeat will still kick in and trigger a reset. Which is rather defeating the purpose of the test - to verify that the heartbeat will kick in if the GuC itself has died. The test is deliberately killing the GuC, so it should never hit the case of a non-dead GuC. But it is not impossible that the kill might fail at some future point due to other driver re-work. So, make the spinner pre-emptible. That way the heartbeat can get through if the GuC is alive and context switching. Thus a reset only happens if the GuC dies. Thus, if the kill should stop working the test will now fail rather than claim to pass. Signed-off-by:
-
Zhanjun Dong authored
The gt wedged could be triggered by missing guc firmware file, HW not working, etc. Once triggered, it means all gt usage is dead, therefore we can't enable pxp under this fatal error condition. v2: Updated commit message. v3: Updated return code check. Signed-off-by:
Zhanjun Dong <zhanjun.dong@intel.com> Reviewed-by:
-
- 29 Nov, 2023 1 commit
-
-
Alan Previn authored
Debugging PXP issues can't even begin without understanding precedding sequence of important events. Add drm_dbg into the most important PXP events. v5 : - rebase. v4 : - rebase. v3 : - move gt_dbg to after mutex block in function i915_gsc_proxy_component_bind. (Vivaik) v2 : - remove __func__ since drm_dbg covers that (Jani). - add timeout dbg of the restart from front-end (Alan). Signed-off-by:Vivaik Balasubrawmanian <vivaik.balasubrawmanian@intel.com> Signed-off-by:
-
- 28 Nov, 2023 1 commit
-
-
Matt Roper authored
This workaround has been dropped from all DG2 variants in the latest workaround database update. Signed-off-by:
Gustavo Sousa <gustavo.sousa@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20231127190043.4099109-2-matthew.d.roper@intel.com
-
- 21 Nov, 2023 1 commit
-
-
Haridhar Kalvala authored
ATS-M device ID update. BSpec: 44477 Signed-off-by:
Lucas De Marchi <lucas.demarchi@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20231120113731.1570589-1-haridhar.kalvala@intel.com
-
- 20 Nov, 2023 5 commits
-
-
Andrzej Hajda authored
Track every intel_gt_pm_get() until its corresponding release in intel_gt_pm_put() by returning a cookie to the caller for acquire that must be passed by on released. When there is an imbalance, we can see who either tried to free a stale wakeref, or who forgot to free theirs. v2: track recently added calls in gen8_ggtt_bind_get_ce and destroyed_worker_func Signed-off-by: -
Andrzej Hajda authored
Beside reusing existing code, the main advantage of ref_tracker is tracking per instance of wakeref. It allows also to catch double put. On the other side we lose information about the first acquire and the last release, but the advantages outweigh it. Signed-off-by:
-
Tvrtko Ursulin authored
The GSC CS is not exposed to the user, so we skipped assigning a uabi class number for it. However, the trace logs use the uabi class and instance to identify the engine, so leaving uabi class unset makes the GSC CS show up as the RCS in those logs. Given that the engine is not exposed to the user, we can't add a new case in the uabi enum, so we insted internally define a kernel internal class as -1. At the same time remove special handling for the name and complete the uabi_classes array so internal class is automatically correctly assigned. Engine will show as 65535:0 other0 in the logs/traces which should be unique enough. v2: * Fix uabi class u8 vs u16 type confusion. Signed-off-by:
-
Tvrtko Ursulin authored
__rcu annotation is needed to avoid the sparse warnings such as: .../i915_drm_client.c:92:9: sparse: sparse: incompatible types in comparison expression (different address spaces): .../i915_drm_client.c:92:9: sparse: struct list_head [noderef] __rcu * .../i915_drm_client.c:92:9: sparse: struct list_head * Signed-off-by:
kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202311110610.h0m6ydI5-lkp@intel.com/ Cc: Andi Shyti <andi.shyti@linux.intel.com> Cc: Aravind Iddamsetty <aravind.iddamsetty@intel.com> Reviewed-by:
-
Tvrtko Ursulin authored
There is no need to return anything in the version which was merged and also the implementation of the !CONFIG_PROC_FS wasn't returning anything, causing a build failure there. Signed-off-by:
-
- 17 Nov, 2023 2 commits
-
-
Radhakrishna Sripada authored
Commit 78cc55e0 ("drm/i915/mcr: Hold GT forcewake during steering operations") introduced the workaround which was in early stages. With a valid lineage number update Workaround for future tracking. Cc: Matt Roper <matthew.d.roper@intel.com> Signed-off-by:
Radhakrishna Sripada <radhakrishna.sripada@intel.com> Reviewed-by:
-
Matt Roper authored
The workaround database was just updated to extend this workaround to DG2-G11 (whereas previously it applied only to G10 and G12). Signed-off-by:
-
- 16 Nov, 2023 4 commits
-
-
Daniele Ceraolo Spurio authored
On MTL, the HuC is only supported on the media GT, so our validation check on the module parameter detects an inconsistency on the root GT (the modparams asks to enable HuC, but the support is not there) and prints the following info message: [drm] GT0: Incompatible option enable_guc=3 - HuC is not supported! This can be confusing to the user and make them think that something is wrong when it isn't, so we need to silence it. Given that any platform that supports HuC also supports GuC, if a user tries to enable HuC on a platform that really doesn't support it they'll already see a message about GuC not being supported, so instead of just silencing the HuC message on newer platforms we can just get rid of it entirely. Signed-off-by:
-
Andrzej Hajda authored
The only task of intel_gt_release_all is to zero gt table. Calling it on error path prevents intel_gt_driver_late_release_all (called from i915_driver_late_release) to cleanup GTs, causing leakage. After i915_driver_late_release GT array is not used anymore so it does not need cleaning at all. Sample leak report: BUG i915_request (...): Objects remaining in i915_request on __kmem_cache_shutdown() ... Object 0xffff888113420040 @offset=64 Allocated in __i915_request_create+0x75/0x610 [i915] age=18339 cpu=1 pid=1454 kmem_cache_alloc+0x25b/0x270 __i915_request_create+0x75/0x610 [i915] i915_request_create+0x109/0x290 [i915] __engines_record_defaults+0xca/0x440 [i915] intel_gt_init+0x275/0x430 [i915] i915_gem_init+0x135/0x2c0 [i915] i915_driver_probe+0x8d1/0xdc0 [i915] v2: removed whole intel_gt_release_all Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/8489 Fixes: bec68cc9 ("drm/i915: Prepare for multiple GTs") Signed-off-by:
-
Andrzej Hajda authored
Trace requires new-line at the end of message (in opposition to printk), otherwise trace dump becomes messy. Signed-off-by:
Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com> Reviewed-by:
-
Vinay Belgaumkar authored
We read RENDER_HEAD as a part of the flush. If GT is in deeper sleep states, this could lead to read errors since we are not using a forcewake. Safer to read a shadowed register instead. Cc: John Harrison <john.c.harrison@intel.com> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Signed-off-by:
Vinay Belgaumkar <vinay.belgaumkar@intel.com> Reviewed-by:
-
- 10 Nov, 2023 6 commits
-
-
Tvrtko Ursulin authored
Use the newly added drm_print_memory_stats helper to show memory utilisation of our objects in drm/driver specific fdinfo output. To collect the stats we walk the per memory regions object lists and accumulate object size into the respective drm_memory_stats categories. v2: * Only account against the active region. * Use DMA_RESV_USAGE_BOOKKEEP when testing for active. (Tejas) v3: * Update commit text. (Aravind) * Update to use memory regions uabi names. Signed-off-by:
Aravind Iddamsetty <aravind.iddamsetty@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20231107101806.608990-6-tvrtko.ursulin@linux.intel.com
-
Tvrtko Ursulin authored
At the moment memory region names are a bit too varied and too inconsistent to be used for ABI purposes, like for upcoming fdinfo memory stats. System memory can be either system or system-ttm. Local memory has the instance number appended, others do not. Not only incosistent but thi kind of implementation detail is uninteresting for intended users of fdinfo memory stats. Add a stable name always formed as $type$instance. Could have chosen a different stable scheme, but I think any consistent and stable scheme should do just fine. Signed-off-by:
-
Tvrtko Ursulin authored
Account ring buffers and logical context space against the owning client memory usage stats. Signed-off-by:
-
Tvrtko Ursulin authored
Account page table backing store against the owning client memory usage stats. Signed-off-by:
-
Tvrtko Ursulin authored
To enable accounting of indirect client memory usage (such as page tables) in the following patch, lets start recording the creator of each PPGTT. Signed-off-by:
-
Tvrtko Ursulin authored
In order to show per client memory usage lets add some infrastructure which enables tracking buffer objects owned by clients. We add a per client list protected by a new per client lock and to support delayed destruction (post client exit) we make tracked objects hold references to the owning client. Also, object memory region teardown is moved to the existing RCU free callback to allow safe dereference from the fdinfo RCU read section. Signed-off-by:
-
- 09 Nov, 2023 1 commit
-
-
Sam James authored
GCC 14 introduces a new -Walloc-size included in -Wextra which errors out like: ``` drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c: In function ‘eb_copy_relocations’: drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c:1681:24: error: allocation of insufficient size ‘1’ for type ‘struct drm_i915_gem_relocation_entry’ with size ‘32’ [-Werror=alloc-size] 1681 | relocs = kvmalloc_array(size, 1, GFP_KERNEL); | ^ ``` So, just swap the number of members and size arguments to match the prototype, as we're initialising 1 element of size `size`. GCC then sees we're not doing anything wrong. Signed-off-by:Sam James <sam@gentoo.org> Reviewed-by:
-
- 07 Nov, 2023 1 commit
-
-
Gustavo Sousa authored
This workaround applies to all steppings of Xe_LPM+. Implement the KMD part. v2: - Put the definition of VDBOX_CGCTL3F1C() in the correct sort order. (Matt) Reviewed-by:
-
- 06 Nov, 2023 2 commits
-
-
Jonathan Cavitt authored
FIXME: It is suspected that some Address Translation Service (ATS) issue on IOMMU is causing CAT errors to occur on some MTL workloads. Applying a write barrier to the ppgtt set entry functions appeared to have no effect, so we must temporarily use I915_MAP_WC in the map_pt_dma class of functions on MTL until a proper ATS solution is found. Signed-off-by:
Jonathan Cavitt <jonathan.cavitt@intel.com> CC: Chris Wilson <chris.p.wilson@linux.intel.com> Reviewed-by:
-
Kunwu Chan authored
Fix smatch warning: drivers/gpu/drm/i915/gem/i915_gem_context.c:847 set_proto_ctx_sseu() warn: potential spectre issue 'pc->user_engines' [r] (local cap) Fixes: d4433c76 ("drm/i915/gem: Use the proto-context to handle create parameters (v5)") Cc: <stable@vger.kernel.org> # v5.15+ Signed-off-by:
Kunwu Chan <chentao@kylinos.cn> Reviewed-by:
Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Signed-off-by:
-
- 03 Nov, 2023 1 commit
-
-
Dorcas AnonoLitunya authored
Removes space after opening parenthesis. Fixes the checkpatch.pl error: ERROR: space prohibited after that opening parenthesis '(' Signed-off-by:Dorcas AnonoLitunya <anonolitunya@gmail.com> Reviewed-by:
-
