- 12 Nov, 2021 1 commit
-
-
git://anongit.freedesktop.org/drm/drm-miscDave Airlie authored
* dma-buf: name_lock fixes * prime: Keep object ref during mmap * nouveau: Fix a refcount issue; Fix device removal; Protect client list with dedicated mutex; Fix address CE0 address calculation * ttm: Fix race condition during BO eviction Signed-off-by:
Dave Airlie <airlied@redhat.com> From: Thomas Zimmermann <tzimmermann@suse.de> Link: https://patchwork.freedesktop.org/patch/msgid/YYzY6jeox9EeI15i@linux-uq9g.fritz.box
-
- 11 Nov, 2021 3 commits
-
-
Dave Airlie authored
I got a drm-fixes which had some 5.15 stuff in it, so to avoid the mess just backmerge here. Linux 5.15 Signed-off-by: -
xinhui pan authored
BO might sit in a wrong lru list as there is a small period of memory moving and lru list updating. Lets skip eviction if we hit such mismatch. Suggested-by:
Christian König <christian.koenig@amd.com> Signed-off-by:
xinhui pan <xinhui.pan@amd.com> Reviewed-by:
-
Dave Airlie authored
Merge tag 'amd-drm-fixes-5.16-2021-11-10' of https://gitlab.freedesktop.org/agd5f/linux into drm-next amd-drm-fixes-5.16-2021-11-10: amdgpu: - Don't allow partial copy from user for DC debugfs - SRIOV fixes - GFX9 CSB pin count fix - Various IP version check fixes - DP 2.0 fixes - Limit DCN1 MPO fix to DCN1 amdkfd: - SVM fixes - Reset fixes Signed-off-by:
-
- 10 Nov, 2021 6 commits
-
-
Dave Airlie authored
Merge tag 'drm-misc-next-fixes-2021-11-10' of git://anongit.freedesktop.org/drm/drm-misc into drm-next Removed the TTM Huge Page functionnality to address a crash, a timeout fix for udl, CONFIG_FB dependency improvements, a fix for a circular locking depency in imx, a NULL pointer dereference fix for virtio, and a naming collision fix for drm/locking. Signed-off-by:
-
Guchun Chen authored
Fixes: 96b8dd44 ("drm/amdgpu/amdgpu_vcn: convert to IP version checking") Signed-off-by:
Flora Cui <flora.cui@amd.com> Signed-off-by:
Guchun Chen <guchun.chen@amd.com> Reviewed-by:
Alex Deucher <alexander.deucher@amd.com> Signed-off-by:
-
Guchun Chen authored
Fixes: b05b9c59 ("drm/amdgpu: clean up set IP function") Signed-off-by:
-
Shirish S authored
[Why] Video plane gets rejected for non-zero src_y and src_x on DCN2.x. [How] Limit the rejection till DCN1.x and verified MPO, by dragging video playback beyond display's left (0, 0) co-ordinates. Fixes: d89f6048 ("drm/amd/display: Reject non-zero src_y and src_x for video planes") Signed-off-by:
Shirish S <shirish.s@amd.com> Reviewed-by:
Harry Wentland <harry.wentland@amd.com> Signed-off-by:
-
Dave Airlie authored
Merge tag 'drm-intel-next-fixes-2021-11-09' of git://anongit.freedesktop.org/drm/drm-intel into drm-next Couple Reverts, build fix, couple virtualization fixes, blank screen and other display rates fixes, and more. Four patches targeting stable in here. Display Fixes: - DP rates related fixes (Imre, Jani) - A Revert on disaling dual eDP that was causing state readout problems (Jani) - put the cdclk vtables in const data (Jani) - Fix DVO port type for moder platforms (Ville) - Fix blankscreen by turning DP++ TMDS output buffers on encoder->shutdown (Ville) - CCS FBs related fixes (Imre) GT fixes: - Fix recursive lock in GuC submission (Matt Brost) - Revert guc_id from i915_request tracepoint (Joonas) - Build fix around dmabuf (Matt Auld) Signed-off-by:
-
Dave Airlie authored
Merge tag 'drm-misc-next-fixes-2021-11-05' of git://anongit.freedesktop.org/drm/drm-misc into drm-next A refcounting fix for outstanding fence callbacks. Signed-off-by:
-
- 09 Nov, 2021 12 commits
-
-
Nicholas Kazlauskas authored
[Why] We need HPD IRQ notifications (RX, short pulse) to properly handle DP MST for DPIA connections. [How] A null pointer exception currently occurs when these are received so add a check to validate that we have a handler installed for the notification. Extend the HPD handler to also handle HPD IRQ (RX) since the logic is the same. Fixes: e27c41d5 ("drm/amd/display: Support for DMUB HPD interrupt handling") Reviewed-by:
Wayne Lin <Wayne.Lin@amd.com> Reviewed-by:
Jude Shih <shenshih@amd.com> Acked-by:
Anson Jacob <Anson.Jacob@amd.com> Tested-by:
Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by:
Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Signed-off-by:
-
Nicholas Kazlauskas authored
[Why] Per DRM spec we only need to hold that lock when touching connector->state - which we do not do in that handler. Taking this locking introduces unnecessary dependencies with other threads which is bad for performance and opens up the potential for a deadlock since there are multiple locks being held at once. [How] Remove the connection_mutex lock/unlock routine and just iterate over the drm connectors normally. The iter helpers implicitly lock the connection list so this is safe to do. DC link access also does not need to be guarded since the link table is static at creation - we don't dynamically add or remove links, just streams. Fixes: e27c41d5 ("drm/amd/display: Support for DMUB HPD interrupt handling") Reviewed-by:
-
Anson Jacob authored
Trivial patch which adds a comment for macro endif's in amdgpu_dm.c Reviewed-by:
Ariel Bernstein <Eric.Bernstein@amd.com> Reviewed-by:
-
Felix Kuehling authored
The check for whether to drain retry faults must be under the mmap write lock to serialize with munmap notifier callbacks. We were also missing checks on child ranges. To fix that, simplify the logic by using a flag rather than checking on each prange. That also allows draining less freqeuntly when many ranges are unmapped at once. Signed-off-by:
Felix Kuehling <Felix.Kuehling@amd.com> Tested-by:
Philip Yang <Philip.Yang@amd.com> Tested-by:
Alex Sierra <Alex.Sierra@amd.com> Reviewed-by:
-
Alex Sierra authored
The low 16MB of virtual address space are currently reserved for kernel mode allocations mapped into user virtual address space. This causes conflicts with HMM/SVM mappings at low virtual addresses. We tried to move those kernel mode allocations to the upper half of the 64-bit virtual address space for GFX9, which is naturally reserved for kernel use. However, TBA (trap handler code) has problems to access addresses in the high virtual space. We have decided to set this to 8KB of the lower address space as a temporary fix, while investigate TBA address problem. It is very unlikely for user space to map memory at this low region. Signed-off-by:
-
Shirish S authored
make action upon failure in "drm_atomic_add_affected_connectors()" consistent with the rest of failures in amdgpu_dm_atomic_check(). Signed-off-by:
-
shaoyunl authored
The KFD pre_reset should be called before reset been executed, it will hold the lock to prevent other rocm process to sent the packlage to hiq during host execute the real reset on the HW Signed-off-by:
shaoyunl <shaoyun.liu@amd.com> Reviewed-by:
-
Evan Quan authored
There was a change(below) target for such issue: d82e2c24 ("drm/amdgpu: Fix crash on device remove/driver unload") But the fix for VI ASICs was missing there. This is a supplement for that. Fixes: d82e2c24 ("drm/amdgpu: Fix crash on device remove/driver unload") Signed-off-by:
Evan Quan <evan.quan@amd.com> Acked-by:
-
Imre Deak authored
So far the remapped view size in GTT/DPT was padded to the next aligned offset unnecessarily after the last color plane with an unaligned size. Remove the unnecessary padding. Cc: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com> Fixes: 3d1adc3d ("drm/i915/adlp: Add support for remapping CCS FBs") Signed-off-by:
Imre Deak <imre.deak@intel.com> Reviewed-by:
Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com> Link: https://patchwork.freedesktop.org/patch/msgid/20211026225105.2783797-3-imre.deak@intel.com (cherry picked from commit 6b6636e1) Signed-off-by:
Rodrigo Vivi <rodrigo.vivi@intel.com>
-
Imre Deak authored
For NV12 FBs with odd main surface tile-row height the CCS surface height was incorrectly calculated 1 less than the actual value. Fix this by rounding up the result of divison. For consistency do the same for the CCS surface width calculation. Fixes: b3e57bcc ("drm/i915/tgl: Gen-12 render decompression") Signed-off-by:
-
Ville Syrjälä authored
Looks like our VBIOS/GOP generally fail to turn the DP dual mode adater TMDS output buffers back on after a reboot. This leads to a black screen after reboot if we turned the TMDS output buffers off prior to reboot. And if i915 decides to do a fastboot the black screen will persist even after i915 takes over. Apparently this has been a problem ever since commit b2ccb822 ("drm/i915: Enable/disable TMDS output buffers in DP++ adaptor as needed") if one rebooted while the display was turned off. And things became worse with commit fe0f1e3b ("drm/i915: Shut down displays gracefully on reboot") since now we always turn the display off before a reboot. This was reported on a RKL, but I confirmed the same behaviour on my SNB as well. So looks pretty universal. Let's fix this by explicitly turning the TMDS output buffers back on in the encoder->shutdown() hook. Note that this gets called after irqs have been disabled, so the i2c communication with the DP dual mode adapter has to be performed via polling (which the gmbus code is perfectly happy to do for us). We also need a bit of care in handling DDI encoders which may or may not be set up for HDMI output. Specifically ddc_pin will not be populated for a DP only DDI encoder, in which case we don't want to call intel_gmbus_get_adapter(). We can handle that by simply doing the dual mode adapter type check before calling intel_gmbus_get_adapter(). Cc: <stable@vger.kernel.org> # v5.11+ Fixes: fe0f1e3b ("drm/i915: Shut down displays gracefully on reboot") Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/4371Signed-off-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20211029191802.18448-2-ville.syrjala@linux.intel.comReviewed-by:
Stanislav Lisovskiy <stanislav.lisovskiy@intel.com> (cherry picked from commit 49c55f7b) Signed-off-by:
-
Stephen Rothwell authored
Commit cd06ab2f ("drm/locking: add backtrace for locking contended locks without backoff") added functions named __stack_depot_* in drm which conflict with stack depot. Rename to __drm_stack_depot_*. v2 by Jani: - Also rename __stack_depot_print References: https://lore.kernel.org/r/20211015202648.258445ef@canb.auug.org.au Fixes: cd06ab2f ("drm/locking: add backtrace for locking contended locks without backoff") Cc: Daniel Vetter <daniel@ffwll.ch> Reviewed-by:
Daniel Vetter <daniel@ffwll.ch> Signed-off-by:
Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by:
Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20211018085113.27033-1-jani.nikula@intel.com (cherry picked from commit c4f08d72)
-
- 08 Nov, 2021 1 commit
-
-
Vivek Kasireddy authored
When virgl is not enabled, vfpriv pointer would not be allocated. Therefore, check for a valid value before dereferencing. Reported-by:
Christian Zigotzky <chzigotzky@xenosoft.de> Cc: Gurchetan Singh <gurchetansingh@chromium.org> Cc: Gerd Hoffmann <kraxel@redhat.com> Signed-off-by:
Vivek Kasireddy <vivek.kasireddy@intel.com> Tested-by:
Gerd Hoffmann <kraxel@redhat.com>
-
- 05 Nov, 2021 17 commits
-
-
Alex Deucher authored
Properly handle SI DC support when CONFIG_DRM_AMD_DC_SI is not set. Fixes: f7f12b25 ("drm/amdgpu: default to true in amdgpu_device_asic_has_dc_support") Reviewed-by:
-
Felix Kuehling authored
If a kfd_bo was shared (e.g. a dmabuf export), the original kfd_bo may be freed when the amdgpu_bo still lives on. Free the kfd_bo struct in the release_notify callback then the amdgpu_bo is freed. Signed-off-by:
Ramesh Errabolu <Ramesh.Errabolu@amd.com> Reviewed-by:
-
shaoyunl authored
When kfd need to be reset, sent command to HWS might cause hang and get unnecessary timeout. This change try not to touch HW in pre_reset and keep queues to be in the evicted state when the reset is done, so they are not put back on the runlist. These queues will be destroied on process termination. Signed-off-by:
-
Evan Quan authored
As part of the ib padding process, accessing the RLC_SPM_* register may trigger gfx hang. Since gfxoff may be already kicked during the whole period. To address that, we manually toggle gfx on/off around the RLC_SPM_* register access. This can resolve the gfx hang issue observed on running Talos with RDP launched in parallel. Signed-off-by:
-
Tao Zhou authored
The error count reset for xgmi3x16 pcs is missed. Signed-off-by:
Tao Zhou <tao.zhou1@amd.com> Reviewed-by:
Hawking Zhang <Hawking.Zhang@amd.com> Signed-off-by:
-
Mario Limonciello authored
Previously there was a check based on chip # for chips that aligned to >=CHIP_NAVI10 to have RLC stopped as part of DPMS check. This was because of gfxclk being controlled by RLC in the newer designs. As part of IP version checking though, this got changed to match IP version for SMU. Because Renoir designs also include smu11 that meant that even GFX9 started to stop RLC earlier. Adjust to match GFX IP version instead of SMU IP version to restore the previous behavior. Fixes: a8967967 ("drm/amdgpu/amdgpu_smu: convert to IP version checking") Signed-off-by:
Mario Limonciello <mario.limonciello@amd.com> Reviewed-by:
-
YuBiao Wang authored
[Why] csb bo is not unpinned in gfx 9. It will lead to pin_count leak on driver unload. [How] Call bo_free_kernel corresponding to bo_create_kernel in gfx_rlc_init_csb. This will also unify the code path with other gfx versions. Signed-off-by:
YuBiao Wang <YuBiao.Wang@amd.com> Reviewed-by:
-
YuBiao Wang authored
[Why] For Vega10, disabling gart of gfxhub could mess up KIQ and PSP under sriov mode, and lead to DMAR on host side. [How] Skip writing GMC registers under sriov. Signed-off-by:
-
Alex Deucher authored
sysfs_emit and sysfs_emit_at requrie a page boundary aligned buf address. Make them happy! v2: fix sysfs_emit -> sysfs_emit_at missed conversions Cc: Lang Yu <lang.yu@amd.com> Cc: Darren Powell <darren.powell@amd.com> Fixes: 6db0c87a ("amdgpu/pm: Replace hwmgr smu usage of sprintf with sysfs_emit") Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1774Reviewed-by:
Lang Yu <lang.yu@amd.com> Signed-off-by:
-
Kent Russell authored
BOs need to be reserved before they are added or removed, so ensure that they are reserved during kfd_mem_attach and kfd_mem_detach Signed-off-by:
Kent Russell <kent.russell@amd.com> Reviewed-by:
-
Alex Sierra authored
[Why]: When we call hmm_range_fault to map memory after a migration, we don't expect memory to be migrated again as a result of hmm_range_fault. The driver ensures that all memory is in GPU-accessible locations so that no migration should be needed. However, there is one corner case where hmm_range_fault can unexpectedly cause a migration from DEVICE_PRIVATE back to system memory due to a write-fault when a system memory page in the same range was mapped read-only (e.g. COW). Ranges with individual pages in different locations are usually the result of failed page migrations (e.g. page lock contention). The unexpected migration back to system memory causes a deadlock from recursive locking in our driver. [How]: Creating a task reference new member under svm_range_list struct. Setting this with "current" reference, right before the hmm_range_fault is called. This member is checked against "current" reference at svm_migrate_to_ram callback function. If equal, the migration will be ignored. Signed-off-by:
-
Harry Wentland authored
There is no reason to allow for partial buffers from userspace in our debugfs. In this particular case callers will zero out the wr_buf but if callers in the future don't do that we might be looking at corrupt data. Linus puts it better than I can in https://lkml.org/lkml/2021/10/26/993Reviewed-by:
Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by:
-
Fabio Estevam authored
This reverts commit f4b34faa. Since commit f4b34faa ("drm/imx: Annotate dma-fence critical section in commit path") the following possible circular dependency is detected: [ 5.001811] ====================================================== [ 5.001817] WARNING: possible circular locking dependency detected [ 5.001824] 5.14.9-01225-g45da36cc6fcc-dirty #1 Tainted: G W [ 5.001833] ------------------------------------------------------ [ 5.001838] kworker/u8:0/7 is trying to acquire lock: [ 5.001848] c1752080 (regulator_list_mutex){+.+.}-{3:3}, at: regulator_lock_dependent+0x40/0x294 [ 5.001903] [ 5.001903] but task is already holding lock: [ 5.001909] c176df78 (dma_fence_map){++++}-{0:0}, at: imx_drm_atomic_commit_tail+0x10/0x160 [ 5.001957] [ 5.001957] which lock already depends on the new lock. ... Revert it for now. Tested on a imx6q-sabresd. Fixes: f4b34faa ("drm/imx: Annotate dma-fence critical section in commit path") Signed-off-by:
Fabio Estevam <festevam@gmail.com> Signed-off-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20211104001112.4035691-1-festevam@gmail.com
-
Arnd Bergmann authored
My previous patch correctly addressed the possible link failure, but as Jani points out, the dependency is now stricter than it needs to be. Change it again, to allow DRM_FBDEV_EMULATION to be used when DRM_KMS_HELPER and FB are both loadable modules and DRM is linked into the kernel. As a side-effect, the option is now only visible when at least one DRM driver makes use of DRM_KMS_HELPER. This is better, because the option has no effect otherwise. Fixes: 606b1028 ("drm: fb_helper: fix CONFIG_FB dependency") Suggested-by:
Javier Martinez Canillas <javierm@redhat.com> Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Daniel Vetter authored
Somehow we only have a list of subdirectories, which apparently made it harder for folks to find the gpu maintainers. Fix that. References: https://lore.kernel.org/dri-devel/YXrAAZlxxStNFG%2FK@phenom.ffwll.local/Signed-off-by:
Daniel Vetter <daniel.vetter@intel.com> Reviewed-by:
-
Johan Hovold authored
USB control-message timeouts are specified in milliseconds and should specifically not vary with CONFIG_HZ. Fixes: 5320918b ("drm/udl: initial UDL driver (v4)") Cc: stable@vger.kernel.org # 3.4 Signed-off-by:
Johan Hovold <johan@kernel.org> Signed-off-by:
-
Jason Gunthorpe authored
The huge page functionality in TTM does not work safely because PUD and PMD entries do not have a special bit. get_user_pages_fast() considers any page that passed pmd_huge() as usable: if (unlikely(pmd_trans_huge(pmd) || pmd_huge(pmd) || pmd_devmap(pmd))) { And vmf_insert_pfn_pmd_prot() unconditionally sets entry = pmd_mkhuge(pfn_t_pmd(pfn, prot)); eg on x86 the page will be _PAGE_PRESENT | PAGE_PSE. As such gup_huge_pmd() will try to deref a struct page: head = try_grab_compound_head(pmd_page(orig), refs, flags); and thus crash. Thomas further notices that the drivers are not expecting the struct page to be used by anything - in particular the refcount incr above will cause them to malfunction. Thus everything about this is not able to fully work correctly considering GUP_fast. Delete it entirely. It can return someday along with a proper PMD/PUD_SPECIAL bit in the page table itself to gate GUP_fast. Fixes: 314b6580 ("drm/ttm, drm/vmwgfx: Support huge TTM pagefaults") Signed-off-by:Jason Gunthorpe <jgg@nvidia.com> Reviewed-by:
Thomas Hellström <thomas.helllstrom@linux.intel.com> Reviewed-by:
-
