kernel/linux.git/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c, branch v6.6.149

drm/amdgpu: fix aperture mapping leak

2026-08-03T09:15:32+00:00

commit ea772a440d56b285f4d491affac50ecd41f6b402 upstream. amdgpu_pci_remove() calls drm_dev_unplug() before invoking the driver fini routines. This causes drm_dev_enter() in amdgpu_ttm_fini() to always return false, so iounmap(aper_base_kaddr) never runs on normal driver unload, leaving an orphaned entry in the x86 PAT interval tree. On connected_to_cpu hardware, the aperture is mapped write-back (WB) via ioremap_cache(). On reload, IP discovery calls memremap(..., MEMREMAP_WC) over the same range. The WC vs WB conflict causes: ioremap error for 0x..., requested 0x1, got 0x0 amdgpu: discovery failed: -2 Fix by switching to devres-managed mappings so cleanup is guaranteed regardless of drm_dev_enter() state: - connected_to_cpu path: devm_memremap(MEMREMAP_WB). For IORESOURCE_SYSTEM_RAM ranges this takes the try_ram_remap() shortcut, returning __va(offset) from the existing kernel direct map. No new ioremap VA or PAT entry is created, so there is nothing to orphan. - dGPU path: devm_ioremap_wc() registers iounmap() as a devres action, guaranteeing cleanup at device_del() time. Also remove iounmap(aper_base_kaddr) from amdgpu_device_unmap_mmio() since the mapping is now devres-owned. v2: Remove redundant x86_64 guard (Lijo) Fixes: 9d0af8b4def0 ("drm/amdgpu: pre-map device buffer as cached for A+A config") Signed-off-by: Asad Kamal Reviewed-by: Christian König Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher (cherry picked from commit d871e99879cb5fd1fa798b006b4888887e63a17a) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: Fix amdgpu_bo_move() when old_mem and new_mem are both GTT

2026-08-03T09:15:23+00:00

commit ee94a65f192c05c543b4d3ad7137cd696b5c18fc upstream. The UVD code relies on GTT to GTT moves in order to ensure that its BOs don't cross 256M segments. Fixes: bfe5e585b44f ("drm/ttm: move last binding into the drivers.") Signed-off-by: Timur Kristóf Reviewed-by: Christian König Signed-off-by: Alex Deucher (cherry picked from commit 21fd45e5e2628d00b478590bcc3d14d3de5d45b6) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: fix zero-size GDS range init on RDNA4

2026-05-17T15:13:34+00:00

commit 095a8b0ad3c3b5cdc3850d961adb8a8f735220bb upstream. RDNA4 (GFX 12) hardware removes the GDS, GWS, and OA on-chip memory resources. The gfx_v12_0 initialisation code correctly leaves adev->gds.gds_size, adev->gds.gws_size, and adev->gds.oa_size at zero to reflect this. amdgpu_ttm_init() unconditionally calls amdgpu_ttm_init_on_chip() for each of these resources regardless of size. When the size is zero, amdgpu_ttm_init_on_chip() forwards the call to ttm_range_man_init(), which calls drm_mm_init(mm, 0, 0). drm_mm_init() immediately fires DRM_MM_BUG_ON(start + size <= start) -- trivially true when size is zero -- crashing the kernel during modprobe of amdgpu on an RX 9070 XT. Guard against this by returning 0 early from amdgpu_ttm_init_on_chip() when size_in_page is zero. This skips TTM resource manager registration for hardware resources that are absent, without affecting any other GPU type. DRM_MM_BUG_ON() only asserts if CONFIG_DRM_DEBUG_MM is enabled in the kernel config. This is apparently rarely enabled as these chips have been in the market for over a year and this issue was only reported now. Link: https://lore.kernel.org/all/bug-221376-2300@https.bugzilla.kernel.org%2F/ Link: https://bugzilla.kernel.org/show_bug.cgi?id=221376 Oops-Analysis: http://oops.fenrus.org/reports/bugzilla.korg/221376/report.html Assisted-by: GitHub Copilot:Claude Sonnet 4.6 linux-kernel-oops-x86. Signed-off-by: Arjan van de Ven Cc: Alex Deucher Cc: "Christian König" Cc: amd-gfx@lists.freedesktop.org Cc: dri-devel@lists.freedesktop.org Cc: linux-kernel@vger.kernel.org Signed-off-by: Alex Deucher (cherry picked from commit 5719ce5865279cad4fd5f01011fe037168503f2d) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: add missing lock to amdgpu_ttm_access_memory_sdma

2026-01-11T14:22:19+00:00

commit 4fa944255be521b1bbd9780383f77206303a3a5c upstream. Users of ttm entities need to hold the gtt_window_lock before using them to guarantee proper ordering of jobs. Cc: stable@vger.kernel.org Fixes: cb5cc4f573e1 ("drm/amdgpu: improve debug VRAM access performance using sdma") Signed-off-by: Pierre-Eric Pelloux-Prayer Reviewed-by: Christian König Signed-off-by: Alex Deucher Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: tear down ttm range manager for doorbell in amdgpu_ttm_fini()

2025-02-08T08:51:48+00:00

[ Upstream commit 60a2c0c12b644450e420ffc42291d1eb248bacb7 ] Tear down ttm range manager for doorbell in function amdgpu_ttm_fini(), to avoid memory leakage. Fixes: 792b84fb9038 ("drm/amdgpu: initialize ttm for doorbells") Signed-off-by: Jiang Liu Signed-off-by: Kent Russell Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Signed-off-by: Sasha Levin

drm/amdgpu: set the right AMDGPU sg segment limitation

2024-12-14T19:00:08+00:00

[ Upstream commit e2e97435783979124ba92d6870415c57ecfef6a5 ] The driver needs to set the correct max_segment_size; otherwise debug_dma_map_sg() will complain about the over-mapping of the AMDGPU sg length as following: WARNING: CPU: 6 PID: 1964 at kernel/dma/debug.c:1178 debug_dma_map_sg+0x2dc/0x370 [ 364.049444] Modules linked in: veth amdgpu(OE) amdxcp drm_exec gpu_sched drm_buddy drm_ttm_helper ttm(OE) drm_suballoc_helper drm_display_helper drm_kms_helper i2c_algo_bit rpcsec_gss_krb5 auth_rpcgss nfsv4 nfs lockd grace netfs xt_conntrack xt_MASQUERADE nf_conntrack_netlink xfrm_user xfrm_algo iptable_nat xt_addrtype iptable_filter br_netfilter nvme_fabrics overlay nfnetlink_cttimeout nfnetlink openvswitch nsh nf_conncount nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 libcrc32c bridge stp llc amd_atl intel_rapl_msr intel_rapl_common sunrpc sch_fq_codel snd_hda_codec_realtek snd_hda_codec_generic snd_hda_scodec_component snd_hda_codec_hdmi snd_hda_intel snd_intel_dspcfg edac_mce_amd binfmt_misc snd_hda_codec snd_pci_acp6x snd_hda_core snd_acp_config snd_hwdep snd_soc_acpi kvm_amd snd_pcm kvm snd_seq_midi snd_seq_midi_event crct10dif_pclmul ghash_clmulni_intel sha512_ssse3 snd_rawmidi sha256_ssse3 sha1_ssse3 aesni_intel snd_seq nls_iso8859_1 crypto_simd snd_seq_device cryptd snd_timer rapl input_leds snd [ 364.049532] ipmi_devintf wmi_bmof ccp serio_raw k10temp sp5100_tco soundcore ipmi_msghandler cm32181 industrialio mac_hid msr parport_pc ppdev lp parport drm efi_pstore ip_tables x_tables pci_stub crc32_pclmul nvme ahci libahci i2c_piix4 r8169 nvme_core i2c_designware_pci realtek i2c_ccgx_ucsi video wmi hid_generic cdc_ether usbnet usbhid hid r8152 mii [ 364.049576] CPU: 6 PID: 1964 Comm: rocminfo Tainted: G OE 6.10.0-custom #492 [ 364.049579] Hardware name: AMD Majolica-RN/Majolica-RN, BIOS RMJ1009A 06/13/2021 [ 364.049582] RIP: 0010:debug_dma_map_sg+0x2dc/0x370 [ 364.049585] Code: 89 4d b8 e8 36 b1 86 00 8b 4d b8 48 8b 55 b0 44 8b 45 a8 4c 8b 4d a0 48 89 c6 48 c7 c7 00 4b 74 bc 4c 89 4d b8 e8 b4 73 f3 ff <0f> 0b 4c 8b 4d b8 8b 15 c8 2c b8 01 85 d2 0f 85 ee fd ff ff 8b 05 [ 364.049588] RSP: 0018:ffff9ca600b57ac0 EFLAGS: 00010286 [ 364.049590] RAX: 0000000000000000 RBX: ffff88b7c132b0c8 RCX: 0000000000000027 [ 364.049592] RDX: ffff88bb0f521688 RSI: 0000000000000001 RDI: ffff88bb0f521680 [ 364.049594] RBP: ffff9ca600b57b20 R08: 000000000000006f R09: ffff9ca600b57930 [ 364.049596] R10: ffff9ca600b57928 R11: ffffffffbcb46328 R12: 0000000000000000 [ 364.049597] R13: 0000000000000001 R14: ffff88b7c19c0700 R15: ffff88b7c9059800 [ 364.049599] FS: 00007fb2d3516e80(0000) GS:ffff88bb0f500000(0000) knlGS:0000000000000000 [ 364.049601] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 364.049603] CR2: 000055610bd03598 CR3: 00000001049f6000 CR4: 0000000000350ef0 [ 364.049605] Call Trace: [ 364.049607] [ 364.049609] ? show_regs+0x6d/0x80 [ 364.049614] ? __warn+0x8c/0x140 [ 364.049618] ? debug_dma_map_sg+0x2dc/0x370 [ 364.049621] ? report_bug+0x193/0x1a0 [ 364.049627] ? handle_bug+0x46/0x80 [ 364.049631] ? exc_invalid_op+0x1d/0x80 [ 364.049635] ? asm_exc_invalid_op+0x1f/0x30 [ 364.049642] ? debug_dma_map_sg+0x2dc/0x370 [ 364.049647] __dma_map_sg_attrs+0x90/0xe0 [ 364.049651] dma_map_sgtable+0x25/0x40 [ 364.049654] amdgpu_bo_move+0x59a/0x850 [amdgpu] [ 364.049935] ? srso_return_thunk+0x5/0x5f [ 364.049939] ? amdgpu_ttm_tt_populate+0x5d/0xc0 [amdgpu] [ 364.050095] ttm_bo_handle_move_mem+0xc3/0x180 [ttm] [ 364.050103] ttm_bo_validate+0xc1/0x160 [ttm] [ 364.050108] ? amdgpu_ttm_tt_get_user_pages+0xe5/0x1b0 [amdgpu] [ 364.050263] amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu+0xa12/0xc90 [amdgpu] [ 364.050473] kfd_ioctl_alloc_memory_of_gpu+0x16b/0x3b0 [amdgpu] [ 364.050680] kfd_ioctl+0x3c2/0x530 [amdgpu] [ 364.050866] ? __pfx_kfd_ioctl_alloc_memory_of_gpu+0x10/0x10 [amdgpu] [ 364.051054] ? srso_return_thunk+0x5/0x5f [ 364.051057] ? tomoyo_file_ioctl+0x20/0x30 [ 364.051063] __x64_sys_ioctl+0x9c/0xd0 [ 364.051068] x64_sys_call+0x1219/0x20d0 [ 364.051073] do_syscall_64+0x51/0x120 [ 364.051077] entry_SYSCALL_64_after_hwframe+0x76/0x7e [ 364.051081] RIP: 0033:0x7fb2d2f1a94f Signed-off-by: Prike Liang Reviewed-by: Christian König Signed-off-by: Alex Deucher Signed-off-by: Sasha Levin

drm/amdgpu: refine error handling in amdgpu_ttm_tt_pin_userptr

2024-12-14T19:00:06+00:00

[ Upstream commit 46186667f98fb7158c98f4ff5da62c427761ffcd ] Free sg table when dma_map_sgtable() failed to avoid memory leak. Signed-off-by: Lang Yu Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Signed-off-by: Sasha Levin

drm/amdgpu: once more fix the call oder in amdgpu_ttm_move() v2

2024-05-17T10:02:27+00:00

commit d3a9331a6591e9df64791e076f6591f440af51c3 upstream. This reverts drm/amdgpu: fix ftrace event amdgpu_bo_move always move on same heap. The basic problem here is that after the move the old location is simply not available any more. Some fixes were suggested, but essentially we should call the move notification before actually moving things because only this way we have the correct order for DMA-buf and VM move notifications as well. Also rework the statistic handling so that we don't update the eviction counter before the move. v2: add missing NULL check Signed-off-by: Christian König Fixes: 94aeb4117343 ("drm/amdgpu: fix ftrace event amdgpu_bo_move always move on same heap") Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3171 Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher CC: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: Fix comparison in amdgpu_res_cpu_visible

2024-05-17T10:02:27+00:00

commit 8d2c930735f850e5be6860aeb39b27ac73ca192f upstream. It incorrectly claimed a resource isn't CPU visible if it's located at the very end of CPU visible VRAM. Fixes: a6ff969fe9cb ("drm/amdgpu: fix visible VRAM handling during faults") Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3343 Reviewed-by: Christian König Reviewed-by: Alex Deucher Reported-and-Tested-by: Jeremy Day Signed-off-by: Michel Dänzer Signed-off-by: Alex Deucher CC: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: fix doorbell regression

2024-05-17T10:02:05+00:00

[ Upstream commit 705d0480e6ae5a73ca3a9c04316d0678e19a46ed ] This patch adds a missed handling of PL domain doorbell while handling VRAM faults. Cc: Christian Koenig Cc: Alex Deucher Fixes: a6ff969fe9cb ("drm/amdgpu: fix visible VRAM handling during faults") Reviewed-by: Christian Koenig Signed-off-by: Shashank Sharma Signed-off-by: Arvind Yadav Signed-off-by: Alex Deucher Signed-off-by: Sasha Levin