[PATCH 5/5] drm/amdgpu: Acquire BO locks only during BO processing
Christian König
christian.koenig at amd.com
Fri May 2 13:07:56 UTC 2025
On 4/9/25 07:48, Arunpravin Paneer Selvam wrote:
> Fix lockdep warnings.
>
> [ +0.000024] WARNING: CPU: 10 PID: 1909 at drivers/gpu/drm/drm_syncobj.c:456 drm_syncobj_find_fence+0x58c/0x6e0
> [ +0.000519] CPU: 10 UID: 1000 PID: 1909 Comm: gnome-shel:cs0 Tainted: G W OE 6.12.0+ #18
> [ +0.000008] Tainted: [W]=WARN, [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
> [ +0.000004] Hardware name: ASUS System Product Name/TUF GAMING B650-PLUS, BIOS 3072 12/20/2024
> [ +0.000004] RIP: 0010:drm_syncobj_find_fence+0x58c/0x6e0
> [ +0.000006] RSP: 0018:ffff88846d9ef680 EFLAGS: 00010202
> [ +0.000008] RAX: 0000000000000000 RBX: 0000000000001388 RCX: 0000000000000001
> [ +0.000004] RDX: 1ffff1108f5ad1da RSI: 0000000000000001 RDI: ffff88847ad68ed0
> [ +0.000005] RBP: ffff88846d9ef770 R08: 0000000000000000 R09: 0000000000000000
> [ +0.000004] R10: 0000000000000000 R11: 0000000000000000 R12: ffff88847ad68000
> [ +0.000004] R13: 0000000000000002 R14: ffff888149353d00 R15: 000000000000000f
> [ +0.000005] FS: 00007269977fe6c0(0000) GS:ffff888409500000(0000) knlGS:0000000000000000
> [ +0.000005] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [ +0.000004] CR2: 0000643a866b50d0 CR3: 0000000469eb2000 CR4: 0000000000f50ef0
> [ +0.000005] PKRU: 55555554
> [ +0.000004] Call Trace:
> [ +0.000004] <TASK>
> [ +0.000005] ? show_regs+0x6c/0x80
> [ +0.000010] ? __warn+0xd2/0x2d0
> [ +0.000008] ? drm_syncobj_find_fence+0x58c/0x6e0
> [ +0.000009] ? report_bug+0x282/0x2f0
> [ +0.000014] ? handle_bug+0x6e/0xc0
> [ +0.000008] ? exc_invalid_op+0x18/0x50
> [ +0.000007] ? asm_exc_invalid_op+0x1b/0x20
> [ +0.000020] ? drm_syncobj_find_fence+0x58c/0x6e0
> [ +0.000012] ? __pfx_drm_syncobj_find_fence+0x10/0x10
> [ +0.000012] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000006] ? lock_is_held_type+0xa3/0x130
> [ +0.000016] amdgpu_userq_wait_ioctl+0x92d/0x2200 [amdgpu]
> [ +0.000257] ? amdgpu_userq_wait_ioctl+0x92d/0x2200 [amdgpu]
> [ +0.000180] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000006] ? __lock_acquire+0x1b19/0x69c0
> [ +0.000022] ? __pfx_amdgpu_userq_wait_ioctl+0x10/0x10 [amdgpu]
> [ +0.000179] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000006] ? __kasan_check_read+0x11/0x20
> [ +0.000007] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? mark_lock+0xfd/0x17c0
> [ +0.000012] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? lock_acquire.part.0+0x116/0x360
> [ +0.000006] ? drm_dev_enter+0x51/0x190
> [ +0.000008] ? __pfx___lock_acquire+0x10/0x10
> [ +0.000008] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? find_held_lock+0x36/0x140
> [ +0.000021] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? lock_acquire+0x7c/0xc0
> [ +0.000005] ? drm_dev_enter+0x51/0x190
> [ +0.000027] drm_ioctl_kernel+0x18b/0x330
> [ +0.000008] ? __pfx_amdgpu_userq_wait_ioctl+0x10/0x10 [amdgpu]
> [ +0.000181] ? __pfx_drm_ioctl_kernel+0x10/0x10
> [ +0.000005] ? lock_acquire+0x7c/0xc0
> [ +0.000009] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? __kasan_check_write+0x14/0x30
> [ +0.000005] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000010] drm_ioctl+0x589/0xd00
> [ +0.000006] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000006] ? __pfx_amdgpu_userq_wait_ioctl+0x10/0x10 [amdgpu]
> [ +0.000187] ? __pfx_drm_ioctl+0x10/0x10
> [ +0.000007] ? __pm_runtime_resume+0x80/0x110
> [ +0.000021] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? trace_hardirqs_on+0x53/0x60
> [ +0.000007] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? _raw_spin_unlock_irqrestore+0x51/0x80
> [ +0.000013] amdgpu_drm_ioctl+0xd2/0x1c0 [amdgpu]
> [ +0.000182] __x64_sys_ioctl+0x13a/0x1c0
> [ +0.000012] x64_sys_call+0x11ad/0x25f0
> [ +0.000007] do_syscall_64+0x91/0x180
> [ +0.000007] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? do_syscall_64+0x9d/0x180
> [ +0.000005] ? syscall_exit_to_user_mode+0x95/0x260
> [ +0.000008] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? do_syscall_64+0x9d/0x180
> [ +0.000009] ? __pfx___do_sys_prctl+0x10/0x10
> [ +0.000012] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000009] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000007] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? syscall_exit_to_user_mode+0x95/0x260
> [ +0.000008] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? do_syscall_64+0x9d/0x180
> [ +0.000009] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000009] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000007] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? irqentry_exit_to_user_mode+0x8b/0x260
> [ +0.000007] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? irqentry_exit+0x77/0xb0
> [ +0.000005] ? srso_alias_return_thunk+0x5/0xfbef5
> [ +0.000005] ? exc_page_fault+0x93/0x150
> [ +0.000009] entry_SYSCALL_64_after_hwframe+0x76/0x7e
> [ +0.000005] RIP: 0033:0x7269b0324ded
> [ +0.000006] Code: 04 25 28 00 00 00 48 89 45 c8 31 c0 48 8d 45 10 c7 45 b0 10 00 00 00 48 89 45 b8 48 8d 45 d0 48 89 45 c0 b8 10 00 00 00 0f 05 <89> c2 3d 00 f0 ff ff 77 1a 48 8b 45 c8 64 48 2b 04 25 28 00 00 00
> [ +0.000005] RSP: 002b:00007269977fc9b0 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
> [ +0.000008] RAX: ffffffffffffffda RBX: 00007269977fcb00 RCX: 00007269b0324ded
> [ +0.000004] RDX: 00007269977fcb00 RSI: 00000000c0406458 RDI: 000000000000000d
> [ +0.000004] RBP: 00007269977fca00 R08: 00007269977fcc1c R09: 0000000000000000
> [ +0.000005] R10: 000000000000000f R11: 0000000000000246 R12: 000000000000000d
> [ +0.000004] R13: 00005bce7c309a50 R14: 00007269977fca30 R15: 0000000000000000
> [ +0.000021] </TASK>
> [ +0.000005] irq event stamp: 1359
> [ +0.000004] hardirqs last enabled at (1365): [<ffffffffaa8693a9>] __up_console_sem+0x79/0xa0
> [ +0.000007] hardirqs last disabled at (1370): [<ffffffffaa86938e>] __up_console_sem+0x5e/0xa0
> [ +0.000005] softirqs last enabled at (756): [<ffffffffaa67377e>] __irq_exit_rcu+0x17e/0x1d0
> [ +0.000006] softirqs last disabled at (749): [<ffffffffaa67377e>] __irq_exit_rcu+0x17e/0x1d0
> [ +0.000005] ---[ end trace 0000000000000000 ]---
>
> Signed-off-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam at amd.com>
> ---
> .../gpu/drm/amd/amdgpu/amdgpu_userq_fence.c | 116 +++++++++++-------
> 1 file changed, 69 insertions(+), 47 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c
> index d5b35b5df527..c5de39a8ff98 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c
> @@ -395,6 +395,43 @@ static void amdgpu_userq_fence_cleanup(struct dma_fence *fence)
> dma_fence_put(fence);
> }
>
> +static int amdgpu_userq_exec_lock(struct drm_exec *exec, u32 flags,
> + struct drm_gem_object **rgobj,
> + struct drm_gem_object **wgobj,
> + u32 num_read_handles,
> + u32 num_write_handles,
> + unsigned int num_fences)
num_fences should always be 1 if I'm not completely mistaken.
> +{
> + int r;
> +
> + if (!exec | !rgobj | !wgobj)
> + return -EINVAL;
> +
> + drm_exec_init(exec, flags,
> + (num_read_handles + num_write_handles));
> +
> + /* Lock all BOs with retry handling */
> + drm_exec_until_all_locked(exec) {
> + r = drm_exec_prepare_array(exec, rgobj, num_read_handles, num_fences);
> + drm_exec_retry_on_contention(exec);
> + if (r)
> + drm_exec_fini(exec);
Don't call drm_exec_fini() here. Rather just return the error.
> +
> + r = drm_exec_prepare_array(exec, wgobj, num_write_handles, num_fences);
> + drm_exec_retry_on_contention(exec);
> + if (r)
> + drm_exec_fini(exec);
Same here.
> + }
> +
> + return r;
> +}
> +
> +static void amdgpu_userq_exec_unlock(struct drm_exec *exec)
> +{
> + if (exec)
> + drm_exec_fini(exec);
> +}
> +
That looks superflous. Just call drm_exec_fini() in the caller.
Regards,
Christian.
> int amdgpu_userq_signal_ioctl(struct drm_device *dev, void *data,
> struct drm_file *filp)
> {
> @@ -511,24 +548,14 @@ int amdgpu_userq_signal_ioctl(struct drm_device *dev, void *data,
> queue->last_fence = dma_fence_get(fence);
> mutex_unlock(&userq_mgr->userq_mutex);
>
> - drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT,
> - (num_read_bo_handles + num_write_bo_handles));
> -
> - /* Lock all BOs with retry handling */
> - drm_exec_until_all_locked(&exec) {
> - r = drm_exec_prepare_array(&exec, gobj_read, num_read_bo_handles, 1);
> - drm_exec_retry_on_contention(&exec);
> - if (r) {
> - amdgpu_userq_fence_cleanup(fence);
> - goto exec_fini;
> - }
> -
> - r = drm_exec_prepare_array(&exec, gobj_write, num_write_bo_handles, 1);
> - drm_exec_retry_on_contention(&exec);
> - if (r) {
> - amdgpu_userq_fence_cleanup(fence);
> - goto exec_fini;
> - }
> + r = amdgpu_userq_exec_lock(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT,
> + gobj_read, gobj_write,
> + num_read_bo_handles,
> + num_write_bo_handles,
> + 1);
> + if (r) {
> + amdgpu_userq_fence_cleanup(fence);
> + goto put_gobj_write;
> }
>
> for (i = 0; i < num_read_bo_handles; i++) {
> @@ -546,6 +573,7 @@ int amdgpu_userq_signal_ioctl(struct drm_device *dev, void *data,
> dma_resv_add_fence(gobj_write[i]->resv, fence,
> DMA_RESV_USAGE_WRITE);
> }
> + amdgpu_userq_exec_unlock(&exec);
>
> /* Add the created fence to syncobj/BO's */
> for (i = 0; i < num_syncobj_handles; i++)
> @@ -554,8 +582,6 @@ int amdgpu_userq_signal_ioctl(struct drm_device *dev, void *data,
> /* drop the reference acquired in fence creation function */
> dma_fence_put(fence);
>
> -exec_fini:
> - drm_exec_fini(&exec);
> put_gobj_write:
> while (wentry-- > 0)
> drm_gem_object_put(gobj_write[wentry]);
> @@ -666,26 +692,6 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> }
> }
>
> - drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT,
> - (num_read_bo_handles + num_write_bo_handles));
> -
> - /* Lock all BOs with retry handling */
> - drm_exec_until_all_locked(&exec) {
> - r = drm_exec_prepare_array(&exec, gobj_read, num_read_bo_handles, 1);
> - drm_exec_retry_on_contention(&exec);
> - if (r) {
> - drm_exec_fini(&exec);
> - goto put_gobj_write;
> - }
> -
> - r = drm_exec_prepare_array(&exec, gobj_write, num_write_bo_handles, 1);
> - drm_exec_retry_on_contention(&exec);
> - if (r) {
> - drm_exec_fini(&exec);
> - goto put_gobj_write;
> - }
> - }
> -
> if (!wait_info->num_fences) {
> if (num_points) {
> struct dma_fence_unwrap iter;
> @@ -698,7 +704,7 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT,
> &fence);
> if (r)
> - goto exec_fini;
> + goto put_gobj_write;
>
> dma_fence_unwrap_for_each(f, &iter, fence)
> num_fences++;
> @@ -716,12 +722,20 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT,
> &fence);
> if (r)
> - goto exec_fini;
> + goto put_gobj_write;
>
> num_fences++;
> dma_fence_put(fence);
> }
>
> + r = amdgpu_userq_exec_lock(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT,
> + gobj_read, gobj_write,
> + num_read_bo_handles,
> + num_write_bo_handles,
> + 1);
> + if (r)
> + goto put_gobj_write;
> +
> /* Count GEM objects fence */
> for (i = 0; i < num_read_bo_handles; i++) {
> struct dma_resv_iter resv_cursor;
> @@ -740,7 +754,7 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> DMA_RESV_USAGE_WRITE, fence)
> num_fences++;
> }
> -
> + amdgpu_userq_exec_unlock(&exec);
> /*
> * Passing num_fences = 0 means that userspace doesn't want to
> * retrieve userq_fence_info. If num_fences = 0 we skip filling
> @@ -753,7 +767,7 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> fence_info = kmalloc_array(wait_info->num_fences, sizeof(*fence_info), GFP_KERNEL);
> if (!fence_info) {
> r = -ENOMEM;
> - goto exec_fini;
> + goto put_gobj_write;
> }
>
> /* Array of fences */
> @@ -763,6 +777,14 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> goto free_fence_info;
> }
>
> + r = amdgpu_userq_exec_lock(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT,
> + gobj_read, gobj_write,
> + num_read_bo_handles,
> + num_write_bo_handles,
> + 1);
> + if (r)
> + goto free_fences;
> +
> /* Retrieve GEM read objects fence */
> for (i = 0; i < num_read_bo_handles; i++) {
> struct dma_resv_iter resv_cursor;
> @@ -772,6 +794,7 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> DMA_RESV_USAGE_READ, fence) {
> if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) {
> r = -EINVAL;
> + amdgpu_userq_exec_unlock(&exec);
> goto free_fences;
> }
>
> @@ -789,6 +812,7 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> DMA_RESV_USAGE_WRITE, fence) {
> if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) {
> r = -EINVAL;
> + amdgpu_userq_exec_unlock(&exec);
> goto free_fences;
> }
>
> @@ -796,6 +820,7 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> dma_fence_get(fence);
> }
> }
> + amdgpu_userq_exec_unlock(&exec);
>
> if (num_points) {
> struct dma_fence_unwrap iter;
> @@ -901,7 +926,6 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> kfree(fence_info);
> }
>
> - drm_exec_fini(&exec);
> for (i = 0; i < num_read_bo_handles; i++)
> drm_gem_object_put(gobj_read[i]);
> kfree(gobj_read);
> @@ -924,8 +948,6 @@ int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
> kfree(fences);
> free_fence_info:
> kfree(fence_info);
> -exec_fini:
> - drm_exec_fini(&exec);
> put_gobj_write:
> while (wentry-- > 0)
> drm_gem_object_put(gobj_write[wentry]);
More information about the amd-gfx
mailing list