On Sun, Mar 23, 2025 at 12:25:58AM +0300, Dmitry Osipenko wrote:
Hi,
This a continuation of a year-old series that adds generic DRM-shmem
shrinker [1]. The old series became too big with too many patches, more
reasonable to split it up into multiple smaller patchsets. Here is
the firtst part that makes preparatory DRM changes.
[1]
https://lore.kernel.org/dri-devel/20240105184624.508603-1-dmitry.osipe...@collabora.com/
After these patches got merged I started seeing this on ast driver
and a similar one qemu-cirrus:
[ 88.731560] fbcon: astdrmfb (fb0) is primary device
[ 88.767100] Console: switching to colour frame buffer device 128x48
[ 88.768982] ------------[ cut here ]------------
[ 88.768989] ast 0000:02:00.0: [drm] Dirty helper failed: ret=-12
[ 88.769016] WARNING: CPU: 0 PID: 10 at drivers/gpu/drm/drm_fbdev_shmem.c:118
drm_fbdev_shmem_helper_fb_dirty+0xcc/0xd0 [drm_shmem_helper]
...
[ 88.769092] CPU: 0 UID: 0 PID: 10 Comm: kworker/0:1 Not tainted 6.14.0-xe+ #2
...
[ 88.769095] Workqueue: events drm_fb_helper_damage_work [drm_kms_helper]
[ 88.769109] RIP: 0010:drm_fbdev_shmem_helper_fb_dirty+0xcc/0xd0
[drm_shmem_helper]
[ 88.769112] Code: 4d 8b 6c 24 50 4d 85 ed 75 04 4d 8b 2c 24 4c 89 e7 e8 98 36 a9
e0 89 d9 4c 89 ea 48 c7 c7 d8 65 54 a1 48 89 c6 e8 64 f4 ee df <0f> 0b eb 8b 90
90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 0f 1f
[ 88.769113] RSP: 0018:ffa00000003e7da0 EFLAGS: 00010246
[ 88.769115] RAX: 0000000000000000 RBX: 00000000fffffff4 RCX: 0000000000000000
[ 88.769117] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
[ 88.769117] RBP: ffa00000003e7db8 R08: 0000000000000000 R09: 0000000000000000
[ 88.769118] R10: 0000000000000000 R11: 0000000000000000 R12: ff11000122c190c8
[ 88.769119] R13: ff11000118588a10 R14: ff1100010f4a04c0 R15: ff1100019750b588
[ 88.769120] FS: 0000000000000000(0000) GS:ff11007e7d000000(0000)
knlGS:0000000000000000
[ 88.769121] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 88.769122] CR2: 00007e43af4056f0 CR3: 00000080c89f0003 CR4: 0000000000f73ef0
[ 88.769123] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 88.769124] DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
[ 88.769125] PKRU: 55555554
[ 88.769126] Call Trace:
[ 88.769127] <TASK>
[ 88.769129] ? show_regs.cold+0x19/0x20
[ 88.769132] ? drm_fbdev_shmem_helper_fb_dirty+0xcc/0xd0 [drm_shmem_helper]
[ 88.769134] ? __warn.cold+0xd2/0x17f
[ 88.769136] ? drm_fbdev_shmem_helper_fb_dirty+0xcc/0xd0 [drm_shmem_helper]
[ 88.769139] ? report_bug+0x114/0x160
[ 88.769143] ? handle_bug+0x6e/0xb0
[ 88.769146] ? exc_invalid_op+0x18/0x80
[ 88.769147] ? asm_exc_invalid_op+0x1b/0x20
[ 88.769153] ? drm_fbdev_shmem_helper_fb_dirty+0xcc/0xd0 [drm_shmem_helper]
[ 88.769156] drm_fb_helper_damage_work+0x8a/0x190 [drm_kms_helper]
[ 88.769164] process_one_work+0x224/0x5f0
[ 88.769170] worker_thread+0x1d4/0x3e0
[ 88.769173] ? __pfx_worker_thread+0x10/0x10
[ 88.769175] kthread+0x10b/0x260
[ 88.769178] ? __pfx_kthread+0x10/0x10
[ 88.769181] ret_from_fork+0x44/0x70
[ 88.769184] ? __pfx_kthread+0x10/0x10
[ 88.769186] ret_from_fork_asm+0x1a/0x30
[ 88.769193] </TASK>
[ 88.769194] irq event stamp: 430275
[ 88.769195] hardirqs last enabled at (430281): [<ffffffff81505836>]
vprintk_emit+0x416/0x470
[ 88.769198] hardirqs last disabled at (430286): [<ffffffff81505868>]
vprintk_emit+0x448/0x470
[ 88.769200] softirqs last enabled at (428924): [<ffffffff8143977d>]
__irq_exit_rcu+0x12d/0x150
[ 88.769203] softirqs last disabled at (428861): [<ffffffff8143977d>]
__irq_exit_rcu+0x12d/0x150
[ 88.769205] ---[ end trace 0000000000000000 ]---
[ 88.769375] ast 0000:02:00.0: [drm] fb0: astdrmfb frame buffer device
[ 88.802554] ast 0000:02:00.0: [drm:drm_fb_helper_hotplug_event [drm_kms_helper]]
I confirmed that reverting the series (together with 2 patches for accel/ivpu
on top)
makes the warning go away. Any idea what's going on?
Lucas De Marchi
