[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20240712131931.20207-1-00107082@163.com>
Date: Fri, 12 Jul 2024 21:19:31 +0800
From: David Wang <00107082@....com>
To: peterx@...hat.com
Cc: akpm@...ux-foundation.org,
david@...hat.com,
linux-kernel-mentees@...ts.linuxfoundation.org,
linux-kernel@...r.kernel.org,
linux-mm@...ck.org,
peili.dev@...il.com,
skhan@...uxfoundation.org,
syzbot+35a4414f6e247f515443@...kaller.appspotmail.com,
syzkaller-bugs@...glegroups.com
Subject: Re: [PATCH] mm: Fix mmap_assert_locked() in follow_pte()
Hi,
> Ah yes, I had one rfc patch for that, I temporarily put that aside as it
> seemed nobody cared except myself.. it's here:
>
> https://lore.kernel.org/all/20240523223745.395337-2-peterx@redhat.com
>
> I didn't know it can already cause real trouble. It looks like that patch
> should fix this.
>
> Thanks,
>
> --
> Peter Xu
Just add another user scenario concering this kernel warning.
Ever since 6.10-rc1, when I suspend my system via `systemctl suspend`, nvidia gpu driver would trigger a warning:
Call Trace:
<TASK>
? __warn+0x7c/0x120
? follow_pte+0x15b/0x170
? report_bug+0x18d/0x1c0
? handle_bug+0x3c/0x80
? exc_invalid_op+0x13/0x60
? asm_exc_invalid_op+0x16/0x20
? follow_pte+0x15b/0x170
follow_phys+0x3a/0xf0
untrack_pfn+0x53/0x120
unmap_single_vma+0xa6/0xe0
zap_page_range_single+0xe4/0x190
? _nv002569kms+0x17b/0x210 [nvidia_modeset]
? srso_return_thunk+0x5/0x5f
? kfree+0x257/0x290
unmap_mapping_range+0x10d/0x130
nv_revoke_gpu_mappings_locked+0x43/0x70 [nvidia]
nv_set_system_power_state+0x1c9/0x470 [nvidia]
nv_procfs_write_suspend+0xd3/0x140 [nvidia]
proc_reg_write+0x58/0xa0
? srso_return_thunk+0x5/0x5f
vfs_write+0xf6/0x440
? __count_memcg_events+0x73/0x110
? srso_return_thunk+0x5/0x5f
? count_memcg_events.constprop.0+0x1a/0x30
? srso_return_thunk+0x5/0x5f
? handle_mm_fault+0xa9/0x2d0
? srso_return_thunk+0x5/0x5f
? preempt_count_add+0x47/0xa0
ksys_write+0x63/0xe0
do_syscall_64+0x4b/0x110
entry_SYSCALL_64_after_hwframe+0x76/0x7e
RIP: 0033:0x7f34a3914240
Code: 40 00 48 8b 15 c1 9b 0d 00 f7 d8 64 89 02 48 c7 c0 ff ff ff ff eb b7 0f 1f 00 80 3d a1 23 0e 00 00 74 17 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 58 c3 0f 1f 80 00 00 00 00 48 83 ec 28 48 89
RSP: 002b:00007ffca2aa2688 EFLAGS: 00000202 ORIG_RAX: 0000000000000001
RAX: ffffffffffffffda RBX: 0000000000000008 RCX: 00007f34a3914240
RDX: 0000000000000008 RSI: 000055a02968ed80 RDI: 0000000000000001
RBP: 000055a02968ed80 R08: 00007f34a39eecd0 R09: 00007f34a39eecd0
R10: 0000000000000000 R11: 0000000000000202 R12: 0000000000000008
R13: 00007f34a39ef760 R14: 0000000000000008 R15: 00007f34a39ea9e0
</TASK>
---[ end trace 0000000000000000 ]---
PM: suspend entry (deep)
Considering out-of-tree nature of nvidia gpu driver, and nobody reported this kernel warning before with in-trees,
I had almost convinced myself that nvidia driver may need "big" rework to live with those "PTE" changes.
So glad to see this thread of discussion/issue/fix now, I have been patching my system manually ever since 6.10-rc1,
hope things got fixed soon...
FYI
David
Powered by blists - more mailing lists