mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] dma-buf/udmabuf: reinstate a reasonable default size limit
@ 2026-09-22  1:23 Jiakai Xu
  2026-09-22  7:22 ` Christian König
  0 siblings, 1 reply; 2+ messages in thread
From: Jiakai Xu @ 2026-09-22  1:23 UTC (permalink / raw)
  To: kraxel, vivek.kasireddy, sumit.semwal, christian.koenig
  Cc: robert.mader, dri-devel, linux-media, linaro-mm-sig,
	linux-kernel, Jiakai Xu

Commit 44e9eb5a7621 ("dma-buf/udmabuf: Disable the size limit by
default") raised the default size_limit_mb from 64 to INT_MAX.

With 4 KiB pages on a 64-bit system, creating a buffer larger than
about 1 TiB makes the two per-page pointer arrays in init_udmabuf()
exceed what kvmalloc() can handle.  __kvmalloc_node_noprof() warns and
fails when the requested size exceeds INT_MAX bytes, which is
trivially triggerable by unprivileged users through the UDMABUF_CREATE
ioctls:

  WARNING: CPU: 0 PID: 412 at mm/slub.c:7021 __kvmalloc_node_noprof
  Call Trace:
   udmabuf_create
   udmabuf_ioctl_create_list

Instead of using a fixed megabyte value, derive the default limit from
the kvmalloc() threshold:

  (((INT_MAX / sizeof(struct folio *)) - 1) >> 20) << PAGE_SHIFT

On a 64-bit system with 4 KiB pages, this evaluates to 1044480 MB
(1020 GiB), allowing buffers a little below 1 TiB while keeping the
per-page pointer array allocations below INT_MAX bytes.

Fixes: 44e9eb5a7621 ("dma-buf/udmabuf: Disable the size limit by default")
Suggested-by: Christian König <christian.koenig@amd.com>
Assisted-by: OpenCode:DeepSeek-V4-Flash
Signed-off-by: Jiakai Xu <xujiakai24@mails.ucas.ac.cn>
---
V2 -> V3:
- Replace the fixed 256 MB default with a limit derived from the
  kvmalloc() INT_MAX threshold, as suggested by Christian König.  On
  64-bit systems with 4 KiB pages, this allows buffers just below
  1 TiB.

V1 -> V2:
- Instead of adding a pgcnt check in udmabuf_create(), restore a
  reasonable default size limit (256 MB), as suggested by Christian
  König.
- Drop the second Fixes tag; the regression is solely due to
  44e9eb5a7621.
- Drop the "Cc: stable" tag; 44e9eb5a7621 is only in v7.3-rc1 and has
  not shipped in any released kernel yet.

V1: https://lore.kernel.org/all/20260918013434.1849135-1-xujiakai24@mails.ucas.ac.cn/
V2: https://lore.kernel.org/all/20260919015731.2077604-1-xujiakai24@mails.ucas.ac.cn/
---
 drivers/dma-buf/udmabuf.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/dma-buf/udmabuf.c b/drivers/dma-buf/udmabuf.c
index df6dd00462423..7ec269809fb49 100644
--- a/drivers/dma-buf/udmabuf.c
+++ b/drivers/dma-buf/udmabuf.c
@@ -20,9 +20,10 @@ static int list_limit = 1024;
 module_param(list_limit, int, 0644);
 MODULE_PARM_DESC(list_limit, "udmabuf_create_list->count limit. Default is 1024.");
 
-static int size_limit_mb = INT_MAX;
+static int size_limit_mb = (((INT_MAX / sizeof(struct folio *)) - 1) >> 20) << PAGE_SHIFT;
 module_param(size_limit_mb, int, 0644);
-MODULE_PARM_DESC(size_limit_mb, "Max size of a dmabuf, in megabytes. Default is INT_MAX.");
+MODULE_PARM_DESC(size_limit_mb,
+		 "Max size of a dmabuf, in megabytes. Default is derived from the kvmalloc() limit.");
 
 struct udmabuf {
 	pgoff_t pagecount;
-- 
2.34.1


Below is the crash report:

------------[ cut here ]------------
WARNING: mm/slub.c:7013 at __kvmalloc_node_noprof+0x516/0x5d0 home/zzzrrll/tmp/kf_src/linux-7.3-rc2/mm/slub.c:7013, CPU#0: syz.5.27938/167363
Modules linked in:
CPU: 0 UID: 0 PID: 167363 Comm: syz.5.27938 Tainted: G        W    L      7.3.0-rc2 #1 PREEMPT(full) 
Tainted: [W]=WARN, [L]=SOFTLOCKUP
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014
RIP: 0010:__kvmalloc_node_noprof+0x516/0x5d0 home/zzzrrll/tmp/kf_src/linux-7.3-rc2/mm/slub.c:7013
Code: 05 bf fa 2d 08 49 8b 0f c6 04 08 00 65 48 ff 0d 90 fa 2d 08 0f 85 4c fd ff ff e8 05 43 8c ff e9 42 fd ff ff e8 9b 86 0e 04 90 <0f> 0b 90 31 db e9 d2 fe ff ff f3 0f 1e fa 65 8b 05 8d fa 2d 08 89
RSP: 0018:ffffc90002bc3cf8 EFLAGS: 00010246
RAX: 0000000000000100 RBX: 0000000000000000 RCX: 0000008000000000
RDX: 0000008000000000 RSI: 0000000000000000 RDI: ffffffff835e3226
RBP: 00000000ffffffff R08: 00000000000028c0 R09: 00000000ffffffff
R10: 0000000000000006 R11: 0000000000000002 R12: 0000008000000000
R13: 00000000000028c0 R14: 0000001000000000 R15: 0000008000000000
FS:  00007fc89ae45640(0000) GS:ffff8880f3de4000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000110c2ff305 CR3: 000000002b3d4000 CR4: 0000000000752ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000006000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
PKRU: 80000000
Call Trace:
 <TASK>
 init_udmabuf home/zzzrrll/tmp/kf_src/linux-7.3-rc2/drivers/dma-buf/udmabuf.c:193 [inline]
 udmabuf_create+0x1b6/0x750 home/zzzrrll/tmp/kf_src/linux-7.3-rc2/drivers/dma-buf/udmabuf.c:386
 udmabuf_ioctl_create home/zzzrrll/tmp/kf_src/linux-7.3-rc2/drivers/dma-buf/udmabuf.c:465 [inline]
 udmabuf_ioctl+0x132/0x1d0 home/zzzrrll/tmp/kf_src/linux-7.3-rc2/drivers/dma-buf/udmabuf.c:496
 vfs_ioctl home/zzzrrll/tmp/kf_src/linux-7.3-rc2/fs/ioctl.c:51 [inline]
 __do_sys_ioctl home/zzzrrll/tmp/kf_src/linux-7.3-rc2/fs/ioctl.c:597 [inline]
 __se_sys_ioctl+0xbc/0x130 home/zzzrrll/tmp/kf_src/linux-7.3-rc2/fs/ioctl.c:583
 do_syscall_x64 home/zzzrrll/tmp/kf_src/linux-7.3-rc2/arch/x86/entry/syscall_64.c:61 [inline]
 do_syscall_64+0x12b/0x350 home/zzzrrll/tmp/kf_src/linux-7.3-rc2/arch/x86/entry/syscall_64.c:84
 entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x590d6d
Code: 02 b8 ff ff ff ff c3 66 0f 1f 44 00 00 f3 0f 1e fa 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 a8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007fc89ae44fd8 EFLAGS: 00000216 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 0000000000611e7f RCX: 0000000000590d6d
RDX: 00002000000023c0 RSI: 0000000040187542 RDI: 000000000000000a
RBP: 00007fc89ae45010 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000216 R12: 00007fc89ae45640
R13: 000000000000000b R14: 0000000000528d40 R15: 00007fc89ae25000
 </TASK>
---[ end trace 0000000000000000 ]---


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH] dma-buf/udmabuf: reinstate a reasonable default size limit
  2026-09-22  1:23 [PATCH] dma-buf/udmabuf: reinstate a reasonable default size limit Jiakai Xu
@ 2026-09-22  7:22 ` Christian König
  0 siblings, 0 replies; 2+ messages in thread
From: Christian König @ 2026-09-22  7:22 UTC (permalink / raw)
  To: Jiakai Xu, kraxel, vivek.kasireddy, sumit.semwal
  Cc: robert.mader, dri-devel, linux-media, linaro-mm-sig, linux-kernel

On 9/22/26 03:23, Jiakai Xu wrote:
> Commit 44e9eb5a7621 ("dma-buf/udmabuf: Disable the size limit by
> default") raised the default size_limit_mb from 64 to INT_MAX.
> 
> With 4 KiB pages on a 64-bit system, creating a buffer larger than
> about 1 TiB makes the two per-page pointer arrays in init_udmabuf()
> exceed what kvmalloc() can handle.  __kvmalloc_node_noprof() warns and
> fails when the requested size exceeds INT_MAX bytes, which is
> trivially triggerable by unprivileged users through the UDMABUF_CREATE
> ioctls:
> 
>   WARNING: CPU: 0 PID: 412 at mm/slub.c:7021 __kvmalloc_node_noprof
>   Call Trace:
>    udmabuf_create
>    udmabuf_ioctl_create_list
> 
> Instead of using a fixed megabyte value, derive the default limit from
> the kvmalloc() threshold:
> 
>   (((INT_MAX / sizeof(struct folio *)) - 1) >> 20) << PAGE_SHIFT
> 
> On a 64-bit system with 4 KiB pages, this evaluates to 1044480 MB
> (1020 GiB), allowing buffers a little below 1 TiB while keeping the
> per-page pointer array allocations below INT_MAX bytes.
> 
> Fixes: 44e9eb5a7621 ("dma-buf/udmabuf: Disable the size limit by default")
> Suggested-by: Christian König <christian.koenig@amd.com>
> Assisted-by: OpenCode:DeepSeek-V4-Flash
> Signed-off-by: Jiakai Xu <xujiakai24@mails.ucas.ac.cn>

Reviewed-by: Christian König <christian.koenig@amd.com>

Going to push that to drm-misc-fixes later today.

Regards,
Christian.

> ---
> V2 -> V3:
> - Replace the fixed 256 MB default with a limit derived from the
>   kvmalloc() INT_MAX threshold, as suggested by Christian König.  On
>   64-bit systems with 4 KiB pages, this allows buffers just below
>   1 TiB.
> 
> V1 -> V2:
> - Instead of adding a pgcnt check in udmabuf_create(), restore a
>   reasonable default size limit (256 MB), as suggested by Christian
>   König.
> - Drop the second Fixes tag; the regression is solely due to
>   44e9eb5a7621.
> - Drop the "Cc: stable" tag; 44e9eb5a7621 is only in v7.3-rc1 and has
>   not shipped in any released kernel yet.
> 
> V1: https://lore.kernel.org/all/20260918013434.1849135-1-xujiakai24@mails.ucas.ac.cn/
> V2: https://lore.kernel.org/all/20260919015731.2077604-1-xujiakai24@mails.ucas.ac.cn/
> ---
>  drivers/dma-buf/udmabuf.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/dma-buf/udmabuf.c b/drivers/dma-buf/udmabuf.c
> index df6dd00462423..7ec269809fb49 100644
> --- a/drivers/dma-buf/udmabuf.c
> +++ b/drivers/dma-buf/udmabuf.c
> @@ -20,9 +20,10 @@ static int list_limit = 1024;
>  module_param(list_limit, int, 0644);
>  MODULE_PARM_DESC(list_limit, "udmabuf_create_list->count limit. Default is 1024.");
>  
> -static int size_limit_mb = INT_MAX;
> +static int size_limit_mb = (((INT_MAX / sizeof(struct folio *)) - 1) >> 20) << PAGE_SHIFT;
>  module_param(size_limit_mb, int, 0644);
> -MODULE_PARM_DESC(size_limit_mb, "Max size of a dmabuf, in megabytes. Default is INT_MAX.");
> +MODULE_PARM_DESC(size_limit_mb,
> +		 "Max size of a dmabuf, in megabytes. Default is derived from the kvmalloc() limit.");
>  
>  struct udmabuf {
>  	pgoff_t pagecount;


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2026-09-22  7:22 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-09-22  1:23 [PATCH] dma-buf/udmabuf: reinstate a reasonable default size limit Jiakai Xu
2026-09-22  7:22 ` Christian König

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®