RE: [PATCH] exfat: Fix uninit-value access in __exfat_write_inode()
From: Sungjong Seo
Date: Tue Nov 07 2023 - 23:35:32 EST
Hello,
A similar fix has already been queued in the dev branch.
Please refer to below commit.
Commit fc12a722e6b7 ("exfat: fix setting uninitialized time to
ctime/atime"):
https://git.kernel.org/pub/scm/linux/kernel/git/linkinjeon/exfat.git/commit/
?h=dev&id=fc12a722e6b799d1d3c1520dc9ba9aab4fda04bf
Thanks.
B.R.
Sungjong Seo
> KMSAN reported the following uninit-value access issue:
>
> =====================================================
> BUG: KMSAN: uninit-value in exfat_set_entry_time+0x309/0x360
> fs/exfat/misc.c:99
> exfat_set_entry_time+0x309/0x360 fs/exfat/misc.c:99
> __exfat_write_inode+0x7ae/0xdb0 fs/exfat/inode.c:59
> __exfat_truncate+0x70e/0xb20 fs/exfat/file.c:163
> exfat_truncate+0x121/0x540 fs/exfat/file.c:211
> exfat_setattr+0x116c/0x1a40 fs/exfat/file.c:312
> notify_change+0x1934/0x1a30 fs/attr.c:499
> do_truncate+0x224/0x2a0 fs/open.c:66
> handle_truncate fs/namei.c:3280 [inline] do_open fs/namei.c:3626
[inline]
> path_openat+0x56c6/0x5f20 fs/namei.c:3779
> do_filp_open+0x21c/0x5a0 fs/namei.c:3809
> do_sys_openat2+0x1ba/0x2f0 fs/open.c:1440 do_sys_open fs/open.c:1455
> [inline] __do_sys_creat fs/open.c:1531 [inline] __se_sys_creat
> fs/open.c:1525 [inline]
> __x64_sys_creat+0xe3/0x140 fs/open.c:1525
> do_syscall_x64 arch/x86/entry/common.c:51 [inline]
> do_syscall_64+0x44/0x110 arch/x86/entry/common.c:82
> entry_SYSCALL_64_after_hwframe+0x63/0x6b
>
> Uninit was stored to memory at:
> exfat_set_entry_time+0x302/0x360 fs/exfat/misc.c:99
> __exfat_write_inode+0x7ae/0xdb0 fs/exfat/inode.c:59
> __exfat_truncate+0x70e/0xb20 fs/exfat/file.c:163
> exfat_truncate+0x121/0x540 fs/exfat/file.c:211
> exfat_setattr+0x116c/0x1a40 fs/exfat/file.c:312
> notify_change+0x1934/0x1a30 fs/attr.c:499
> do_truncate+0x224/0x2a0 fs/open.c:66
> handle_truncate fs/namei.c:3280 [inline] do_open fs/namei.c:3626
[inline]
> path_openat+0x56c6/0x5f20 fs/namei.c:3779
> do_filp_open+0x21c/0x5a0 fs/namei.c:3809
> do_sys_openat2+0x1ba/0x2f0 fs/open.c:1440 do_sys_open fs/open.c:1455
> [inline] __do_sys_creat fs/open.c:1531 [inline] __se_sys_creat
> fs/open.c:1525 [inline]
> __x64_sys_creat+0xe3/0x140 fs/open.c:1525
> do_syscall_x64 arch/x86/entry/common.c:51 [inline]
> do_syscall_64+0x44/0x110 arch/x86/entry/common.c:82
> entry_SYSCALL_64_after_hwframe+0x63/0x6b
>
> Local variable ts created at:
> __exfat_write_inode+0x102/0xdb0 fs/exfat/inode.c:29
> __exfat_truncate+0x70e/0xb20 fs/exfat/file.c:163
>
> CPU: 0 PID: 13839 Comm: syz-executor.7 Not tainted 6.6.0-14500-
> g1c41041124bd #10 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
> BIOS 1.16.2-1.fc38 04/01/2014
> =====================================================
>
> Commit 4c72a36edd54 ("exfat: convert to new timestamp accessors") changed
> __exfat_write_inode() to use new timestamp accessor functions.
>
> As for mtime, inode_set_mtime_to_ts() is called after
> exfat_set_entry_time(). This causes the above issue because `ts` is not
> initialized when exfat_set_entry_time() is called. The same issue can
> occur for atime.
>
> This patch resolves this issue by calling inode_get_mtime() and
> inode_get_atime() before exfat_set_entry_time() to initialize `ts`.
>
> Fixes: 4c72a36edd54 ("exfat: convert to new timestamp accessors")
> Signed-off-by: Shigeru Yoshida <syoshida@xxxxxxxxxx>
> ---
> fs/exfat/inode.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/fs/exfat/inode.c b/fs/exfat/inode.c index
> 875234179d1f..e7ff58b8e68c 100644
> --- a/fs/exfat/inode.c
> +++ b/fs/exfat/inode.c
> @@ -56,18 +56,18 @@ int __exfat_write_inode(struct inode *inode, int sync)
> &ep->dentry.file.create_time,
> &ep->dentry.file.create_date,
> &ep->dentry.file.create_time_cs);
> + ts = inode_get_mtime(inode);
> exfat_set_entry_time(sbi, &ts,
> &ep->dentry.file.modify_tz,
> &ep->dentry.file.modify_time,
> &ep->dentry.file.modify_date,
> &ep->dentry.file.modify_time_cs);
> - inode_set_mtime_to_ts(inode, ts);
> + ts = inode_get_atime(inode);
> exfat_set_entry_time(sbi, &ts,
> &ep->dentry.file.access_tz,
> &ep->dentry.file.access_time,
> &ep->dentry.file.access_date,
> NULL);
> - inode_set_atime_to_ts(inode, ts);
>
> /* File size should be zero if there is no cluster allocated */
> on_disk_size = i_size_read(inode);
> --
> 2.41.0