| Message ID | 20201222011634.124180-1-yuchao0@huawei.com |
|---|---|
| State | New |
| Headers | show |
| Series | [5.10.y] f2fs: fix to seek incorrect data offset in inline data file | expand |
On Tue, Dec 22, 2020 at 09:36:35AM +0100, Greg KH wrote: > On Tue, Dec 22, 2020 at 09:16:34AM +0800, Chao Yu wrote: > > As kitestramuort reported: > > > > F2FS-fs (nvme0n1p4): access invalid blkaddr:1598541474 > > [ 25.725898] ------------[ cut here ]------------ > > [ 25.725903] WARNING: CPU: 6 PID: 2018 at f2fs_is_valid_blkaddr+0x23a/0x250 > > [ 25.725923] Call Trace: > > [ 25.725927] ? f2fs_llseek+0x204/0x620 > > [ 25.725929] ? ovl_copy_up_data+0x14f/0x200 > > [ 25.725931] ? ovl_copy_up_inode+0x174/0x1e0 > > [ 25.725933] ? ovl_copy_up_one+0xa22/0xdf0 > > [ 25.725936] ? ovl_copy_up_flags+0xa6/0xf0 > > [ 25.725938] ? ovl_aio_cleanup_handler+0xd0/0xd0 > > [ 25.725939] ? ovl_maybe_copy_up+0x86/0xa0 > > [ 25.725941] ? ovl_open+0x22/0x80 > > [ 25.725943] ? do_dentry_open+0x136/0x350 > > [ 25.725945] ? path_openat+0xb7e/0xf40 > > [ 25.725947] ? __check_sticky+0x40/0x40 > > [ 25.725948] ? do_filp_open+0x70/0x100 > > [ 25.725950] ? __check_sticky+0x40/0x40 > > [ 25.725951] ? __check_sticky+0x40/0x40 > > [ 25.725953] ? __x64_sys_openat+0x1db/0x2c0 > > [ 25.725955] ? do_syscall_64+0x2d/0x40 > > [ 25.725957] ? entry_SYSCALL_64_after_hwframe+0x44/0xa9 > > > > llseek() reports invalid block address access, the root cause is if > > file has inline data, f2fs_seek_block() will access inline data regard > > as block address index in inode block, which should be wrong, fix it. > > > > Fixes: 788e96d1d3994 ("f2fs: code cleanup by removing unnecessary check") > > Fixes: 267378d4de69 ("f2fs: introduce f2fs_seek_block to support SEEK_{DATA, HOLE} in llseek") > > Cc: stable <stable@vger.kernel.org> # 3.16+ > > Reported-by: kitestramuort <kitestramuort@autistici.org> > > Signed-off-by: Chao Yu <yuchao0@huawei.com> > > Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> > > --- > > > > This will cause boot failure in f2fs image, which was reported in gentoo > > community, I'd like to fix them in stable kernel 5.10 first as request in > > bugzilla: > > > > https://bugzilla.kernel.org/show_bug.cgi?id=210825 > > <formletter> > > This is not the correct way to submit patches for inclusion in the > stable kernel tree. Please read: > https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html > for how to do this properly. > > </formletter> Specifically, what is the git commit id of this patch in Linus's tree? thanks, greg k-h
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index ee861c6d9ff0..fe39e591e5b4 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -412,9 +412,14 @@ static loff_t f2fs_seek_block(struct file *file, loff_t offset, int whence) goto fail; /* handle inline data case */ - if (f2fs_has_inline_data(inode) && whence == SEEK_HOLE) { - data_ofs = isize; - goto found; + if (f2fs_has_inline_data(inode)) { + if (whence == SEEK_HOLE) { + data_ofs = isize; + goto found; + } else if (whence == SEEK_DATA) { + data_ofs = offset; + goto found; + } } pgofs = (pgoff_t)(offset >> PAGE_SHIFT);
As kitestramuort reported: F2FS-fs (nvme0n1p4): access invalid blkaddr:1598541474 [ 25.725898] ------------[ cut here ]------------ [ 25.725903] WARNING: CPU: 6 PID: 2018 at f2fs_is_valid_blkaddr+0x23a/0x250 [ 25.725923] Call Trace: [ 25.725927] ? f2fs_llseek+0x204/0x620 [ 25.725929] ? ovl_copy_up_data+0x14f/0x200 [ 25.725931] ? ovl_copy_up_inode+0x174/0x1e0 [ 25.725933] ? ovl_copy_up_one+0xa22/0xdf0 [ 25.725936] ? ovl_copy_up_flags+0xa6/0xf0 [ 25.725938] ? ovl_aio_cleanup_handler+0xd0/0xd0 [ 25.725939] ? ovl_maybe_copy_up+0x86/0xa0 [ 25.725941] ? ovl_open+0x22/0x80 [ 25.725943] ? do_dentry_open+0x136/0x350 [ 25.725945] ? path_openat+0xb7e/0xf40 [ 25.725947] ? __check_sticky+0x40/0x40 [ 25.725948] ? do_filp_open+0x70/0x100 [ 25.725950] ? __check_sticky+0x40/0x40 [ 25.725951] ? __check_sticky+0x40/0x40 [ 25.725953] ? __x64_sys_openat+0x1db/0x2c0 [ 25.725955] ? do_syscall_64+0x2d/0x40 [ 25.725957] ? entry_SYSCALL_64_after_hwframe+0x44/0xa9 llseek() reports invalid block address access, the root cause is if file has inline data, f2fs_seek_block() will access inline data regard as block address index in inode block, which should be wrong, fix it. Fixes: 788e96d1d3994 ("f2fs: code cleanup by removing unnecessary check") Fixes: 267378d4de69 ("f2fs: introduce f2fs_seek_block to support SEEK_{DATA, HOLE} in llseek") Cc: stable <stable@vger.kernel.org> # 3.16+ Reported-by: kitestramuort <kitestramuort@autistici.org> Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> --- This will cause boot failure in f2fs image, which was reported in gentoo community, I'd like to fix them in stable kernel 5.10 first as request in bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=210825 fs/f2fs/file.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)