Skip to content

xfs: fix random format verification issues

Dave Chinner linux-xfs lkml llvm oe-lkp 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest4 patches 16 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/4xfs: detect self referencing btree sibling pointersfs/xfs/libxfs/xfs_btree.c+102 -357 replies
2/4xfs: validate inode fork size against fork formatfs/xfs/libxfs/xfs_inode_buf.c+26 -92 replies
3/4xfs: set XFS_FEAT_NLINK correctlyfs/xfs/libxfs/xfs_sb.c+2 -02 replies
4/4xfs: validate v5 feature fieldsfs/xfs/libxfs/xfs_sb.c+57 -105 replies

Reviews trailers per patch

All 4 patches have a Reviewed-by or Acked-by.

PatchChristoph HellwigDarrick J. Wong
1/4R-b R-b
2/4R-b R-b
3/4R-b R-b
4/4R-b R-b

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.

Read the v1 thread

Cover letter [PATCH 0/4] xfs: fix random format verification issues

Hi folks,

This series contains a handful of fixes for the malicious format verification attacks that were dumped into the kernel.org bugzilla over the weekend.

I'm not going to give any credit to the reporter - they are not disclosing these issues responsibly and so I'm not going to encourage them by giving them any credit for causing me unnecessary stress by forcing me to drop everything to investigate the reported problems.

As I said back in my response to a similar malicious bug reporter recently at https://bugzilla.kernel.org/show_bug.cgi?id=215783 :

"If you are going to run some scripted tool to randomly corrupt the filesystem to find failures, then you have an ethical and moral responsibility to do some of the work to narrow down and identify the cause of the failure, not just throw them at someone to do all the work."

I'll add that *responsible disclosure* is also necessary. Giving us a chance to determine the severity and impact of the format verification/corruption problem before they are made public would take a lot of the angst and stress out of this situation.

</rant>

Anyway, one of the bug fixes "breaks" xfs/348, which is actually broken because it encodes observed (broken) behaviour as success rather than encoding correct behaviour and then failing. This series ensures that the data fork is in the correct form for symlinks, directories and regular files and so the test now fails. Fixes for that test will need to be done.

-Dave.