Skip to contentSkip to conversation

A list of HFS+ kernel module bugs found in 4.18

DISCUSSION Wen Xu 3 participants 4 messages Jun 27, 2018 – Jul 22, 2018 1 message not in the archive VFS

lore
Share & tools
lore

Wen Xu

author
#0 lore raw

Hi all,

Here are a list of bugs found in HFS+ filesystem by fuzzing. You can get the image and POC that cause kernel panic in the following links:

200299 Kernel panic because mount() hfsplus image does not always return correct value
https://bugzilla.kernel.org/show_bug.cgi?id=200299

200297 Kernel panic in hfsplus_lookup() when open a file in a corrupted hfs+ filesystem
https://bugzilla.kernel.org/show_bug.cgi?id=200297

200295 BUG() in hfsplus_create_attributes_file() when calling setxattr()
https://bugzilla.kernel.org/show_bug.cgi?id=200295

200293 Out-of-bound access in hfsplus_bnode_read()
https://bugzilla.kernel.org/show_bug.cgi?id=200293

200291 Kernel panic when invoking setxattr() on a hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200291

200289 Kernel panic when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200289

200287 Out-of-bound access in hfsplus_bnode_copy() when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200287

200285 Out-of-bound access in hfsplus_bmap_alloc() when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200285

I would like to provide any further help to analyze the crashes and fix the bugs. I am also willing to test the patches.

Thanks,
Wen

A message that is not in the archive (sent off-list or not yet received): [email protected]

#2 lore raw

Hi, thanks for the reports:

Quoted 10 lines · Wen

Hi Ernesto,

I reported the following bugs weeks before but did not get and respond. I saw you commit patches for HFS+, so could you please take a look on these issue?

Here are a list of bugs found in HFS+ filesystem by fuzzing. You can get the image and POC that cause kernel panic in the following links:

200299 Kernel panic because mount() hfsplus image does not always return correct value
https://bugzilla.kernel.org/show_bug.cgi?id=200299

This was also found by syzbot a couple of months ago. It's already fixed in the -mm tree if you want to test it.

Quoted 5 lines

200297 Kernel panic in hfsplus_lookup() when open a file in a corrupted hfs+ filesystem
https://bugzilla.kernel.org/show_bug.cgi?id=200297

200293 Out-of-bound access in hfsplus_bnode_read()
https://bugzilla.kernel.org/show_bug.cgi?id=200293

I'll try to fix these two and get back to you. It may take a couple of days.

Quoted 6 lines

200295 BUG() in hfsplus_create_attributes_file() when calling setxattr()
https://bugzilla.kernel.org/show_bug.cgi?id=200295

200291 Kernel panic when invoking setxattr() on a hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200291

Quoted 8 lines

200289 Kernel panic when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200289

200287 Out-of-bound access in hfsplus_bnode_copy() when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200287

200285 Out-of-bound access in hfsplus_bmap_alloc() when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200285

The xattr implementation is a mess. I found many bugs myself just by attempting to use it. I will take a look, but I can't promise to be able to help. If it was up to me I would much rather get rid of xattr support entirely, although I suppose somebody might be using it.

Ernest

Quoted 4 lines

I would like to provide any further help to analyze the crashes and fix the bugs. I am also willing to test the patches.

Thanks,
Wen

#3 lore raw

Hi Ernesto,

I checked the patch for
200297 Kernel panic in hfsplus_lookup() when open a file in a corrupted hfs+ filesystem
https://bugzilla.kernel.org/show_bug.cgi?id=200297

Is

200293 Out-of-bound access in hfsplus_bnode_read()
https://bugzilla.kernel.org/show_bug.cgi?id=200293

Also handled with any patch?

Thanks,
Wen

Quoted 53 lines · Ernesto A. Fernández

Hi, thanks for the reports:

Quoted 10 lines · Wen

Hi Ernesto,

I reported the following bugs weeks before but did not get and respond. I saw you commit patches for HFS+, so could you please take a look on these issue?

Here are a list of bugs found in HFS+ filesystem by fuzzing. You can get the image and POC that cause kernel panic in the following links:

200299 Kernel panic because mount() hfsplus image does not always return correct value
https://bugzilla.kernel.org/show_bug.cgi?id=200299

This was also found by syzbot a couple of months ago. It's already fixed in the -mm tree if you want to test it.

Quoted 5 lines

200297 Kernel panic in hfsplus_lookup() when open a file in a corrupted hfs+ filesystem
https://bugzilla.kernel.org/show_bug.cgi?id=200297

200293 Out-of-bound access in hfsplus_bnode_read()
https://bugzilla.kernel.org/show_bug.cgi?id=200293

I'll try to fix these two and get back to you. It may take a couple of days.

Quoted 6 lines

200295 BUG() in hfsplus_create_attributes_file() when calling setxattr()
https://bugzilla.kernel.org/show_bug.cgi?id=200295

200291 Kernel panic when invoking setxattr() on a hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200291

Quoted 8 lines

200289 Kernel panic when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200289

200287 Out-of-bound access in hfsplus_bnode_copy() when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200287

200285 Out-of-bound access in hfsplus_bmap_alloc() when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200285

The xattr implementation is a mess. I found many bugs myself just by attempting to use it. I will take a look, but I can't promise to be able to help. If it was up to me I would much rather get rid of xattr support entirely, although I suppose somebody might be using it.

Ernest

Quoted 4 lines

I would like to provide any further help to analyze the crashes and fix the bugs. I am also willing to test the patches.

Thanks,
Wen

#4 lore raw
Quoted 6 lines · Wen

Hi Ernesto,

I checked the patch for
200297 Kernel panic in hfsplus_lookup() when open a file in a corrupted hfs+ filesystem
https://bugzilla.kernel.org/show_bug.cgi?id=200297

Quoted 6 lines

Is

200293 Out-of-bound access in hfsplus_bnode_read()
https://bugzilla.kernel.org/show_bug.cgi?id=200293

Also handled with any patch?

Not yet, I'm afraid. As it turns out, that is also a bug in the xattr implementation. Fixing those will most likely require a big rewrite.

Thanks,
Ernest

Quoted 58 lines

Thanks,
Wen

Quoted 53 lines · Ernesto A. Fernández

Hi, thanks for the reports:

Quoted 10 lines · Wen

Hi Ernesto,

I reported the following bugs weeks before but did not get and respond. I saw you commit patches for HFS+, so could you please take a look on these issue?

Here are a list of bugs found in HFS+ filesystem by fuzzing. You can get the image and POC that cause kernel panic in the following links:

200299 Kernel panic because mount() hfsplus image does not always return correct value
https://bugzilla.kernel.org/show_bug.cgi?id=200299

This was also found by syzbot a couple of months ago. It's already fixed in the -mm tree if you want to test it.

Quoted 5 lines

200297 Kernel panic in hfsplus_lookup() when open a file in a corrupted hfs+ filesystem
https://bugzilla.kernel.org/show_bug.cgi?id=200297

200293 Out-of-bound access in hfsplus_bnode_read()
https://bugzilla.kernel.org/show_bug.cgi?id=200293

I'll try to fix these two and get back to you. It may take a couple of days.

Quoted 6 lines

200295 BUG() in hfsplus_create_attributes_file() when calling setxattr()
https://bugzilla.kernel.org/show_bug.cgi?id=200295

200291 Kernel panic when invoking setxattr() on a hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200291

Quoted 8 lines

200289 Kernel panic when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200289

200287 Out-of-bound access in hfsplus_bnode_copy() when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200287

200285 Out-of-bound access in hfsplus_bmap_alloc() when calling setxattr() on a corrupted hfs+ image
https://bugzilla.kernel.org/show_bug.cgi?id=200285

The xattr implementation is a mess. I found many bugs myself just by attempting to use it. I will take a look, but I can't promise to be able to help. If it was up to me I would much rather get rid of xattr support entirely, although I suppose somebody might be using it.

Ernest

Quoted 4 lines

I would like to provide any further help to analyze the crashes and fix the bugs. I am also willing to test the patches.

Thanks,
Wen