Re: [PATCH v2 0/5] fs/iso9660: Fix out-of-bounds read

[Thomas Schmitt]

Hi,

Lidong Chen wrote:
> I ran grub-fastest with both ce_loop ISO files. The endless loops were
> detected and Grub exited accordingly.

Good.


> I didn't know where the grub error message
> were stored in case of grub-fastest.

So you don't see an error message ?

I had the same problem a while ago, when i tried to check that my thoughts
about the loop end condition in grub_iso9660_susp_iterate() are correct.
(This is now covered by your patch 2.)


> But, I traced with gdb, and saw the  code reported the error.

It's on my todo list to learn how to prepare grub-fstest for working with
gdb. Currently gdb says "No debugging symbols found in ./grub-fstest".


> If the diff looks good, I will send the v3 patches set.

I have no objections.

If patches 1 to 4 are included in v3, please tell whether they have changed
towards v2. (I see no reason why they should change. But if they do, i'll
have to compare them with the earlier versions.)

---------------------------------------------------------------------

I still riddle about how the error message can become visible to the user.
I don't get ideas for that from
  https://www.gnu.org/software/grub/manual/grub-dev/grub-dev.html#Error-Handling

I wonder what is supposed to happen to the "textual message" component of
a grub_error() call. Under which conditions will it be displayed ?
And where ?


Have a nice day :)

Thomas