https://bugs.kde.org/show_bug.cgi?id=510200

[email protected] changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #185547|0                           |1
        is obsolete|                            |

--- Comment #9 from [email protected] ---
Created attachment 185660
  --> https://bugs.kde.org/attachment.cgi?id=185660&action=edit
updated patch

(In reply to Paul Floyd from comment #5)
> Does this also affect other syscalls where an input path is longer than
> PATH_MAX?

I think so.  I think PRE_MEM_RASCIIZ() only works for properly null terminated
strings.

> The call to  ML_(fd_at_check_allowed) should be passing in "mknodat". Also I
> prefer to use the same argument names as the man page (where syscalls have a
> direct libc equivalent). If the user sees our error and then looks at "man
> {error-syscall -function}" then its clearer if all the names match. So that
> would be "mknodat(dirfd)".

Understood, makes sense.  Attaching updated patch.   And it's now being tested
here: https://builder.sourceware.org/buildbot/#/changes/98002 .

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to