https://bugs.kde.org/show_bug.cgi?id=345077

--- Comment #7 from Mark Wielaard <[email protected]> ---
(In reply to Alexandra Hajkova from comment #6)
> Created attachment 127385 [details]
> patch

We discussed this on irc a bit:

- Nicely split the generic code from the execve and execveat PRE handlers.
- The test for whether it is an absolute PATH is wrong, should be (path[0] !=
'/').
- abs_path needs to be allocate before use.
- Because abs_path and buf are allocated they need to be VG_(free) at the end
after handle_pre_sys_execve returns (which means the syscall failed, otherwise
the function doesn't return).
- Because path might now be replaced the am_is_valid_for_client check should be
moved from the generic code to the execve and execveat PRE handlers.
- Since glibc doesn't have a wrapper yet, best to have our own syscall wrapper
in the testcase (with a prereq check case, then you can unconditionally
add/compile the test and you can get rid of the configure.ac tests and
HAVE_EXECVEAT conditional).
- Testcase needs to test errors before success case (because success never
returns).

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

Reply via email to