Re: [LTP] syscalls: add v7.3 syscall numbers

[Andrea Cervesato via ltp <ltp@lists.linux.it>]

> Could this subject be prefixed with [STAGING]? Linux 7.2 is the latest
> stable release, while this test targets fchroot() from Linux 7.3.

Please ignore this, it's handled by patchwork and there's no way to obtain it.

> >  * :manpage:`fchroot(2) `was introduced in Linux v7.3.
> 
> Could the space be moved after the closing backtick? As written, it becomes
> part of the man-page role target.

will fix before merge.

> > 	off += snprintf(upwards + off, sizeof(upwards) - off, "/..");
> 
> Could truncation be handled here? LTP accepts an arbitrary absolute TMPDIR.
> With a sufficiently deep path, off grows past sizeof(upwards), the size
> subtraction underflows, and the next iteration passes an out-of-bounds
> pointer and a huge size to snprintf().

this can be fixed indeed.

> > 	SAFE_STAT(tst_tmpdir_path(), &st);
> > 	TST_EXP_EXPR(st.st_dev == realcwd.st_dev &&
> > 		st.st_ino == realcwd.st_ino,
> > 		"working directory restored after setns()");
> 
> Could this stat "." and compare it with realroot instead? mntns_install()
> sets both fs->pwd and fs->root to the mount namespace root. Calling stat on
> an absolute tmpdir path succeeds regardless of the current working directory,
> so a regression that leaves pwd unchanged would pass this check.

this is funny, it was "." before and it was complaining. I will use "." before
merge.

--
Andrea Cervesato
SUSE QE Automation Engineer Linux
andrea.cervesato@suse.com

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp