This is the mail archive of the
mailing list for the glibc project.
Re: [PATCH 3/4] nptl: Fix testcases for new pthread cancellation, mechanism
- From: Adhemerval Zanella <azanella at linux dot vnet dot ibm dot com>
- To: "Joseph S. Myers" <joseph at codesourcery dot com>
- Cc: "GNU C. Library" <libc-alpha at sourceware dot org>
- Date: Mon, 22 Sep 2014 10:14:53 -0300
- Subject: Re: [PATCH 3/4] nptl: Fix testcases for new pthread cancellation, mechanism
- Authentication-results: sourceware.org; auth=none
- References: <541C2901 dot 1050609 at linux dot vnet dot ibm dot com> <541C2B18 dot 4090107 at linux dot vnet dot ibm dot com> <541C2DDA dot 6050204 at linux dot vnet dot ibm dot com> <Pine dot LNX dot 4 dot 64 dot 1409191659440 dot 11496 at digraph dot polyomino dot org dot uk>
Hi Joseph, thanks for the review.
On 19-09-2014 14:02, Joseph S. Myers wrote:
> On Fri, 19 Sep 2014, Adhemerval Zanella wrote:
>> This patches fixes the NPTL testcase that assumes the old behavior and
> Do the fixes result in tests that work both with and without the main
> changes? If so, they should go first. (In general, for bisectability
> either you fix tests first to work both with and without the main changes,
> or the test fixes need to go in the same commit as the main changes even
> if posted separately for convenience.)
The debug/tst-backtrace5.c is adjusted to take in consideration the new stack chain
call for syscalls. The tst-cancelX modification work on both schemes (the thread
in cancelled in the syscall itself).
>> also remove the tst-cancel-wrappers.sh test (which checks for symbols
>> that does not exist anymore).
> Is there a better way of achieving the substance of this test (i.e.
> verifying that all those functions are cancellable) that still works after
> the patch?
I will double check, but later time I checked all the cancelable entry point in
GLIBC are already being fairly tested by other nptl/tst-cancelX cases.