Bug 11799 - si_code is not SI_USER on raise()
Summary: si_code is not SI_USER on raise()
Status: RESOLVED FIXED
Alias: None
Product: glibc
Classification: Unclassified
Component: nptl (show other bugs)
Version: 2.12
: P2 minor
Target Milestone: ---
Assignee: Ulrich Drepper
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-07-08 18:22 UTC by Jan Kratochvil
Modified: 2014-06-30 17:32 UTC (History)
2 users (show)

See Also:
Host:
Target: x86_64-fedora13-linux-gnu
Build:
Last reconfirmed:
fweimer: security-

Attachments
Reproducer. (386 bytes, text/plain)
2010-07-08 18:22 UTC, Jan Kratochvil 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Jan Kratochvil 2010-07-08 18:22:02 UTC 
si_code is SI_TKILL in such case but:
./sysdeps/unix/sysv/linux/bits/siginfo.h:
  SI_USER,			/* Sent by kill, sigsend, raise.  */

kernel-2.6.33.5-124.fc13.x86_64
glibc-2.12-2.x86_64 and HEAD (5be9d05d3136460cf6cedd3823fd236aeebe83ef)

Just noticed, unaware of any problem due to it.
Comment 1 Jan Kratochvil 2010-07-08 18:22:26 UTC 
Created attachment 4876 [details]
Reproducer.
Comment 2 Ulrich Drepper 2011-05-30 00:06:23 UTC 
That is no problem except in the documentation.  No standard specifies anything about the si_code used for raise.  I changed the documentation.
Comment 3 Michael Kerrisk 2012-04-20 01:18:23 UTC 
Around glibc 2.3.2/2.3.3, the implementation of raise() changed, resulting in a silent ABI change to the si_code behavior. Once upon a time, the code comment was true.