I am running a 2.4 kernel, so the failure I reported is apparently not a
bug. I also read somewhere that TLS requires package libc6-i686, which is
not installed on my system.
It would be nice if a diagnostic message was generated in this scenario,
instead of TLS failing silently.
--
To
Jeroen N. Witmond [EMAIL PROTECTED] writes:
Oops, my bad. In simplifying the testcase I removed an essential
part. In the new testcase (attached), the address of variable
'local' is taken in the thread. Unfortunately, this does not remove
the problem. The testcase still exits with 2 instead
On Thu, May 19, 2005 at 10:41:31PM +0200, Jeroen N. Witmond wrote:
Also, note that this is an unsound way to test for TLS. The compiler,
assembler, linker, C library, and on some platforms kernel must all
support it.
I am running debian/sarge (Linux DoornRoosje 2.4.26-1-386 #1 Tue
Also, note that this is an unsound way to test for TLS. The compiler,
assembler, linker, C library, and on some platforms kernel must all
support it.
I am running debian/sarge (Linux DoornRoosje 2.4.26-1-386 #1 Tue Aug 24
13:31:19 JST 2004 i686 GNU/Linux). The versions of gcc and libgcc1
Package: gcc-3.3
Version: 1:3.3.5-12
Severity: normal
File: /usr/bin/gcc
In the configure.in for the program I am working on, the reaction of
the compiler to the __thread keyword is used to determine the
availability of TLS. However, the compiler does not complain but the
variables marked with
5 matches
Mail list logo