MySQL Bugs: #10712: CONFIG_NR_CPUS not defined for kernel header percpu.h

Bug #10712	CONFIG_NR_CPUS not defined for kernel header percpu.h
Submitted:	18 May 2005 14:28	Modified:	19 Jun 2005 9:09
Reporter:	Dmitry Lunin	Email Updates:
Status:	No Feedback	Impact on me:	None
Category:	MySQL Server: Compiling	Severity:	S3 (Non-critical)
Version:	4.1.12	OS:	Linux (Slackware Linux 10.1)
Assigned to:		CPU Architecture:	Any

Description:
Slackware Linux 10.1, gcc 3.3.4 (or gcc 3.3.5), kernel 2.6.10

gcc -c  -o strings-x86.o strings-x86.s
gcc -c  -o longlong2str-x86.o longlong2str-x86.s
gcc -c  -o my_strtoll10-x86.o my_strtoll10-x86.s
if gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../include    -O3 -DDBUG_OFF    -MT bfill.o -MD -MP -MF ".deps/bfill.Tpo" -c -o bfill.o bfill.c; \
then mv -f ".deps/bfill.Tpo" ".deps/bfill.Po"; else rm -f ".deps/bfill.Tpo"; exit 1; fi
In file included from /usr/include/asm/percpu.h:4,
                 from /usr/include/asm/processor.h:22,
                 from /usr/include/asm/atomic.h:6,
                 from ../include/my_global.h:294,
                 from bfill.c:32:
/usr/include/asm-generic/percpu.h:8: error: `CONFIG_NR_CPUS' undeclared here (not in a function)
make[2]: *** [bfill.o] Error 1
make[2]: Leaving directory `/z/src/mysql-4.1.12/strings'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/z/src/mysql-4.1.12'
make: *** [all] Error 2

How to repeat:
./configure --prefix=/usr --with-mysqld-user=mysql --with-unix-socket-path=/var/run/mysql/mysql.sock --localstatedir=/var/lib/mysql --enable-assembler --with-raid --without-debug --enable-thread-safe-client --without-bench --with-extra-charsets=complex --program-prefix="" --program-suffix="" i686-slackware-linux

make

Are you sure it's 4.1.12 ?
This bug was fixed in 4.1.12 - see http://bugs.mysql.com/bug.php?id=7970

Could you attach ([Files] tab) config.log ?

Yes. I am assured. MySQL 4.1.12.
This link saw.

Config Log

Attachment: config.log.gz (application/x-gzip, text), 31.18 KiB.

The file is attached

Unfirtunately, we don't have a configuration similar to yours and cannot repeat this bug.

So, you need to help us to find the reason.

configure contains a check whether <asm/atomic.h> is usable - it tries to compile and run a small program using atomic operations. But in your case configure succeeds, but later compilation of MySQL fails. And we need to know why configure doesn't see the problem.

What you can try to do:

1. check how configure does the test (e.g. by putting "set -x" in configure script right above atomic test and "set +x" when the test is finished)
2. perform the same test manually (create the same test program, compile it with the same switches)
3. check the preprocessed source of this test program (gcc <all other switches> -E -dD)
4. check the preprocessed source of bfill.c (gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../include    -O3 -DDBUG_OFF -E -dD bfill.c)

at this point you'll be able to see why CONFIG_NR_CPUS is a problem for the main build but not for the configure.

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".