/usr/vac/bin/cc -c -DHAVE_CONFIG_H -g -I. -I/opt/build/gcc-3.3/libiberty/../include /opt/build/gcc-3.3/libiberty/physmem.c -o physmem.o "/opt/build/gcc-3.3/libiberty/physmem.c", line 149.32: 1506-022 (S) "physmem" is not a member of "struct {...}". gmake[1]: *** [physmem.o] Error 1 gmake[1]: Leaving directory `/opt/build/gcc-3.3-objdir/libiberty' On AIX 5.1, the _system_configuration struct in <sys/systemcfg.h> does have the physmem member: long long physmem; /* bytes of OS available memory */ However, on 4.3.2, it doesn't exist (it exists on AIX 4.3.3 and above). And, I don't see anything else in _system_configuration to get physical memory.
Created attachment 4071 [details] Patch to correctly set HAVE__SYSTEM_CONFIGURATION if physmem member exists
David, I believe this is you area of expertise. Would it be possible for you to take a look at the patch here and check whether this is a correct fix to the problem? Thanks, Dara
You should sent your patch to gcc-patches@gcc.gnu.org, with a descripition of the bug.
Subject: Re: [Bug bootstrap/10974] Bootstrap failure Kaveh, Would you plesae consider updating the libiberty physical memory check on AIX with this patch? David
Subject: Re: Bootstrap failure > From: David Edelsohn <dje@watson.ibm.com> > > Kaveh, > Would you plesae consider updating the libiberty physical memory > check on AIX with this patch? > David Sure. Though I don't have any AIX boxes at all for testing it. Can I assume that you verified physmem is still correctly activated on AIX4.3.3 and AIX5? Thanks, --Kaveh -- Kaveh R. Ghazi ghazi@caip.rutgers.edu
Subject: Re: Bootstrap failure >>>>> Kaveh R Ghazi writes: Kaveh> Sure. Though I don't have any AIX boxes at all for testing it. Can I Kaveh> assume that you verified physmem is still correctly activated on Kaveh> AIX4.3.3 and AIX5? The patch tests for the existence of the structure member, not a specific AIX version. Why are you asking about AIX version numbers? David
Subject: Re: Bootstrap failure > From: David Edelsohn <dje@watson.ibm.com> > > >>>>> Kaveh R Ghazi writes: > > Kaveh> Sure. Though I don't have any AIX boxes at all for testing it. Can I > Kaveh> assume that you verified physmem is still correctly activated on > Kaveh> AIX4.3.3 and AIX5? > > The patch tests for the existence of the structure member, not a > specific AIX version. Why are you asking about AIX version numbers? > David I want to know if someone had tested the patch in both the negative *and* the positive case of whether the structure member exists. I assume albert checked the negative case on his 4.3.2 box. I listed the AIX version numbers of where I thought the positive case lies. -- Kaveh R. Ghazi ghazi@caip.rutgers.edu
I tested on AIX 5.1 but not 4.3.3.
Subject: Re: Bootstrap failure >>>>> ghazi@caip rutgers edu writes: ghazi> I want to know if someone had tested the patch in both the negative ghazi> *and* the positive case of whether the structure member exists. I ghazi> assume albert checked the negative case on his 4.3.2 box. I listed ghazi> the AIX version numbers of where I thought the positive case lies. The patch looks correct to me. It tests assigning the structure member to a double instead of a "long long", but that is what the function in libiberty/physmem.c returns. David
Hello, the patch for this doesn't yet seem to have been committed yet. Is the problem that nobody has tested it yet on AIX 4.3.3? Thanks, Dara
Subject: Re: Bootstrap failure >>>>> dhazeghi@yahoo com writes: Dara> the patch for this doesn't yet seem to have been committed yet. Is the problem that nobody has Dara> tested it yet on AIX 4.3.3? I do not know why Kaveh has not merged the patch yet. David
Subject: Re: Bootstrap failure > From: David Edelsohn <dje@watson.ibm.com> > > >>>>> dhazeghi@yahoo com writes: > > Dara> the patch for this doesn't yet seem to have been committed yet. Is the problem that nobody has > Dara> tested it yet on AIX 4.3.3? > > I do not know why Kaveh has not merged the patch yet. > David For some reason, the followups to my question about testing didn't get cc'ed to me directly. So I never saw the replies until I checked the PR audit trail just now. I'm convinced by the replies, so I'll check in the patch tomorrow (Sunday) on mainline and 3.3 branch. -- Kaveh R. Ghazi ghazi@caip.rutgers.edu
Waiting on Kaveh Ghazi to check it in.
Subject: Bug 10974 CVSROOT: /cvs/gcc Module name: gcc Changes by: ghazi@gcc.gnu.org 2003-06-09 05:20:02 Modified files: libiberty : ChangeLog physmem.c configure.in config.in configure Log message: PR bootstrap/10974 * physmem.c: Update comment. * configure.in: Modify test for _system_configuration for older AIX systems. * config.in, configure: Regenerated. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libiberty/ChangeLog.diff?cvsroot=gcc&r1=1.446&r2=1.447 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libiberty/physmem.c.diff?cvsroot=gcc&r1=1.10&r2=1.11 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libiberty/configure.in.diff?cvsroot=gcc&r1=1.68&r2=1.69 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libiberty/config.in.diff?cvsroot=gcc&r1=1.30&r2=1.31 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libiberty/configure.diff?cvsroot=gcc&r1=1.68&r2=1.69
fix installed