Re: 2.4.12 compile errors on Sparc32
Update: On Thu, Oct 25, 2001 at 07:18:54PM +0200, Peter Keel wrote: > On Mon, Oct 22, 2001 at 06:41:22AM -0400, Joe Stevens wrote: > > Yea, the recent CVS versions should work, I only keep that 2.4.8 up there > > because its 'stable' (I use it on a number of production machines, uptimes > > of more than a month...) > > Ok. Got it. 2.4.13-pre3, rsync'ed from the cvs at http://vger.samba.org/ > WORKS. On a Sparc10: > > ncpus probed: 2 > ncpus active: 2 > Cpu0Bogo: 74.75 > Cpu1Bogo: 74.95 It works, somehow. But about every two days, I get a spinlock- deadlock between the two processors. This _might_ be a problem related to my disks/controller, since in 2.2.19 I loose disks sometimes (disk can't be accessed anymore, not even with probe-scs-all, until I turn of the machine and turn it on again). Then again, this might be an issue with the kernel. I'll try it on a single-cpu Sparc 10 the next few days. Peter -- Those who give up essential liberties for temporary safety deserve neither liberty nor safety. -- Benjamin Franklin
More 2.4.12 compile errors on Sparc32
On Thu, 25 Oct 2001, Peter Keel wrote: > Ok. Got it. 2.4.13-pre3, rsync'ed from the cvs at http://vger.samba.org/ > WORKS. On a Sparc10: Hm, I try to compile kernel 2.4.13 and the kernel from vger. Both compile till the end, then following error is reported: objdump -x vmlinux.o | ./btfixupprep > btfix.s Wrong use of 'disable_irq' in '.text.exit' section. It can be only used in .text, .text.init, .fixup and __ksymtab make[1]: *** [btfix.s] Error 1 make[1]: Leaving directory `/home/andras/src/vger/linux/arch/sparc/boot' make: *** [_dir_arch/sparc/boot] Error 2 What's the problem? The system is a SparcStation 5 with Debian unstable (upgraded from potato). GCC and binutils should be up to date. I experienced this on all kernels >2.4.2 if there was not another problem ;) Thanks, Andras
Re: 2.4.12 compile errors on Sparc32
On Mon, Oct 22, 2001 at 06:41:22AM -0400, Joe Stevens wrote: > Yea, the recent CVS versions should work, I only keep that 2.4.8 up there > because its 'stable' (I use it on a number of production machines, uptimes > of more than a month...) Ok. Got it. 2.4.13-pre3, rsync'ed from the cvs at http://vger.samba.org/ WORKS. On a Sparc10: cpu : Texas Instruments, Inc. - SuperSparc-(II) fpu : SuperSparc on-chip FPU promlib : Version 3 Revision 2 prom: 2.22 type: sun4m ncpus probed: 2 ncpus active: 2 Cpu0Bogo: 74.75 Cpu1Bogo: 74.95 MMU type: TI Viking/MXCC contexts: 65536 nocache total : 1048576 nocache used: 772864 CPU0: online CPU1: online Whooohooo! > On Mon, 22 Oct 2001, Peter Keel wrote: > > On Sun, Oct 21, 2001 at 01:22:39AM -0400, Joe Stevens wrote: > > > I have a complete source tree of 2.4.8 that I have patched up to work on > > > sparc32 available at http://home.att.ne.jp/iota/joe/ . It have it running > > > on a number of Sparc 5s. > > > > I managed to compile the kernel from the cvs -- cvs-access was impossible > > the user/password mentionned on http://vger.samba.org/ don't match, but > > rsync works. I haven't tested it yet, since I'm always reluctant to boot > > a machine. I will report back if anything interesting happens. Cheers Peter -- Those who give up essential liberties for temporary safety deserve neither liberty nor safety. -- Benjamin Franklin
Re: 2.4.12 compile errors on Sparc32
Yea, the recent CVS versions should work, I only keep that 2.4.8 up there because its 'stable' (I use it on a number of production machines, uptimes of more than a month...) Joe. On Mon, 22 Oct 2001, Peter Keel wrote: > On Sun, Oct 21, 2001 at 01:22:39AM -0400, Joe Stevens wrote: > > > > I have a complete source tree of 2.4.8 that I have patched up to work on > > sparc32 available at http://home.att.ne.jp/iota/joe/ . It have it running > > on a number of Sparc 5s. > > I managed to compile the kernel from the cvs -- cvs-access was impossible > the user/password mentionned on http://vger.samba.org/ don't match, but > rsync works. I haven't tested it yet, since I'm always reluctant to boot > a machine. I will report back if anything interesting happens. > > Peter >
Re: 2.4.12 compile errors on Sparc32
On Sun, Oct 21, 2001 at 01:22:39AM -0400, Joe Stevens wrote: > > I have a complete source tree of 2.4.8 that I have patched up to work on > sparc32 available at http://home.att.ne.jp/iota/joe/ . It have it running > on a number of Sparc 5s. I managed to compile the kernel from the cvs -- cvs-access was impossible the user/password mentionned on http://vger.samba.org/ don't match, but rsync works. I haven't tested it yet, since I'm always reluctant to boot a machine. I will report back if anything interesting happens. Peter -- Those who give up essential liberties for temporary safety deserve neither liberty nor safety. -- Benjamin Franklin
Re: 2.4.12 compile errors on Sparc32
>... 2.4 is supposed to be a stable kernel, so the first thing I'd expect >is that it compiles; and that it won't break compilation on a complete >architecture every other version. The trouble is: There is no maintainer for sparc32 on 2.4. Do you want to jump in? There seem to be enough people who'd really appreciate a stable 2.4. kernel on this architecture - alas none of us is a kernel guru. Fortunately from time to time there kind souls have provided that got us a little further each time (thanks DaveM, Anton, Uzi, Collin and all those I was so unkind to forget). You'll find these patches in the archives of debian-sparc, sparcliunx and [EMAIL PROTECTED] Good luck! Ingo
2.4.12 compile errors on Sparc32
Indeed, 2.4.11 compiled on Sparc32, the first since a long time. I couldn't run the one one I tried because it told me it was too big. Before I could shrink it, 2.4.12 appeared, and of course: /usr/src/linux-2.4.12/include/linux/interrupt.h:77: warning: `__cpu_raise_softirq' redefined /usr/src/linux-2.4.12/include/asm/softirq.h:31: warning: this is the location of the previous definition In file included from /usr/src/linux-2.4.12/include/linux/pagemap.h:16, from /usr/src/linux-2.4.12/include/linux/locks.h:8, from /usr/src/linux-2.4.12/include/linux/devfs_fs_kernel.h:6, from init/main.c:16: /usr/src/linux-2.4.12/include/linux/highmem.h: In function `clear_user_highpage': /usr/src/linux-2.4.12/include/linux/highmem.h:48: `KM_USER0' undeclared (first use in this function) /usr/src/linux-2.4.12/include/linux/highmem.h:48: (Each undeclared identifier is reported only once /usr/src/linux-2.4.12/include/linux/highmem.h:48: for each function it appears in.) /usr/src/linux-2.4.12/include/linux/highmem.h: In function `copy_user_highpage': /usr/src/linux-2.4.12/include/linux/highmem.h:89: `KM_USER0' undeclared (first use in this function) /usr/src/linux-2.4.12/include/linux/highmem.h:90: `KM_USER1' undeclared (first use in this function) init/main.c: In function `smp_init': init/main.c:515: warning: implicit declaration of function `cpu_relax' make: *** [init/main.o] Error 1 So what would be course of action? I tried 2.4.13-pre3, and 2.4.12-ac4. Same result. ... 2.4 is supposed to be a stable kernel, so the first thing I'd expect is that it compiles; and that it won't break compilation on a complete architecture every other version. If there are hardware-issues, that's one thing, but _compilation_ issues, that's a no-no. If I ever meet Linus, he'll be getting a brown paper bag from me... . Thank you Peter -- Those who give up essential liberties for temporary safety deserve neither liberty nor safety. -- Benjamin Franklin