|Summary:||Error occurs compiling kernel source|
|Product:||Red Hat Enterprise Linux 3||Reporter:||Rick Hester <rick.hester>|
|Component:||kernel||Assignee:||Anil S Keshavamurthy <akeshava>|
|Status:||CLOSED WONTFIX||QA Contact:||Brian Brock <bbrock>|
|Version:||3.0||CC:||alext, grgustaf, jbaron, kaccardi, luyu, michael.raskey, peterm, petrides, stephane.eranian, suzanne.pherigo|
|Fixed In Version:||Doc Type:||Bug Fix|
|Doc Text:||Story Points:||---|
|Last Closed:||2007-10-19 19:05:20 UTC||Type:||---|
|oVirt Team:||---||RHEL 7.3 requirements from Atomic Host:|
Description Rick Hester 2005-04-06 15:55:08 UTC
Description of problem: Error occurs during 'make compressed' step of re-compiling kernel source and compilation aborts. I don't have the text of the error, but I am trying to get that from the team reporting the error. Version-Release number of selected component (if applicable): Linux onyx.test 2.4.21-31.EL #1 SMP Thu Mar 10 19:36:07 EST 2005 ia64 ia64 ia64 GNU/Linux How reproducible: Steps to Reproduce: In the /usr/src/linux-2.4/.config 1.make mrproper 2.cp /boot/config-2.4.21-31.EL /usr/src/linux-2.4/.config 3.make menuconfig, remove the smp support feature and save the .config file. 4.make dep 5.make clean 6.make compressed - the error will occures. And the process of compile kernel aborts Actual results: Compilation of kernel build aborts Expected results: Kernel to build without errors. Additional info:
Comment 1 Ernie Petrides 2005-04-12 20:31:46 UTC
Steps #5-6 don't look right to me. This is what I do (working from a clean source tree, not necessarily from installation of kernel-source-*): cp configs/kernel-2.4.21-i686-smp.config .config make oldconfig make dep make bzImage make modules Please let me know if "make bzImage" works in place of "make clean" and "make compressed".
Comment 2 Tim Chambers 2005-04-18 21:08:14 UTC
[forwarding on behalf of email@example.com] I think the steps that I provide are right, because on the IA64 platform the command should be "make compressed": In the /usr/src/linux-2.4/.config 1.make mrproper 2.cp /boot/config-2.4.21-31.EL /usr/src/linux-2.4/.config 3.make menuconfig, remove the smp support feature and save the .config file. 4.make dep 5.make clean 6.make compressed the error will occure. And the process of compile kernel abort. I didn't use the make oldconfig because I want to change some options. In this case I removed the smp support option. The make bzImage should be used on IA32 platform.
Comment 3 Ernie Petrides 2005-04-19 21:27:20 UTC
Hello, Tim. You are quite right. I didn't originally notice that this was an ia64 kernel build. The build error is as follows: perfmon.c:3491: conflicting types for `pfm_lazy_save_regs' perfmon.c:464: previous declaration of `pfm_lazy_save_regs' perfmon.c: In function `pfm_load_regs': perfmon.c:3750: `this_cpu' undeclared (first use in this function) perfmon.c:3750: (Each undeclared identifier is reported only once perfmon.c:3750: for each function it appears in.) make: *** [perfmon.o] Error 1 This problem was caused by the enabling of the perfmon() system call for ia64 in U5, with a patch to fix the code submitted by Stephane Eranian of HP (now added to the cc: list). Strictly speaking, we don't support non-SMP configurations on ia64. But if you folks could work with Stephane to develop and test an incremental ia64 perfmon patch that does work on non-SMP configs, then we'd be willing to take it into U6. Otherwise, we'd have to close this bug as WONTFIX. For now, I'll leave this bug in NEEDINFO state. Please let us know which way you'd like to go. If you'd like this problem fixed, please have Stephane attach a patch tested against the latest U5 to this bug. Thanks. -ernie
Comment 4 Ernie Petrides 2005-04-19 21:30:31 UTC
I forgot to mention that the ia64 perfmon update was on behalf of bug 112988.
Comment 5 Tim Chambers 2005-04-19 22:22:57 UTC
Adding Suzanne to the list. I know of no customer requirements that would motivate HP to support non-SMP configurations of ia64. I propose we close this. Suzanne, what do you say?
Comment 6 Stephane Eranian 2005-04-20 09:16:03 UTC
If you point me at the sources for kernel-2.4.21-31 I could take a quick look but as you said, I don't think UP configs are ever used.
Comment 7 Ernie Petrides 2005-04-20 19:15:30 UTC
Stephane, please get these through the RHN beta channels. The latest RPMs there are for 2.4.21-31.EL, but there was another respin Friday (for -32.EL), which should appear in the beta channels soon.
Comment 8 Tim Chambers 2005-04-25 17:40:32 UTC
I assume "get these through the RHN beta channels" means we should simply subscribe a system to the beta channel and use up2date to download the RPM? If not, please explain how you would do it. I tried that today, and -31.EL is still the latest that I see. Can you be a little more specific about "soon?" Even better, could you just put the respin somwhere where Stephane could get it right away?
Comment 9 Ernie Petrides 2005-04-25 18:07:08 UTC
You can download RPMs directly via rhn.redhat.com (although you'd need a valid login), and partners should already have been informed how to download RPMs directly from partners.redhat.com. Using RHN, just choose the appropriate RHEL v. 3 (arch) beta channel and then select packages (probably want to use "kernel-" as a filter). You can probably use the following URL (once logged in) to isolate the "kernel-source-2.4.21-31.EL.i386.rpm" RPM and then select "Download RPM" near the bottom: https://rhn.redhat.com/network/software/packages/details.pxt?pid=301945 I've just found out that the -32.EL U5 beta kernel is likely to be loaded into the associated RHN beta channels later today. (It's already on the Partners' FTP site.)
Comment 10 Ernie Petrides 2005-09-12 19:26:55 UTC
Stephane, are you going to follow up on this at some point?
Comment 11 Stephane Eranian 2005-09-16 12:40:20 UTC
Ernie, I will take a look at this on Monday. But I thought you said it does not really matter because you do not ship UP kernels.
Comment 12 Ernie Petrides 2005-09-16 19:24:29 UTC
Stephane, you are correct: fixing this is not a requirement of Red Hat's. But as I wrote in comment #3, if you'd like to help out your HP colleague Rick Hester by fixing this, then we are open to applying a fix of yours to the current Update in progress (which is now U7). If not, just let me know and I'll close this bug as WONTFIX.
Comment 13 Stephane Eranian 2005-09-19 09:08:29 UTC
Created attachment 118964 [details] Fix perfmon UP compile Ernie, Here is a patch that should fix your compilation problem in UP mode. It is relative to the U6 kernel source.
Comment 14 Ernie Petrides 2005-09-19 18:33:25 UTC
Thanks, Stephane. Geoff, when you test this, please make sure it compiles in both the normal (SMP) build and in the re-config'ed non-SMP build (as described in the initial comment of this bug report).
Comment 15 Michael Raskey 2006-03-03 18:15:08 UTC
Tried to compile the kernel on RHEL3 U7 beta, and the following error occured: gcc -D__KERNEL__ -I/usr/src/linux-2.4.21-38.EL/include -Wall -Wstrict- prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -Wno-unused - fomit-frame-pointer -pipe -ffixed-r13 -mfixed-range=f10-f15,f32-f127 -falign- functions=32 -frename-registers --param max-inline-insns=5000 -mconstant-gp - nostdinc -iwithprefix include -DKBUILD_BASENAME=ia64_ksyms - DEXPORT_SYMTAB -c ia64_ksyms.c ia64_ksyms.c:210: `pm_power_off' undeclared here (not in a function) ia64_ksyms.c:210: initializer element is not constant ia64_ksyms.c:210: (near initialization for `__ksymtab_pm_power_off.value') make: *** [ia64_ksyms.o] Error 1 make: Leaving directory `/usr/src/linux-2.4.21-38.EL/arch/ia64/kernel' make: *** [_dir_arch/ia64/kernel] Error 2 [root@rx2600-2 linux-2.4]# uname -a Linux rx2600-2 2.4.21-39.EL #1 SMP Fri Jan 20 18:40:17 EST 2006 ia64 ia64 ia64 GNU/Linux Was the patch applied? Was is supposed to make it into Update 7?
Comment 16 Ernie Petrides 2006-03-03 20:37:24 UTC
This bug was recently reassigned from Geoff to Anil. No patch was ever posted internally, and so no, the fix did not make it into U7.
Comment 17 Anil S Keshavamurthy 2006-03-03 21:23:48 UTC
In my view, these kinds of fixes should not be encouraged. Once we fix this particular compile issues, we might see some other compile bug for UP and it is not worth atacking all those bugs. Since this UP configuration is totally unsupported by Red Hat and their could very well be boot or stability issues too with the built UP kernel. Unless the priority of this bug is raised high, I will not going to look at this bug for now.
Comment 18 RHEL Product and Program Management 2007-10-19 19:05:20 UTC
This bug is filed against RHEL 3, which is in maintenance phase. During the maintenance phase, only security errata and select mission critical bug fixes will be released for enterprise products. Since this bug does not meet that criteria, it is now being closed. For more information of the RHEL errata support policy, please visit: http://www.redhat.com/security/updates/errata/ If you feel this bug is indeed mission critical, please contact your support representative. You may be asked to provide detailed information on how this bug is affecting you.