[Linux-ia64] [[email protected]: Re: kernel make install - why don't we make it easy like on ia32?]
How do I build:
/boot/module-info-${KERNELRELEASE}
/boot/kernel.h-${KERNELRELEASE}
/boot/initrd-${KERNELRELEASE}In upgrading from RedHat ia64 2.4.9-18 to 2.4.17 and 2.4.18 may require changes in these files. I thought linux/arch/i386/boot/Makefile would have been a resonable place to have found out how i32 does it but it's not there.
I had problems last time I tried to run 2.4.17 on my 2.4.7-10 Red Hat 7.2 i386 systems. I was wondering of the /boot/module-info file had to be updated or that the stuff in /lib/modules that 'Make modules-install' sets up might need to be tweaked.
-piet
Quoted 6 lines · Piet/Pete Delaney
How do I build:
/boot/module-info-${KERNELRELEASE} /boot/kernel.h-${KERNELRELEASE} /boot/initrd-${KERNELRELEASE}
Quoted 7 lines
In upgrading from RedHat ia64 2.4.9-18 to 2.4.17 and 2.4.18 may require changes in these files. I thought linux/arch/i386/boot/Makefile would have been a resonable place to have found out how i32 does it but it's not there.
I had problems last time I tried to run 2.4.17 on my 2.4.7-10 Red Hat 7.2 i386 systems. I was wondering of the
/boot/module-infofile had to be updated or that the stuff in/lib/modulesthat 'Make modules-install' sets up might need to be tweaked.
The module problem I'm getting in trying to boot a generic 2.4.17 kernel is with /lib/ext3.0. It's getting a lot of undefined symbols like:
getblk_Rsmp_0fdb96c8 journal_destroy_Rsmp_e6169b31 register_filesystem_Rsmp_462b6e36 __wait_on_buffer_Rsmp_e7187873 generic_read_dir_Rsmp_... journal_try_start_Rsmp_... journal_check_available_features_Rsmp_...
It fails in /bin/insmod in mounting the /proc filesystem.
Then it fails to mount the (ext3) root filesystem.
The last time I had this on a ia32 laptop I thought the problem was with the root being a ext3 filesystem but the ext3 filesystem being a loadable module. So I updated the kernel .config file to make ext3 a part of the kernel instead of a loadable module. I re-installed the modules just to be safe and the kernel continued to get the same error. The only thing I didn't do is to do a make mrproper. I'll try that this time.
Any suggestions?
-piet
Quoted 30 lines · Piet/Pete Delaney
Quoted 6 lines · Piet/Pete Delaney
How do I build:
/boot/module-info-${KERNELRELEASE} /boot/kernel.h-${KERNELRELEASE} /boot/initrd-${KERNELRELEASE}Quoted 7 lines
In upgrading from RedHat ia64 2.4.9-18 to 2.4.17 and 2.4.18 may require changes in these files. I thought linux/arch/i386/boot/Makefile would have been a resonable place to have found out how i32 does it but it's not there.
I had problems last time I tried to run 2.4.17 on my 2.4.7-10 Red Hat 7.2 i386 systems. I was wondering of the
/boot/module-infofile had to be updated or that the stuff in/lib/modulesthat 'Make modules-install' sets up might need to be tweaked.The module problem I'm getting in trying to boot a generic 2.4.17 kernel is with
/lib/ext3.0. It's getting a lot of undefined symbols like:getblk_Rsmp_0fdb96c8 journal_destroy_Rsmp_e6169b31 register_filesystem_Rsmp_462b6e36 __wait_on_buffer_Rsmp_e7187873 generic_read_dir_Rsmp_... journal_try_start_Rsmp_... journal_check_available_features_Rsmp_...It fails in
/bin/insmodin mounting the /proc filesystem.Then it fails to mount the (ext3) root filesystem.
Quoted 7 lines
The last time I had this on a ia32 laptop I thought the problem was with the root being a ext3 filesystem but the ext3 filesystem being a loadable module. So I updated the kernel .config file to make ext3 a part of the kernel instead of a loadable module. I re-installed the modules just to be safe and the kernel continued to get the same error. The only thing I didn't do is to do a make mrproper. I'll try that this time.
Quoted 4 lines · Piet/Pete Delaney
How do I build:
/boot/module-info-${KERNELRELEASE} /boot/kernel.h-${KERNELRELEASE}
Redhat special files, not part of the kernel build.
Quoted 1 line
/boot/initrd-${KERNELRELEASE}mkinitrd, if you really insist. initrd is useful for distributors and when you need a binary only module at boot time. Otherwise initrd is a pain in the neck, build boot drivers into the kernel instead.