Re: compiling a bootable kernel for ia64 (itanium2, mckinley, rx2620)
On Wed, 31 Jan 2018, Dániel Kasza wrote:
I think I also ran into this issue on my RX2600 a few months ago.
Are you working on resolving this issue? Do you need any help with testing?
Yes, in some sense. We have tried to compile and boot a few kernels on
real hardware. (Don't you mind me replying to the debian-ia64 list, too,
so that the information gets to more people who may need it.)
Gleb glebfm@ has been doing the attempts; the last one being compiling the
squeeze mckinley kernel (which boots successfully for us with a newer
gcc-4.6 from wheezy. The result: it didn't boot.
On Wed, Jan 31, 2018 at 6:33 PM, Ivan Zakharyaschev <email@example.com>
Issue 711135 seems to never have been resolved yet (since Debian 6
"squeeze", which has the last kernel which boots on rx2620:
Maybe this works:
https://lists.debian.org/debian-ia64/2013/07/msg00009.html (Will Deacon):
Ok, after some more experimentation, this is looking more and more like
a compiler problem. Using 4.6.3, I *can* build a bootable kernel from the
Squeeze sources but only if I hack the kernel Makefile to pass -O1 instead
of -O2 or -Os.
Last night I learned the idea with -O1; so in the same build environment
that Gleb has created and with the same unpacked package sources (modified
to use gcc-4.6), I did the change for -O1:
root@rx2620:~# diff -du
--- /media/home/imz/linux-2.6-2.6.32/Makefile.orig_Os 2009-12-03
+++ /media/home/imz/linux-2.6-2.6.32/Makefile 2018-02-01
@@ -520,11 +520,7 @@
# Defaults vmlinux but it is usually overridden in the arch makefile
-KBUILD_CFLAGS += -Os
-KBUILD_CFLAGS += -O2
+KBUILD_CFLAGS += -O1
and started debuild -b -us -uc. The packages have just been built, but not
tested yet. Let's see (a bit later) what the results will be.