View previous topic :: View next topic |
Author |
Message |
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
Posted: Wed Feb 17, 2010 3:16 am Post subject: |
|
|
Thaidog wrote: | ebrsc wrote: | here's what I get:
Code: | ipo: warning #11009: file format not recognized for arch/x86/vdso/vdso32/note.o
ipo: warning #11009: file format not recognized for arch/x86/vdso/vdso32/int80.o
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: cannot find -lirc_s
make[2]: *** [arch/x86/vdso/vdso32-int80.so.dbg] Error 1
make[1]: *** [arch/x86/vdso] Error 2
make: *** [arch/x86] Error 2 |
disclaimer: first time ever that I use that icc stuff, so I'm super noob on the subject and may have done something extremely dumb.
linux 2.6.32.8
dna-2.6.32.8-intel64.patch
icc 11.1.064-1
on x86_64 machine
/edit: just saw this http://groups.google.com/group/linuxdna/browse_thread/thread/2cbf65f5303104a7?pli=1
retrying with 32bit-compiler installed as well. |
A few more questions:
What cpu are you using? What command do you use to source the compiler? |
cpu: Intel(R) Core(TM)2 Duo CPU T7300
command:
Code: | source /opt/intel/Compiler/11.1/064/bin/iccvars.sh intel64 |
one question on my side: is this [1] still needed ? if yes, I guess it should point to the 32bit version libirc_s.a, is it right ?
also, just to give more input, I tried to build linux-dna with a PKGBUILD (arch linux package building script) using:
Code: | make bzImage modules || return 1 |
and it fails the same way as my former post. if I use:
the same warnings appear during the build but the build is succesful but all I get is a nice kernel panic.
[1] http://www.linuxdna.com/changeicc.html |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Wed Feb 17, 2010 4:42 am Post subject: |
|
|
Our latest patches are for 64bit kernels only. Also [1] should not be needed. Nor should you have to use the wrapper script. Also you should be able to invoke the build by simply issuing "make"... and not need to point anything to the wrapper. (The Gentoo install section on the mirror is only for the ICC version that is package.keyword masked in portage - more than likely you can not build any current kernels with that version)
Simply source the icc compiler (like you have been doing with intel64) patch the kernel, run make menuconfig and then finally make. _________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
Posted: Wed Feb 17, 2010 5:15 am Post subject: |
|
|
Thaidog wrote: | Simply source the icc compiler (like you have been doing with intel64) patch the kernel, run make menuconfig and then finally make. |
ok, I started from scratch and did exactly this. the only machine I had available has a AMD Athlon(tm) 7750 cpu. I suppose it's not really smart to use an amd cpu but given the fact I get exactly the same error as with the C2D, that should be unrelated. here's what happens:
Code: | VDSO arch/x86/vdso/vdso32-int80.so.dbg
ipo: warning #11009: file format not recognized for arch/x86/vdso/vdso32/note.o
ipo: warning #11009: file format not recognized for arch/x86/vdso/vdso32/int80.o
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: cannot find -lirc_s
make[2]: *** [arch/x86/vdso/vdso32-int80.so.dbg] Error 1
make[1]: *** [arch/x86/vdso] Error 2
make: *** [arch/x86] Error |
|
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Wed Feb 17, 2010 6:00 am Post subject: |
|
|
ebrsc wrote: | Thaidog wrote: | Simply source the icc compiler (like you have been doing with intel64) patch the kernel, run make menuconfig and then finally make. |
ok, I started from scratch and did exactly this. the only machine I had available has a AMD Athlon(tm) 7750 cpu. I suppose it's not really smart to use an amd cpu but given the fact I get exactly the same error as with the C2D, that should be unrelated. here's what happens:
Code: | VDSO arch/x86/vdso/vdso32-int80.so.dbg
ipo: warning #11009: file format not recognized for arch/x86/vdso/vdso32/note.o
ipo: warning #11009: file format not recognized for arch/x86/vdso/vdso32/int80.o
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: skipping incompatible /opt/intel/Compiler/11.1/064/lib/intel64/libirc_s.a when searching for -lirc_s
ld: cannot find -lirc_s
make[2]: *** [arch/x86/vdso/vdso32-int80.so.dbg] Error 1
make[1]: *** [arch/x86/vdso] Error 2
make: *** [arch/x86] Error |
|
AMD's will not work out of the box... there are some thing's that need to be patched. Do you have ia32 emulation enabled? It could be the issue. Add this to the bottom of the patch for that:
Code: |
diff -Nur linux-2.6.32.7/arch/x86/vdso/Makefile
linux-dna-2.6.32.7/arch/x86/vdso/Makefile
--- linux-2.6.32.7/arch/x86/vdso/Makefile 2010-01-29 07:06:20.000000000 +0800
+++ linux-dna-2.6.32.7/arch/x86/vdso/Makefile 2010-02-09
11:39:39.053639278 +0800
@@ -118,7 +118,7 @@
# The DSO images are built using a special linker script.
#
quiet_cmd_vdso = VDSO $@
- cmd_vdso = $(CC) -nostdlib -o $@ \
+ cmd_vdso = gcc -nostdlib -o $@ \
$(VDSO_LDFLAGS) $(VDSO_LDFLAGS_$(filter %.lds,$(^F))) \
-Wl,-T,$(filter %.lds,$^) $(filter %.o,$^)
|
_________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
|
Back to top |
|
|
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
Posted: Wed Feb 17, 2010 7:16 am Post subject: |
|
|
Thaidog wrote: | Do you have ia32 emulation enabled? It could be the issue. |
Code: | $zcat /proc/config.gz | grep IA32_EMU
CONFIG_IA32_EMULATION=y |
Thaidog wrote: | Code: |
diff -Nur linux-2.6.32.7/arch/x86/vdso/Makefile
linux-dna-2.6.32.7/arch/x86/vdso/Makefile
--- linux-2.6.32.7/arch/x86/vdso/Makefile 2010-01-29 07:06:20.000000000 +0800
+++ linux-dna-2.6.32.7/arch/x86/vdso/Makefile 2010-02-09
11:39:39.053639278 +0800
@@ -118,7 +118,7 @@
# The DSO images are built using a special linker script.
#
quiet_cmd_vdso = VDSO $@
- cmd_vdso = $(CC) -nostdlib -o $@ \
+ cmd_vdso = gcc -nostdlib -o $@ \
$(VDSO_LDFLAGS) $(VDSO_LDFLAGS_$(filter %.lds,$(^F))) \
-Wl,-T,$(filter %.lds,$^) $(filter %.o,$^)
|
|
This patch solved the compiling issue, thanks.
Thanks. If I understand correctly, this is only a runtime issue, am I right ? Actually, the AMD machine is my build machine but I intend to run the linux-dna kernel on a core2duo. I should be fine without doing anything special, right ?
TIA |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Wed Feb 17, 2010 7:33 am Post subject: |
|
|
Yes I think that should be fine to compile on the AMD and then move it over. _________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
Posted: Wed Feb 17, 2010 8:00 am Post subject: |
|
|
OK thanks. Can't wait to try that at home
BTW, it builds fine with Kolivas' BFS... "I believe I can flyyyyyyyyyyyy" |
|
Back to top |
|
|
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
|
Back to top |
|
|
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
Posted: Thu Feb 18, 2010 7:20 am Post subject: |
|
|
build fails
EDIT: it fails the same way you reported on google groups
Code: | compilation aborted for fs/compat_ioctl.c (code 2)
make[1]: *** [fs/compat_ioctl.o] Error 2
make: *** [fs] Error 2 |
I'll try linuxrouter's patch.
kernel panic
more details coming later today... |
|
Back to top |
|
|
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
Posted: Thu Feb 18, 2010 3:56 pm Post subject: |
|
|
ok I'll try to focus on .33 as it should be out soon. so I used:
http://www.linuxdna.com/dna-2.6.33-rc8-intel64-1.patch on 2.6.33-rc8 vanilla with
Code: | --- linux-2.6/fs/compat_ioctl.c 2010-02-11 23:52:47.341194876 -0700
+++ linux-2.6.dna/fs/compat_ioctl.c 2010-02-18 01:08:14.537439544 -0700
@@ -1660,7 +1660,7 @@
int i;
const int max = ARRAY_SIZE(ioctl_pointer) - 1;
- BUILD_BUG_ON(max >= (1 << 16));
+ MAYBE_BUILD_BUG_ON(max >= (1 << 16));
/* guess initial offset into table, assuming a
normalized distribution */ |
build went fine without any scary message but I still get a kernel panic:
http://omploader.org/vM2xkeQ/DSC_0006.JPG
Quote: | You may need to add the ia32 emulation to it again. |
if you're talking about the patch you posted in https://forums.gentoo.org/viewtopic-p-6175189.html#6175189 , it's already included in dna-2.6.33-rc8-intel64-1.patch |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Thu Feb 18, 2010 7:15 pm Post subject: |
|
|
What does the top of that kernel panic say? something like: kernel panic not syncing.... _________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
Posted: Thu Feb 18, 2010 9:21 pm Post subject: |
|
|
Thaidog wrote: | What does the top of that kernel panic say? something like: kernel panic not syncing.... |
Code: | Kernel panic - not syncing: Attempted to kill init!
Pid: 1, comm: Swapper Tainted: G D 2.6.33-rc8-dna #1 |
Two additional info that may or may not be important:
- This is a i686 install but I'm running all kind of x86_64 kernel on it (64on32) without any issue since years
- ICC is not installed on this machine |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Fri Feb 19, 2010 4:47 am Post subject: |
|
|
ebrsc wrote: | Thaidog wrote: | What does the top of that kernel panic say? something like: kernel panic not syncing.... |
Code: | Kernel panic - not syncing: Attempted to kill init!
Pid: 1, comm: Swapper Tainted: G D 2.6.33-rc8-dna #1 |
Two additional info that may or may not be important:
- This is a i686 install but I'm running all kind of x86_64 kernel on it (64on32) without any issue since years
- ICC is not installed on this machine |
Linuxrouter over at the google group suggested this patch to try:
http://www.linuxforge.net/linux/kernel/dna-2.6.33-rc8-intel64-1.patch _________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Sat Feb 20, 2010 10:49 pm Post subject: |
|
|
Thaidog wrote: | tranquilcool wrote: | make AR=xiar LD=xild && make AR=xiar LD=xild modules_install
with patch make hangs at;
xild: executing 'ld'
CC arch/x86/kernel/cpu/intel_cacheinfo.o
CC arch/x86/kernel/cpu/addon_cpuid_features.o
CC arch/x86/kernel/cpu/proc.o
MKCAP arch/x86/kernel/cpu/capflags.c
CC arch/x86/kernel/cpu/capflags.o
CC arch/x86/kernel/cpu/powerflags.o
CC arch/x86/kernel/cpu/common.o |
No need to add AR=xiar LD=xild to the command any more. |
I'm sorry this advice is incorrect. I thought Luyi had added this in to the latest patch but it is not. _________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Sun Feb 21, 2010 1:07 am Post subject: |
|
|
FYI we got ebrsc's kernel to compile by turning off Oprofile support:
General Setup --> OProfile system profiling (EXPERIMENTAL)
Here is the latest patch:
wget http://www.linuxdna.com/dna-2.6.33-rc8-intel64-2.patch _________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
Posted: Sun Feb 21, 2010 2:20 am Post subject: |
|
|
well it compiled fine but was panicking. now that I can boot, I'm getting issues... just some random examples:
- I can't access any directory mounted with sshfs
- playing anything from youtube will result in a firefox crash at best or a complete system lockup.
Now is the time to check logs/debugging to see if anything interesting shows up... |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Sun Feb 21, 2010 5:16 am Post subject: |
|
|
ebrsc wrote: |
well it compiled fine but was panicking. now that I can boot, I'm getting issues... just some random examples:
- I can't access any directory mounted with sshfs
- playing anything from youtube will result in a firefox crash at best or a complete system lockup.
Now is the time to check logs/debugging to see if anything interesting shows up... |
Let us know what you find out! Also if you get a chance to compile the kernel with icc installed let us know if these issues persist with icc lib linkage. _________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
ebrsc n00b
Joined: 24 Jun 2008 Posts: 31
|
Posted: Mon Feb 22, 2010 9:12 am Post subject: |
|
|
Thaidog wrote: | Also if you get a chance to compile the kernel with icc installed let us know if these issues persist with icc lib linkage. |
I put my 64on32 setup aside and installed a fresh and pure 64bit system with ICC installed and recompiled my kernel with AR=xiar LD=xild. The firefox/flash issue is still there. I didnt try sshfs as the former problem is a blocker anyway. Out of curiosity Thaidog, are you using linux-dna on a desktop or as a server ? Is everything working as it should ? |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Mon Feb 22, 2010 4:16 pm Post subject: |
|
|
ebrsc wrote: | Thaidog wrote: | Also if you get a chance to compile the kernel with icc installed let us know if these issues persist with icc lib linkage. |
I put my 64on32 setup aside and installed a fresh and pure 64bit system with ICC installed and recompiled my kernel with AR=xiar LD=xild. The firefox/flash issue is still there. I didnt try sshfs as the former problem is a blocker anyway. Out of curiosity Thaidog, are you using linux-dna on a desktop or as a server ? Is everything working as it should ? |
I have been using the 32bit 2.6.30/31 patch on my P4 workstation with no issues with flash. I only have the 64bit 2.6.32/33 version on a vm right now. I will add a thread to the google group to see if anybody else reports these issues. What video card are you using? You might want to try the new 64bit flash beta to see if that helps any:
http://labs.adobe.com/technologies/flashplayer10/ _________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
Posted: Wed Feb 24, 2010 4:36 am Post subject: |
|
|
The flash bug seems to be fixed with this patch:
http://www.linuxdna.com/dna-2.6.32.8-intel64-3.patch _________________ Registered Linux User: 437619
"I'm a big believer in technology over politics" - Linus Torvalds |
|
Back to top |
|
|
Thaidog Veteran
Joined: 19 May 2004 Posts: 1053
|
|
Back to top |
|
|
darkbasic Tux's lil' helper
Joined: 06 Sep 2006 Posts: 133
|
|
Back to top |
|
|
|