X86_64 DOM0 SMP sysfs entries for cpu
Anyone running X86_64 DOM0 with CONFIG_SMP=y want to confirm whether they have entries in sysfs for cpus?
run:
ls /sys/devices/system/cpu/
On my two-way Opteron running x86_32, I get:
(bebop) ~ % ls /sys/devices/system/cpu/
cpu0/ cpu1/
When I boot x86_64 SMP I get nothing in there.
Signature · 5 lines
Ryan Harper Software Engineer; Linux Technology Center IBM Corp., Austin, Tx (512) 838-9253 T/L: 678-9253 [email protected]
PATCH 1/1 sparse: link to i386 topology.o
Quoted 5 lines · Ryan Harper
Anyone running
X86_64DOM0 withCONFIG_SMP=ywant to confirm whether they have entries in sysfs for cpus?run:
Quoted 8 lines
ls
/sys/devices/system/cpu/On my two-way Opteron running
x86_32, I get:(bebop) ~ % ls
/sys/devices/system/cpu/
cpu0/ cpu1/When I boot
x86_64SMP I get nothing in there.
The arch_{register/unregister}_cpu() routines are in i386 topology code which x86_64 normally includes. The obj-y rule was commented out in x86_64/kernel/Makefile, maybe on purpose. With the attached patch, values are populated in sysfs correctly.
Signature · 24 lines
Ryan Harper Software Engineer; Linux Technology Center IBM Corp., Austin, Tx (512) 838-9253 T/L: 678-9253 [email protected] diffstat output: Makefile | 2 +- 1 files changed, 1 insertion(+), 1 deletion(-) Signed-off-by: Ryan Harper <[email protected]> --- diff -r 188c782fa9bb linux-2.6-xen-sparse/arch/xen/x86_64/kernel/Makefile --- a/linux-2.6-xen-sparse/arch/xen/x86_64/kernel/Makefile Fri Aug 19 13:05:31 2005 +++ b/linux-2.6-xen-sparse/arch/xen/x86_64/kernel/Makefile Tue Aug 23 11:10:59 2005 @@ -44,7 +44,7 @@ c-obj-$(CONFIG_MODULES) += module.o -#obj-y += topology.o +obj-y += topology.o c-obj-y += intel_cacheinfo.o bootflag-y += ../../../i386/kernel/bootflag.o
Re: [PATCH] sparse: link to i386 topology.o
Quoted 4 lines · * Ryan Harper ([email protected]) wrote:
The arch_{register/unregister}_cpu() routines are in i386 topology code which
x86_64normally includes. The obj-y rule was commented out in x86_64/kernel/Makefile, maybe on purpose. With the attached patch, values are populated in sysfs correctly.
ACK, thanks Ryan.
-chris