Skip to contentSkip to conversation

size of spinlock_t

Linux-ia64 Ray Bryant 2 participants 3 messages Feb 15, 2002

lore
Share & tools
lore

Ray Bryant

author
#0 lore raw

In experimenting with new numa lock implementation, I changed the size of spinlock_t to 16 bytes. The resulting kernel hangs on our SN hardware just after printing "All processors have done init_idle". Now it is possible this is something peculiar to the SGI kernels, but I was wondering if anyone knew of places in the IA64 kernel that "know" that sizeof(spinlock_t)=8 and don't work correctly if this is changed.

Signature · 9 lines
Best Regards,
Ray
-----------------------------------------------
Ray Bryant               SGI
512-453-9679 (work)      512-507-7807 (cell)
[email protected]           [email protected]
     Computers are like air conditioners.
   They stop working when you open Windows.
-----------------------------------------------
#1 lore raw
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line · Ray Bryant

Ray> In experimenting with new numa lock implementation, I changed Ray> the size of spinlock_t to 16 bytes. The resulting kernel hangs Ray> on our SN hardware just after printing "All processors have Ray> done init_idle". Now it is possible this is something peculiar Ray> to the SGI kernels, but I was wondering if anyone knew of Ray> places in the IA64 kernel that "know" that
Ray> sizeof(spinlock_t)=8 and don't work correctly if this is Ray> changed.

Are you sure it's not just include/asm/offsets.h that's out of date? Make sure you have no stray arch/ia64/tools/print_offsets.s file and then rerun make dep.

	--david
author
#2 lore raw
In reply to Ray Bryant

Yes that was it. Thanks. I am totally amazed that "make clean; make vmlinux" did not remake offsets.h.

Quoted 22 lines · David Mosberger
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines · Ray Bryant

Ray> In experimenting with new numa lock implementation, I changed Ray> the size of spinlock_t to 16 bytes. The resulting kernel hangs Ray> on our SN hardware just after printing "All processors have Ray> done init_idle". Now it is possible this is something peculiar Ray> to the SGI kernels, but I was wondering if anyone knew of Ray> places in the IA64 kernel that "know" that
Ray> sizeof(spinlock_t)=8 and don't work correctly if this is Ray> changed.

Are you sure it's not just include/asm/offsets.h that's out of date? Make sure you have no stray arch/ia64/tools/print_offsets.s file and then rerun make dep.

        --david

_______________________________________________
Linux-IA64 mailing list
[email protected]
http://lists.linuxia64.org/lists/listinfo/linux-ia64

Signature · 9 lines
Best Regards,
Ray
-----------------------------------------------
Ray Bryant               SGI
512-453-9679 (work)      512-507-7807 (cell)
[email protected]           [email protected]
     Computers are like air conditioners.
   They stop working when you open Windows.
-----------------------------------------------