2496afbf1e
We need to have a stronger barrier between releasing the lock and checking for any waiting spinners. A compiler barrier is not sufficient because the CPU's ordering rules do not prevent the read xl->spinners from happening before the unlock assignment, as they are different memory locations. We need to have an explicit barrier to enforce the write-read ordering to different memory locations. Because of it, I can't bring up > 4 HVM guests on one SMP machine. [ Code and commit comments expanded -J ] [ Impact: avoid deadlock when using Xen PV spinlocks ] Signed-off-by: Yang Xiaowei <xiaowei.yang@intel.com> Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> |
||
|---|---|---|
| .. | ||
| Kconfig | ||
| Makefile | ||
| debugfs.c | ||
| debugfs.h | ||
| enlighten.c | ||
| grant-table.c | ||
| irq.c | ||
| mmu.c | ||
| mmu.h | ||
| multicalls.c | ||
| multicalls.h | ||
| setup.c | ||
| smp.c | ||
| spinlock.c | ||
| suspend.c | ||
| time.c | ||
| vdso.h | ||
| xen-asm.S | ||
| xen-asm.h | ||
| xen-asm_32.S | ||
| xen-asm_64.S | ||
| xen-head.S | ||
| xen-ops.h | ||