diff options
| author | Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> | 2009-12-09 10:56:41 +0900 |
|---|---|---|
| committer | Tony Luck <tony.luck@intel.com> | 2009-12-14 16:03:36 -0800 |
| commit | 09b366b78c9602f53344c269eac608fd6e24d670 (patch) | |
| tree | 1352edcc54a17157ba4e57e8dcd765d30df7889a /kernel/mutex-debug.h | |
| parent | 3ea6b3d0e6d0ffd91c0f8cadeb69b7133c038b32 (diff) | |
| download | kernel-crypto-09b366b78c9602f53344c269eac608fd6e24d670.tar.gz kernel-crypto-09b366b78c9602f53344c269eac608fd6e24d670.tar.xz kernel-crypto-09b366b78c9602f53344c269eac608fd6e24d670.zip | |
[IA64] preallocate IA64_IRQ_MOVE_VECTOR
Previously, we tried to use IA64_DEF_FIRST_DEVICE_VECTOR (0x30) as the
IA64_IRQ_MOVE_VECTOR. However, we allocate other IRQs from the device
vector range, so there's no guarantee that IA64_DEF_FIRST_DEVICE_VECTOR
will still be available when we register IA64_IRQ_MOVE_VECTOR.
This patch statically allocates 0x30 for IA64_IRQ_MOVE_VECTOR and
removes it from the device vector range.
Without this patch, we crash on machines like the HP rx3600 that use
vector 48 (0x30) as the ACPI SCI interrupt:
kernel BUG at arch/ia64/kernel/irq_ia64.c:647!
swapper[0]: bugcheck! 0 [1]
Modules linked in:
Pid: 0, CPU 0, comm: swapper
psr : 00001010084a2018 ifs : 800000000000030e ip : [<a000000100012ed0>] Not tainted (2.6.32-rc8-00184-gd5d4ec8)
ip is at ia64_native_register_percpu_irq+0x110/0x1e0
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Tested-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Tested-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Diffstat (limited to 'kernel/mutex-debug.h')
0 files changed, 0 insertions, 0 deletions