From: Jacob Pan <[email protected]> This is a backport for 2.6.37 stable. The original commit ID is 6550904ddbc3c286798a87edf95eeebcc62bc58a
Offlining the secondary CPU causes the timer irq affinity to be set to CPU 0. When the secondary CPU is back online again, the wrong irq affinity will be used. This patch ensures secondary per CPU timer always has the correct IRQ affinity when enabled. Signed-off-by: Jacob Pan <[email protected]> LKML-Reference: <[email protected]> Signed-off-by: H. Peter Anvin <[email protected]> Cc: <[email protected]> 2.6.37 --- arch/x86/kernel/apb_timer.c | 8 +++++--- 1 files changed, 5 insertions(+), 3 deletions(-) diff --git a/arch/x86/kernel/apb_timer.c b/arch/x86/kernel/apb_timer.c index 92543c7..5181673 100644 --- a/arch/x86/kernel/apb_timer.c +++ b/arch/x86/kernel/apb_timer.c @@ -313,10 +313,12 @@ static void apbt_setup_irq(struct apbt_dev *adev) if (adev->irq == 0) return; + irq_modify_status(adev->irq, 0, IRQ_MOVE_PCNTXT); + irq_set_affinity(adev->irq, cpumask_of(adev->cpu)); + /* APB timer irqs are set up as mp_irqs, timer is edge type */ + __set_irq_handler(adev->irq, handle_edge_irq, 0, "edge"); + if (system_state == SYSTEM_BOOTING) { - irq_modify_status(adev->irq, 0, IRQ_MOVE_PCNTXT); - /* APB timer irqs are set up as mp_irqs, timer is edge type */ - __set_irq_handler(adev->irq, handle_edge_irq, 0, "edge"); if (request_irq(adev->irq, apbt_interrupt_handler, IRQF_TIMER | IRQF_DISABLED | IRQF_NOBALANCING, adev->name, adev)) { -- 1.7.0.4 _______________________________________________ stable mailing list [email protected] http://linux.kernel.org/mailman/listinfo/stable
