[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[xen staging] x86/irq: remove offline CPUs from old CPU mask when adjusting move_cleanup_count



commit e63209d3ba2fd1b2f232babd14c9c679ffa7b09a
Author:     Roger Pau Monné <roger.pau@xxxxxxxxxx>
AuthorDate: Mon Jun 10 10:33:22 2024 +0200
Commit:     Jan Beulich <jbeulich@xxxxxxxx>
CommitDate: Mon Jun 10 10:33:22 2024 +0200

    x86/irq: remove offline CPUs from old CPU mask when adjusting 
move_cleanup_count
    
    When adjusting move_cleanup_count to account for CPUs that are offline also
    adjust old_cpu_mask, otherwise further calls to fixup_irqs() could subtract
    those again and create an imbalance in move_cleanup_count.
    
    Fixes: 472e0b74c5c4 ('x86/IRQ: deal with move cleanup count state in 
fixup_irqs()')
    Signed-off-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>
    Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>
---
 xen/arch/x86/irq.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/xen/arch/x86/irq.c b/xen/arch/x86/irq.c
index e36e06deaa..d101ffeaf9 100644
--- a/xen/arch/x86/irq.c
+++ b/xen/arch/x86/irq.c
@@ -2563,6 +2563,14 @@ void fixup_irqs(const cpumask_t *mask, bool verbose)
             desc->arch.move_cleanup_count -= cpumask_weight(affinity);
             if ( !desc->arch.move_cleanup_count )
                 release_old_vec(desc);
+            else
+                /*
+                 * Adjust old_cpu_mask to account for the offline CPUs,
+                 * otherwise further calls to fixup_irqs() could subtract those
+                 * again and possibly underflow the counter.
+                 */
+                cpumask_andnot(desc->arch.old_cpu_mask, 
desc->arch.old_cpu_mask,
+                               affinity);
         }
 
         if ( !desc->action || cpumask_subset(desc->affinity, mask) )
--
generated by git-patchbot for /home/xen/git/xen.git#staging



 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.