Skip to contentSkip to conversation

fix irq_vector[] update of c/s 19419

PATCH In review Jan Beulich 2 participants 2 messages Apr 1, 2009

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 fix irq_vector[] update of c/s 19419

The updating of irq_vector[] looks very suspicious - the array is IRQ- indexed, hence you potentially overwrite an already existing entry, or risk your entry to be overwritten later.

Also, what is the reason for overwriting the vector_irq[] initialization assign_irq_vector() already did?

Likewise you should be calling free_irq_vector() in the error path rather than doing any of the updates yourself.

Finally, assign_irq_vector() does not return zero on failure, but a negative error code.

If there is nothing I'm overlooking, here's a patch - untested, as I don't have the needed hardware.

xen/arch/x86/hpet.c +3 -6
@@ -346,17 +346,14 @@ static int hpet_assign_irq(struct hpet_e
     unsigned int vector;
 
     vector = assign_irq_vector(AUTO_ASSIGN_IRQ);
-    if ( !vector )
-        return -EINVAL;
+    if ( (int)vector < 0 )
+        return vector;
 
-    irq_vector[vector] = vector;
-    vector_irq[vector] = vector;
     vector_channel[vector] = ch - &hpet_events[0];
 
     if ( hpet_setup_msi_irq(vector) )
     {
-        irq_vector[vector] = 0;
-        vector_irq[vector] = FREE_TO_ASSIGN_IRQ;
+        free_irq_vector(vector);
         vector_channel[vector] = -1;
         return -EINVAL;
     }
#1 lore raw

All things you found are absolutely right.

I used to think some code paths may need a vector to irq converting, but I can't find a way to dynamically assign a irq, so I did the assignment manually. Careful checking shows it is not necessary, so your correction should be ok.

As to the assign_irq_vector() return code, it should be my incaution.

Thanks for this correction patch.

Jimmy

Quoted 41 lines · Jan Beulich

The updating of irq_vector[] looks very suspicious - the array is IRQ- indexed, hence you potentially overwrite an already existing entry, or risk your entry to be overwritten later.

Also, what is the reason for overwriting the vector_irq[] initialization assign_irq_vector() already did?

Likewise you should be calling free_irq_vector() in the error path rather than doing any of the updates yourself.

Finally, assign_irq_vector() does not return zero on failure, but a negative error code.

If there is nothing I'm overlooking, here's a patch - untested, as I don't have the needed hardware.

xen/arch/x86/hpet.c +3 -6
@@ -346,17 +346,14 @@ static int hpet_assign_irq(struct hpet_e
     unsigned int vector;
 
     vector = assign_irq_vector(AUTO_ASSIGN_IRQ);
-    if ( !vector )
-        return -EINVAL;
+    if ( (int)vector < 0 )
+        return vector;
 
-    irq_vector[vector] = vector;
-    vector_irq[vector] = vector;
     vector_channel[vector] = ch - &hpet_events[0];
 
     if ( hpet_setup_msi_irq(vector) )
     {
-        irq_vector[vector] = 0;
-        vector_irq[vector] = FREE_TO_ASSIGN_IRQ;
+        free_irq_vector(vector);
         vector_channel[vector] = -1;
         return -EINVAL;
     }