Skip to contentSkip to conversation

PATCH vmx.c: add printk_ratelimit in vmx_intr_assist

PATCH Ryan Harper 2 participants 6 messages Dec 13, 2007 – Dec 18, 2007 4 messages not in the archive KVM

lore
Share & tools
lore
#0 lore raw

Add printk_ratelimit check in front of printk. This prevents spamming of the message during 32-bit ubuntu 6.06server install. Previously, it would hang during the partition formatting stage.

Signature · 5 lines
Ryan Harper
Software Engineer; Linux Technology Center
IBM Corp., Austin, Tx
(512) 838-9253   T/L: 678-9253
ryanh-r/[email protected]

For some guests (ubuntu6.06-server) on intel machines, vmx_inter_assist spams a message preventing the guest from making forward progress. Adding in a ratelimit check prevents the spamming.

drivers/kvm/vmx.c +2 -1
index 548e3a5..1c1ad7f 100644
@@ -2261,7 +2261,8 @@ static void vmx_intr_assist(struct kvm_vcpu *vcpu)
 	if (intr_info_field & INTR_INFO_VALID_MASK) {
 		if (idtv_info_field & INTR_INFO_VALID_MASK) {
 			/* TODO: fault when IDT_Vectoring */
-			printk(KERN_ERR "Fault when IDT_Vectoring\n");
+			if (printk_ratelimit())
+				printk(KERN_ERR "Fault when IDT_Vectoring\n");
 		}
 		if (has_ext_irq)
 			enable_irq_window(vcpu);

-------------------------------------------------------------------------
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services
for just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace

_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

#2 lore raw
In reply to a message not in the archive
Quoted 4 lines · Ryan Harper

Add printk_ratelimit check in front of printk. This prevents spamming of the message during 32-bit ubuntu 6.06server install. Previously, it would hang during the partition formatting stage.

Applied, but this is worrying. Do you know whether this is a regression?

Signature · 9 lines
Any sufficiently difficult bug is indistinguishable from a feature.


-------------------------------------------------------------------------
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services
for just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
#4 lore raw
In reply to a message not in the archive
Quoted 7 lines · Avi Kivity
Quoted 4 lines · Ryan Harper

Add printk_ratelimit check in front of printk. This prevents spamming of the message during 32-bit ubuntu 6.06server install. Previously, it would hang during the partition formatting stage.

Applied, but this is worrying. Do you know whether this is a regression?

I'll try out older releases to see if we did regress.

Signature · 12 lines
Ryan Harper
Software Engineer; Linux Technology Center
IBM Corp., Austin, Tx
(512) 838-9253   T/L: 678-9253
ryanh-r/[email protected]

-------------------------------------------------------------------------
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services
for just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
#6 lore raw
In reply to a message not in the archive
Quoted 10 lines · Ryan Harper
Quoted 7 lines · Avi Kivity
Quoted 4 lines · Ryan Harper

Add printk_ratelimit check in front of printk. This prevents spamming of the message during 32-bit ubuntu 6.06server install. Previously, it would hang during the partition formatting stage.

Applied, but this is worrying. Do you know whether this is a regression?

I'll try out older releases to see if we did regress.

Looks like kvm-47 introduced the hang during format, -46 finishes the format just fine and with no printk spamming. I'll start seeing if I can bisect down to a particular changeset.

Signature · 12 lines
Ryan Harper
Software Engineer; Linux Technology Center
IBM Corp., Austin, Tx
(512) 838-9253   T/L: 678-9253
ryanh-r/[email protected]

-------------------------------------------------------------------------
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services
for just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
#8 lore raw
In reply to a message not in the archive
Quoted 8 lines · Ryan Harper
Quoted 7 lines · Ryan Harper
Quoted 6 lines · Avi Kivity
Quoted 4 lines · Ryan Harper

Add printk_ratelimit check in front of printk. This prevents spamming of the message during 32-bit ubuntu 6.06server install. Previously, it would hang during the partition formatting stage.

Quoted 7 lines
Quoted 3 lines
Quoted 1 line

Applied, but this is worrying. Do you know whether this is a regression?

I'll try out older releases to see if we did regress.

Looks like kvm-47 introduced the hang during format, -46 finishes the format just fine and with no printk spamming. I'll start seeing if I can bisect down to a particular changeset.

My git skills are poor and I can't for the life of my figure out how to get kvm and kvm-userspace checked out at kvm-46 and build them together. So, instead I tested the nightly tarballs between Oct 10 (kvm-46 release) and Oct 18 (kvm-47).

kvm-snapshot-20071014.tar.gz is the last known good version.

version:        kvm-snapshot-20071014
srcversion:     5F6E5524AD0A76F65C4CF60

The flooding of: [2592349.069880] Fault when IDT_Vectoring starts with 1015 but 1015 can complete the formatting. 16 and 17 won't compile. 18 of course fails completely.

So, something between 14 and 15 added the flooding, and something between 15 and 18 prevents formatting from progressing.

Signature · 12 lines
Ryan Harper
Software Engineer; Linux Technology Center
IBM Corp., Austin, Tx
(512) 838-9253   T/L: 678-9253
ryanh-r/[email protected]

-------------------------------------------------------------------------
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services
for just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
#9 lore raw
In reply to a message not in the archive
Quoted 11 lines · Ryan Harper
Quoted 9 lines · Ryan Harper
Quoted 7 lines · Avi Kivity
Quoted 5 lines · Ryan Harper

Add printk_ratelimit check in front of printk. This prevents spamming of the message during 32-bit ubuntu 6.06server install. Previously, it would hang during the partition formatting stage.

Quoted 10 lines
Quoted 4 lines
Quoted 2 lines

Applied, but this is worrying. Do you know whether this is a regression?

I'll try out older releases to see if we did regress.

Looks like kvm-47 introduced the hang during format, -46 finishes the format just fine and with no printk spamming. I'll start seeing if I can bisect down to a particular changeset.

Unfortunately there was a window where interrupts were injected outside the critical section, causing these errors. This was introduced in kvm-47, so this doesn't help much.

Signature · 9 lines
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services
for just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace