From: Julia Lawall <[email protected]>
There is a call to local_irq_restore in the normal exit case, so it would
seem that there should be one on an error return as well.
The semantic patch that finds this problem is as follows:
(http://www.emn.fr/x-info/coccinelle/)
// <smpl>
@@
expression l;
expression E,E1,E2;
@@
local_irq_save(l);
... when != local_irq_restore(l)
when != spin_unlock_irqrestore(E,l)
when any
when strict
(
if (...) { ... when != local_irq_restore(l)
when != spin_unlock_irqrestore(E1,l)
+ local_irq_restore(l);
return ...;
}
|
if (...)
+ {local_irq_restore(l);
return ...;
+ }
|
spin_unlock_irqrestore(E2,l);
|
local_irq_restore(l);
)
// </smpl>
Signed-off-by: Julia Lawall <[email protected]>
---
.../julia/linuxcopy}/arch/ia64/kvm/kvm-ia64.c | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/home/julia/linux-2.6/arch/ia64/kvm/kvm-ia64.c b/var/julia/linuxcopy/arch/ia64/kvm/kvm-ia64.c
index 2672f4d..7a37d06 100644
--- a/home/julia/linux-2.6/arch/ia64/kvm/kvm-ia64.c
+++ b/var/julia/linuxcopy/arch/ia64/kvm/kvm-ia64.c
@@ -125,9 +125,9 @@ void kvm_arch_hardware_enable(void *garbage)
PAGE_KERNEL));
local_irq_save(saved_psr);
slot = ia64_itr_entry(0x3, KVM_VMM_BASE, pte, KVM_VMM_SHIFT);
+ local_irq_restore(saved_psr);
if (slot < 0)
return;
- local_irq_restore(saved_psr);
spin_lock(&vp_lock);
status = ia64_pal_vp_init_env(kvm_vsa_base ?
@@ -160,9 +160,9 @@ void kvm_arch_hardware_disable(void *garbage)
local_irq_save(saved_psr);
slot = ia64_itr_entry(0x3, KVM_VMM_BASE, pte, KVM_VMM_SHIFT);
+ local_irq_restore(saved_psr);
if (slot < 0)
return;
- local_irq_restore(saved_psr);
status = ia64_pal_vp_exit_env(host_iva);
if (status)
@@ -1253,6 +1253,7 @@ static int vti_vcpu_setup(struct kvm_vcpu *vcpu, int id)
uninit:
kvm_vcpu_uninit(vcpu);
fail:
+ local_irq_restore(psr);
return r;
}
On Tue, Jul 22, 2008 at 09:38:18PM +0200, Julia Lawall wrote:
> There is a call to local_irq_restore in the normal exit case, so it would
> seem that there should be one on an error return as well.
Patch looks good to me:
Reviewed-by: Matthew Wilcox <[email protected]>
--
Intel are signing my paycheques ... these opinions are still mine
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours. We can't possibly take such
a retrograde step."
Julia Lawall wrote:
> From: Julia Lawall <[email protected]>
>
> There is a call to local_irq_restore in the normal exit case, so it
> would
> seem that there should be one on an error return as well.
>
> The semantic patch that finds this problem is as follows:
> (http://www.emn.fr/x-info/coccinelle/)
Good catch! Excellent patch to find such issues!
Xiantao
>
> // <smpl>
> @@
> expression l;
> expression E,E1,E2;
> @@
>
> local_irq_save(l);
> ... when != local_irq_restore(l)
> when != spin_unlock_irqrestore(E,l)
> when any
> when strict
> (
> if (...) { ... when != local_irq_restore(l)
> when != spin_unlock_irqrestore(E1,l)
> + local_irq_restore(l);
> return ...;
> }
>>
> if (...)
> + {local_irq_restore(l);
> return ...;
> + }
>>
> spin_unlock_irqrestore(E2,l);
>>
> local_irq_restore(l);
> )
> // </smpl>
>
> Signed-off-by: Julia Lawall <[email protected]>
>
> ---
> .../julia/linuxcopy}/arch/ia64/kvm/kvm-ia64.c | 5 +++--
> 1 files changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/home/julia/linux-2.6/arch/ia64/kvm/kvm-ia64.c
> b/var/julia/linuxcopy/arch/ia64/kvm/kvm-ia64.c index 2672f4d..7a37d06
> 100644 --- a/home/julia/linux-2.6/arch/ia64/kvm/kvm-ia64.c
> +++ b/var/julia/linuxcopy/arch/ia64/kvm/kvm-ia64.c
> @@ -125,9 +125,9 @@ void kvm_arch_hardware_enable(void *garbage)
> PAGE_KERNEL));
> local_irq_save(saved_psr);
> slot = ia64_itr_entry(0x3, KVM_VMM_BASE, pte, KVM_VMM_SHIFT);
> + local_irq_restore(saved_psr);
> if (slot < 0)
> return;
> - local_irq_restore(saved_psr);
>
> spin_lock(&vp_lock);
> status = ia64_pal_vp_init_env(kvm_vsa_base ?
> @@ -160,9 +160,9 @@ void kvm_arch_hardware_disable(void *garbage)
>
> local_irq_save(saved_psr);
> slot = ia64_itr_entry(0x3, KVM_VMM_BASE, pte, KVM_VMM_SHIFT);
> + local_irq_restore(saved_psr);
> if (slot < 0)
> return;
> - local_irq_restore(saved_psr);
>
> status = ia64_pal_vp_exit_env(host_iva);
> if (status)
> @@ -1253,6 +1253,7 @@ static int vti_vcpu_setup(struct kvm_vcpu
> *vcpu, int id) uninit:
> kvm_vcpu_uninit(vcpu);
> fail:
> + local_irq_restore(psr);
> return r;
> }
Julia Lawall wrote:
> From: Julia Lawall <[email protected]>
>
> There is a call to local_irq_restore in the normal exit case, so it would
> seem that there should be one on an error return as well.
>
> The semantic patch that finds this problem is as follows:
> (http://www.emn.fr/x-info/coccinelle/)
>
Applied, thanks.
--
Do not meddle in the internals of kernels, for they are subtle and quick to panic.