* [PATCH] Fix loop logic in irq_alloc_virt()
@ 2006-08-02 0:48 Michael Ellerman
2006-08-02 4:20 ` Michael Ellerman
2006-08-09 9:38 ` Benjamin Herrenschmidt
0 siblings, 2 replies; 3+ messages in thread
From: Michael Ellerman @ 2006-08-02 0:48 UTC (permalink / raw)
To: Paul Mackerras; +Cc: linuxppc-dev
There's a bug in irq_alloc_virt() if it's asked for more than 1 interrupt,
if it can't find a slot it might look past the end of the irq_map.
I think this is a fix. No one in the kernel actually calls this with
count > 1, so it's not critical.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
---
arch/powerpc/kernel/irq.c | 19 ++++++++++---------
1 file changed, 10 insertions(+), 9 deletions(-)
Index: to-merge/arch/powerpc/kernel/irq.c
===================================================================
--- to-merge.orig/arch/powerpc/kernel/irq.c
+++ to-merge/arch/powerpc/kernel/irq.c
@@ -728,7 +728,6 @@ unsigned int irq_alloc_virt(struct irq_h
{
unsigned long flags;
unsigned int i, j, found = NO_IRQ;
- unsigned int limit = irq_virq_count - count;
if (count == 0 || count > (irq_virq_count - NUM_ISA_INTERRUPTS))
return NO_IRQ;
@@ -745,14 +744,16 @@ unsigned int irq_alloc_virt(struct irq_h
/* Look for count consecutive numbers in the allocatable
* (non-legacy) space
*/
- for (i = NUM_ISA_INTERRUPTS; i <= limit; ) {
- for (j = i; j < (i + count); j++)
- if (irq_map[j].host != NULL) {
- i = j + 1;
- continue;
- }
- found = i;
- break;
+ for (i = NUM_ISA_INTERRUPTS, j = 0; i < irq_virq_count; i++) {
+ if (irq_map[i].host != NULL)
+ j = 0;
+ else
+ j++;
+
+ if (j == count) {
+ found = i - count + 1;
+ break;
+ }
}
if (found == NO_IRQ) {
spin_unlock_irqrestore(&irq_big_lock, flags);
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] Fix loop logic in irq_alloc_virt()
2006-08-02 0:48 [PATCH] Fix loop logic in irq_alloc_virt() Michael Ellerman
@ 2006-08-02 4:20 ` Michael Ellerman
2006-08-09 9:38 ` Benjamin Herrenschmidt
1 sibling, 0 replies; 3+ messages in thread
From: Michael Ellerman @ 2006-08-02 4:20 UTC (permalink / raw)
To: Paul Mackerras; +Cc: linuxppc-dev
[-- Attachment #1: Type: text/plain, Size: 1319 bytes --]
On Wed, 2006-08-02 at 10:48 +1000, Michael Ellerman wrote:
> There's a bug in irq_alloc_virt() if it's asked for more than 1 interrupt,
> if it can't find a slot it might look past the end of the irq_map.
>
> I think this is a fix. No one in the kernel actually calls this with
> count > 1, so it's not critical.
> Index: to-merge/arch/powerpc/kernel/irq.c
> ===================================================================
> --- to-merge.orig/arch/powerpc/kernel/irq.c
> +++ to-merge/arch/powerpc/kernel/irq.c
> @@ -745,14 +744,16 @@ unsigned int irq_alloc_virt(struct irq_h
> /* Look for count consecutive numbers in the allocatable
> * (non-legacy) space
> */
> - for (i = NUM_ISA_INTERRUPTS; i <= limit; ) {
> - for (j = i; j < (i + count); j++)
> - if (irq_map[j].host != NULL) {
> - i = j + 1;
> - continue;
> - }
> - found = i;
> - break;
To be clear: the bug is that the continue affects the inner for loop,
not the outer one, so i becomes j + 1 and then we continue the inner
loop without checking if i is still <= limit.
cheers
--
Michael Ellerman
IBM OzLabs
wwweb: http://michael.ellerman.id.au
phone: +61 2 6212 1183 (tie line 70 21183)
We do not inherit the earth from our ancestors,
we borrow it from our children. - S.M.A.R.T Person
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 191 bytes --]
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] Fix loop logic in irq_alloc_virt()
2006-08-02 0:48 [PATCH] Fix loop logic in irq_alloc_virt() Michael Ellerman
2006-08-02 4:20 ` Michael Ellerman
@ 2006-08-09 9:38 ` Benjamin Herrenschmidt
1 sibling, 0 replies; 3+ messages in thread
From: Benjamin Herrenschmidt @ 2006-08-09 9:38 UTC (permalink / raw)
To: Michael Ellerman; +Cc: linuxppc-dev list
On Wed, 2006-08-02 at 10:48 +1000, Michael Ellerman wrote:
> There's a bug in irq_alloc_virt() if it's asked for more than 1 interrupt,
> if it can't find a slot it might look past the end of the irq_map.
>
> I think this is a fix. No one in the kernel actually calls this with
> count > 1, so it's not critical.
Good catch. /me stupid.
> Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> ---
>
> arch/powerpc/kernel/irq.c | 19 ++++++++++---------
> 1 file changed, 10 insertions(+), 9 deletions(-)
>
> Index: to-merge/arch/powerpc/kernel/irq.c
> ===================================================================
> --- to-merge.orig/arch/powerpc/kernel/irq.c
> +++ to-merge/arch/powerpc/kernel/irq.c
> @@ -728,7 +728,6 @@ unsigned int irq_alloc_virt(struct irq_h
> {
> unsigned long flags;
> unsigned int i, j, found = NO_IRQ;
> - unsigned int limit = irq_virq_count - count;
>
> if (count == 0 || count > (irq_virq_count - NUM_ISA_INTERRUPTS))
> return NO_IRQ;
> @@ -745,14 +744,16 @@ unsigned int irq_alloc_virt(struct irq_h
> /* Look for count consecutive numbers in the allocatable
> * (non-legacy) space
> */
> - for (i = NUM_ISA_INTERRUPTS; i <= limit; ) {
> - for (j = i; j < (i + count); j++)
> - if (irq_map[j].host != NULL) {
> - i = j + 1;
> - continue;
> - }
> - found = i;
> - break;
> + for (i = NUM_ISA_INTERRUPTS, j = 0; i < irq_virq_count; i++) {
> + if (irq_map[i].host != NULL)
> + j = 0;
> + else
> + j++;
> +
> + if (j == count) {
> + found = i - count + 1;
> + break;
> + }
> }
> if (found == NO_IRQ) {
> spin_unlock_irqrestore(&irq_big_lock, flags);
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2006-08-09 12:32 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-08-02 0:48 [PATCH] Fix loop logic in irq_alloc_virt() Michael Ellerman
2006-08-02 4:20 ` Michael Ellerman
2006-08-09 9:38 ` Benjamin Herrenschmidt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).