net: ipa: don't proceed to out-of-bound write
Subsystems (3)
Jakub Kicinski
maintainerPATCH 1/1 net: ipa: don't proceed to out-of-bound write
GCC 12 seems upset that we check ipa_irq against array bound but then proceed, anyway:
drivers/net/ipa/ipa_interrupt.c: In function ‘ipa_interrupt_add’:drivers/net/ipa/ipa_interrupt.c:196:27: warning: array subscript 30 is above array bounds of ‘void (*[30])(struct ipa *, enum ipa_irq_id)’ [-Warray-bounds]
196 | interrupt->handler[ipa_irq] = handler;
| ~~~~~~~~~~~~~~~~~~^~~~~~~~~drivers/net/ipa/ipa_interrupt.c:42:27: note: while referencing ‘handler’
42 | ipa_irq_handler_t handler[IPA_IRQ_COUNT];
| ^~~~~~~
---
CC: [email protected]
--- drivers/net/ipa/ipa_interrupt.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/net/ipa/ipa_interrupt.c +4 -2
@@ -191,7 +191,8 @@ void ipa_interrupt_add(struct ipa_interrupt *interrupt,struct ipa *ipa = interrupt->ipa;u32 offset; -WARN_ON (ipa_irq >=IPA_IRQ_COUNT ); +if (WARN_ON (ipa_irq >=IPA_IRQ_COUNT )) +return ; interrupt->handler[ipa_irq] = handler;
@@ -208,7 +209,8 @@ ipa_interrupt_remove(struct ipa_interrupt *interrupt, enum ipa_irq_id ipa_irq)struct ipa *ipa = interrupt->ipa;u32 offset; -WARN_ON (ipa_irq >=IPA_IRQ_COUNT ); +if (WARN_ON (ipa_irq >=IPA_IRQ_COUNT )) +return ;/* Update the IPA interrupt mask to disable it */ interrupt->enabled &= ~BIT (ipa_irq);
Signature · 1 line
2.34.3
Quoted 5 lines · Jakub Kicinski
GCC 12 seems upset that we check
ipa_irqagainst array bound but then proceed, anyway:
drivers/net/ipa/ipa_interrupt.c: In function ‘ipa_interrupt_add’:drivers/net/ipa/ipa_interrupt.c:196:27: warning: array subscript 30 is above array bounds of ‘void (*[30])(struct ipa *, enum ipa_irq_id)’ [-Warray-bounds]
Quoted 7 lines
196 | interrupt->handler[ipa_irq] = handler; | ~~~~~~~~~~~~~~~~~~^~~~~~~~~ drivers/net/ipa/ipa_interrupt.c:42:27: note: while referencing ‘handler’ 42 | ipa_irq_handler_t handler[IPA_IRQ_COUNT]; | ^~~~~~~Signed-off-by Jakub Kicinski <[email protected]>
Well, that's a reasonable thing to complain about. I think when I switched to using these WARN*() calls Leon said testing the return value was unusual in the networking code.
In any case, this is a good fix. The problem won't happen anyway, so silencing the error this way is just fine.
Quoted 30 lines
---
CC: [email protected]--- drivers/net/ipa/ipa_interrupt.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)drivers/net/ipa/ipa_interrupt.c +4
-2@@ -191,7 +191,8 @@ void ipa_interrupt_add(struct ipa_interrupt *interrupt,struct ipa *ipa = interrupt->ipa;u32 offset; -WARN_ON (ipa_irq >=IPA_IRQ_COUNT ); +if (WARN_ON (ipa_irq >=IPA_IRQ_COUNT )) +return ; interrupt->handler[ipa_irq] = handler;@@ -208,7 +209,8 @@ ipa_interrupt_remove(struct ipa_interrupt *interrupt, enum ipa_irq_id ipa_irq)struct ipa *ipa = interrupt->ipa;u32 offset; -WARN_ON (ipa_irq >=IPA_IRQ_COUNT ); +if (WARN_ON (ipa_irq >=IPA_IRQ_COUNT )) +return ;/* Update the IPA interrupt mask to disable it */ interrupt->enabled &= ~BIT (ipa_irq);
This patch was applied to netdev/net-next.git (master) by Jakub Kicinski <[email protected]>:
Hello:
This patch was applied to netdev/net-next.git (master) by Jakub Kicinski <[email protected]>:
Quoted 5 lines · On Wed, 18 May 2022 17:44:17 -0700 you wrote:
GCC 12 seems upset that we check
ipa_irqagainst array bound but then proceed, anyway:
drivers/net/ipa/ipa_interrupt.c: In function ‘ipa_interrupt_add’:drivers/net/ipa/ipa_interrupt.c:196:27: warning: array subscript 30 is above array bounds of ‘void (*[30])(struct ipa *, enum ipa_irq_id)’ [-Warray-bounds]
Quoted 7 lines
196 | interrupt->handler[ipa_irq] = handler; | ~~~~~~~~~~~~~~~~~~^~~~~~~~~
drivers/net/ipa/ipa_interrupt.c:42:27: note: while referencing ‘handler’
42 |ipa_irq_handler_thandler[IPA_IRQ_COUNT];
| ^~~~~~~[...]
Here is the summary with links:
- [net-next] net: ipa: don't proceed to out-of-bound write
https://git.kernel.org/netdev/net-next/c/1172aa6e4a19
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html