On Thu, Apr 12, 2007 at 03:57:48PM -0700, Jeremy Fitzhardinge wrote:
> diff -r 2ad8a0729f26 arch/arm/boot/compressed/misc.c
> --- a/arch/arm/boot/compressed/misc.c Thu Apr 12 13:44:02 2007 -0700
> +++ b/arch/arm/boot/compressed/misc.c Thu Apr 12 15:48:43 2007 -0700
> @@ -239,7 +239,7 @@ static ulg
On Thu, Apr 12, 2007 at 01:50:54PM -0700, Jeremy Fitzhardinge wrote:
> Subject: deflate stack usage in lib/inflate.c
>
> inflate_fixed and huft_build together use around 2.7k of stack. When
> using 4k stacks, I saw stack overflows from interrupts arriving while
> unpacking the root initrd:
>
> d
Andi Kleen wrote:
> Yes, but then we should have seen more frequently, shouldn't we? I always
> run with the stack overflow check enabled and I don't think I ever saw
> warnings in inflate.
>
I guess the window is just while decompressing the root filesystem.
Interrupts under Xen might be usi
On Friday 13 April 2007 01:20:40 Jan Engelhardt wrote:
>
> On Apr 12 2007 15:39, Jeremy Fitzhardinge wrote:
> >Andi Kleen wrote:
> >> Hmm, does Xen perhaps not use interrupt stacks? Normally 2.7k should be
> >> still
> >> green as long as there are not too many functions above/below it.
> >
> >Th
Jan Engelhardt wrote:
> On Apr 12 2007 15:39, Jeremy Fitzhardinge wrote:
>
>> Andi Kleen wrote:
>>
>>> Hmm, does Xen perhaps not use interrupt stacks? Normally 2.7k should be
>>> still
>>> green as long as there are not too many functions above/below it.
>>>
>> That's a good point,
On Apr 12 2007 15:39, Jeremy Fitzhardinge wrote:
>Andi Kleen wrote:
>> Hmm, does Xen perhaps not use interrupt stacks? Normally 2.7k should be still
>> green as long as there are not too many functions above/below it.
>
>That's a good point, I'll need to check that. Still, nearly 3k of stack!
I
On Thu, Apr 12, 2007 at 03:57:48PM -0700, Jeremy Fitzhardinge wrote:
> Matt Mackall wrote:
> > On Thu, Apr 12, 2007 at 01:50:54PM -0700, Jeremy Fitzhardinge wrote:
> >
> >> -#define HEAP_SIZE 0x3000
> >> +#define HEAP_SIZE 0x4000
> >>
> >
> > There are a bunch more o
Jeremy Fitzhardinge wrote:
> Andi Kleen wrote:
>>> (This was under Xen, but there's no reason it couldn't happen on bare
>>> hardware.)
>>>
>> Hmm, does Xen perhaps not use interrupt stacks?
>
> Looks like that's all done in do_IRQ, so it should be independent of
> whether its Xen or not.
On Friday 13 April 2007 00:56:56 Jeremy Fitzhardinge wrote:
> Andi Kleen wrote:
> >> (This was under Xen, but there's no reason it couldn't happen on bare
> >> hardware.)
> >>
> >
> > Hmm, does Xen perhaps not use interrupt stacks?
>
> Looks like that's all done in do_IRQ, so it should be i
Matt Mackall wrote:
> On Thu, Apr 12, 2007 at 01:50:54PM -0700, Jeremy Fitzhardinge wrote:
>
>> -#define HEAP_SIZE 0x3000
>> +#define HEAP_SIZE 0x4000
>>
>
> There are a bunch more of these that'll need fixing.
>
Like this?
diff -r 2ad8a0729f26 arch/alpha/boot/m
Andi Kleen wrote:
>> (This was under Xen, but there's no reason it couldn't happen on bare
>> hardware.)
>>
>
> Hmm, does Xen perhaps not use interrupt stacks?
Looks like that's all done in do_IRQ, so it should be independent of
whether its Xen or not. And the stack overflow check is perfo
Andi Kleen wrote:
> Hmm, does Xen perhaps not use interrupt stacks? Normally 2.7k should be still
> green as long as there are not too many functions above/below it.
>
That's a good point, I'll need to check that. Still, nearly 3k of stack!
J
-
To unsubscribe from this list: send the line
On Thu, Apr 12, 2007 at 01:50:54PM -0700, Jeremy Fitzhardinge wrote:
> -#define HEAP_SIZE 0x3000
> +#define HEAP_SIZE 0x4000
There are a bunch more of these that'll need fixing.
--
Mathematics is the supreme nostalgia of our time.
-
To unsubscribe from this list: send the
> (This was under Xen, but there's no reason it couldn't happen on bare
> hardware.)
Hmm, does Xen perhaps not use interrupt stacks? Normally 2.7k should be still
green as long as there are not too many functions above/below it.
-Andi
-
To unsubscribe from this list: send the line "unsubscribe
Subject: deflate stack usage in lib/inflate.c
inflate_fixed and huft_build together use around 2.7k of stack. When
using 4k stacks, I saw stack overflows from interrupts arriving while
unpacking the root initrd:
do_IRQ: stack overflow: 384
[] show_trace_log_lvl+0x1a/0x30
[] show_trace+0x12/0x1
15 matches
Mail list logo
