sparse_irq: userspace ABI breakage in -tip tree

sparse_irq: userspace ABI breakage in -tip tree

[Arjan van de Ven]

commit 70426a06a58561c401e64c5767d71a011d4b978b
Author: Yinghai Lu <yhlu.kernel@gmail.com>
Date:   Tue Aug 19 20:50:27 2008 -0700

    x86: make 32 bit to use sparse_irq

@@ -285,7 +303,7 @@ int show_interrupts(struct seq_file *p, void *v)
                action = desc->action;
                if (!action && !any_count)
                        goto skip;
-               seq_printf(p, "%3d: ",i);
+               seq_printf(p, "%#x: ",i);
 #ifndef CONFIG_SMP
                seq_printf(p, "%10u ", kstat_irqs(i));
 #else


this is a userspace ABI breakage; you're changing /proc/interrupts from
decimal to hex! This file is used by both humans and computer programs,
and the humans will get confused, but the computer programs will break.
Not a good idea to put into mainline.

(I'm not so convinced of sparse_irq as a whole to be honest, but that's
a different story)

Re: sparse_irq: userspace ABI breakage in -tip tree

[Yinghai Lu]

On Fri, Sep 19, 2008 at 3:53 PM, Arjan van de Ven <arjan@infradead.org> wrote:
>
> commit 70426a06a58561c401e64c5767d71a011d4b978b
> Author: Yinghai Lu <yhlu.kernel@gmail.com>
> Date:   Tue Aug 19 20:50:27 2008 -0700
>
>    x86: make 32 bit to use sparse_irq
>
> @@ -285,7 +303,7 @@ int show_interrupts(struct seq_file *p, void *v)
>                action = desc->action;
>                if (!action && !any_count)
>                        goto skip;
> -               seq_printf(p, "%3d: ",i);
> +               seq_printf(p, "%#x: ",i);
>  #ifndef CONFIG_SMP
>                seq_printf(p, "%10u ", kstat_irqs(i));
>  #else
>
>
> this is a userspace ABI breakage; you're changing /proc/interrupts from
> decimal to hex! This file is used by both humans and computer programs,
> and the humans will get confused, but the computer programs will break.
> Not a good idea to put into mainline.

change back to decimal could help?

>
> (I'm not so convinced of sparse_irq as a whole to be honest, but that's
> a different story)

anything need to be addressed?

YH

Re: sparse_irq: userspace ABI breakage in -tip tree

[Arjan van de Ven]

On Fri, 19 Sep 2008 17:10:42 -0700
"Yinghai Lu" <yhlu.kernel@gmail.com> wrote:

> On Fri, Sep 19, 2008 at 3:53 PM, Arjan van de Ven
> <arjan@infradead.org> wrote:
> >
> > commit 70426a06a58561c401e64c5767d71a011d4b978b
> > Author: Yinghai Lu <yhlu.kernel@gmail.com>
> > Date:   Tue Aug 19 20:50:27 2008 -0700
> >
> >    x86: make 32 bit to use sparse_irq
> >
> > @@ -285,7 +303,7 @@ int show_interrupts(struct seq_file *p, void *v)
> >                action = desc->action;
> >                if (!action && !any_count)
> >                        goto skip;
> > -               seq_printf(p, "%3d: ",i);
> > +               seq_printf(p, "%#x: ",i);
> >  #ifndef CONFIG_SMP
> >                seq_printf(p, "%10u ", kstat_irqs(i));
> >  #else
> >
> >
> > this is a userspace ABI breakage; you're changing /proc/interrupts
> > from decimal to hex! This file is used by both humans and computer
> > programs, and the humans will get confused, but the computer
> > programs will break. Not a good idea to put into mainline.
> 
> change back to decimal could help?

that's a minimum; someone (you?) needs to check if there are other
things changed in the format of this file....



-- 
Arjan van de Ven 	Intel Open Source Technology Centre
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org

Re: sparse_irq: userspace ABI breakage in -tip tree

[Yinghai Lu]

On Fri, Sep 19, 2008 at 5:44 PM, Arjan van de Ven <arjan@infradead.org> wrote:
> On Fri, 19 Sep 2008 17:10:42 -0700
> "Yinghai Lu" <yhlu.kernel@gmail.com> wrote:
>
>> On Fri, Sep 19, 2008 at 3:53 PM, Arjan van de Ven
>> <arjan@infradead.org> wrote:
>> >
>> > commit 70426a06a58561c401e64c5767d71a011d4b978b
>> > Author: Yinghai Lu <yhlu.kernel@gmail.com>
>> > Date:   Tue Aug 19 20:50:27 2008 -0700
>> >
>> >    x86: make 32 bit to use sparse_irq
>> >
>> > @@ -285,7 +303,7 @@ int show_interrupts(struct seq_file *p, void *v)
>> >                action = desc->action;
>> >                if (!action && !any_count)
>> >                        goto skip;
>> > -               seq_printf(p, "%3d: ",i);
>> > +               seq_printf(p, "%#x: ",i);
>> >  #ifndef CONFIG_SMP
>> >                seq_printf(p, "%10u ", kstat_irqs(i));
>> >  #else
>> >
>> >
>> > this is a userspace ABI breakage; you're changing /proc/interrupts
>> > from decimal to hex! This file is used by both humans and computer
>> > programs, and the humans will get confused, but the computer
>> > programs will break. Not a good idea to put into mainline.
>>
>> change back to decimal could help?
>
> that's a minimum; someone (you?) needs to check if there are other
> things changed in the format of this file....

does the ABI require the irq number in ascend sequence in /proc/interrupts?

also how about /proc/stat?

YH