*sigh* I really shouldn't have mailed this out in the first place. Every time I look at it I find new bugs despite the fact that it actually works on my machine.
> ld r3,SOFTE(r1)
> - bl .local_irq_restore
> +#ifdef CONFIG_TRACE_IRQFLAGS
> + cmpdi r3,0
> + beq 14f
> + bl .trace_hardirqs_on
> + ld r3,SOFTE(r1)
> +14:
> + bl .raw_local_irq_restore
> + cmpdi r3,0
This should reload r3 before the second compare as raw_local_irq_restore
returns nothing (void). And why does it need a second compare anyway?
Just rewrite as
ld r3, SOFTE(r1)
#ifdef CONFIG_TRACE_IRQFLAGS
cmpdi r3, 0
bne 14f
bl .raw_local_irq_restore
bl .trace_hardirqs_off
b 15f
14:
bl .trace_hardirqs_on
li r3, 1
#endif
bl .raw_local_irq_restore
15:
which has the advantage of having only one conditional branch and less
preprocessor foo.
> +#ifdef CONFIG_TRACE_IRQFLAGS
> + cmpdi r5,0
> + beq 5f
> + bl .trace_hardirqs_on
> + ld r5,SOFTE(r1)
> stb r5,PACASOFTIRQEN(r13)
> -
> + b 6f
> +5:
> + stb r5,PACASOFTIRQEN(r13)
> + bl .trace_hardirqs_off
> +6:
> +#else
> + stb r5,PACASOFTIRQEN(r13)
> +#endif
Similarly, that could be rewritten as
#ifdef CONFIG_TRACE_IRQFLAGS
cmpdi r5, 0
bne 5f
stb r5, PACASOFTIRQEN(r13)
bl .trace_hardirqs_off
b 6f
5:
bl .trace_hardirqs_on
li r5, 1
#endif
stb r5, PACASOFTIRQEN(r13)
6:
I can't test these modifications over the weekend but they should be
fine.
johannes
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Linuxppc-dev mailing list Linuxppc-dev@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-dev
