[PATCH] ia64: spill_mask generation corrections

Jan Beulich JBeulich@novell.com
Tue Jul 26 16:22:00 GMT 2005


>This looks OK, though I think you are trying to be too clever in a few
>places.
>
>For instance, in check_pending_save, you have
>+       else
>+         cur = (prev = cur)->next;
>You are trying too hard to avoid using braces here.
>
>In 5 different places, you have inside a loop
>+      mask &= ~(mask & (~mask + 1));
>+      if (!mask)
>+       return ptr;
>+      cur = alloc_record (fr_mem);
>+      cur->r.record.p.frmask = mask;
>+      prev->r.record.p.frmask ^= mask;
>I understand why you wrote it that way, but I don't think you can
safely
>expect that everyone looking at this code in the next 20 years will
>understand it.  There should be comments explaining what it is doing.

>First you clear off the low order bit of mask.  Then you xor it into
>prev frmask so that prev frmask is left with only the low order bit. 
>Then you repeat until no bits are left.

Jim,

I'm just unsure, did your (above) reply mean once adjusted I can commit
the patch (as posted to the ml a week ago), or do I have to wait for
another approval in such a case?

Jan
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: binutils-mainline-ia64-unwind-spill-mask.patch
URL: <https://sourceware.org/pipermail/binutils/attachments/20050726/191df052/attachment.ksh>


More information about the Binutils mailing list