On Sun, 30 Mar 2008, Pekka Enberg wrote:

quoted text
> Christoph, I think you're overlooking the same thing I was until Linus
> straightened me out. We're calling kmalloc_large() from __slab_alloc()
> for the fall-back case which causes a bug fixed by Linus' revert.

Yes, that means that __GFP_ZERO can be passed to the page allocator there. 
Same thing as happens when kmalloc() calls kmalloc_large() (as you pointed 
out before). The harm is zeroing an allocated object twice in rare cases 
that trigger the fallback path and that is avoided by Linus' patch.

For the fix to get rid of the backtraces that started this discussion see
commit 9c312058b2e530722c7bd30c1b6f26eea35dc5fe

--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/