On Sat, 16 Mar 2002 06:51:35 -0500,
Paul Gortmaker <p_gortmaker@yahoo.com> wrote:
>I had a think on this from the perspective of increasing UP performance
>of SMP kernels, and came up with the following rather interesting (IMHO)
>patch. Executive summary is that when a SMP kernel finds itself on a UP
>box, it modifies itself (ooohh!) by going along and essentially doing a
>sed '_text,_etext s/lock/nop/' :)
>
>Details: Address of each relevant lock opcode is stored (similar to
>the way an exception table is)
Does your patch work with recent binutils? I suspect that any lock
code in discarded exit sections will cause binutils to barf (loudly).
Look at the out of line changes to the lock code in 2.4.18-pre3, I had
to stop using a single lock section because of the more rigorous
binutils checks.
-
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/
This archive was generated by hypermail 2b29 : Sat Mar 23 2002 - 22:00:11 EST