* Andi Kleen <[email protected]> wrote:
> > + : "rax", "rsi", "rdx", "rcx", \
> > + "r8", "r9", "r10", "r11", "memory"); \
>
> I think it would be still better if you used the stubs in
> arch/x86_64/lib/thunk.S and not clobber all the registers. While it
> won't make that much difference for the out of line mutexes it will
> generate better code for inline mutexes, and if someone ever decides
> they're a good idea the code will be ready.
i didnt want to add it to thunk.S because right now it would cause an
unnecessary slowdown for the slowpath, by quite a number of
instructions: due to the indiscriminate register-saving/restoring done
in thunk.S.
even though it's a "slow path" relative to the fastpath, we shouldnt
slow it down unnecessarily. So if someone wants to play with more
inlining later on, this has to be done in context of that effort.
Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
