[PATCH] arm/crypto: Make asm SHA-1 and AES code Thumb-2 compatible

[dave.martin@linaro.org (Dave Martin)]

On Mon, Nov 26, 2012 at 10:45:27AM +1000, David McCullough wrote:
> 
> Nicolas Pitre wrote the following:
> > On Tue, 20 Nov 2012, Dave Martin wrote:
> > 
> > > On Thu, Nov 15, 2012 at 05:26:44PM -0500, Nicolas Pitre wrote:
> > > > On Mon, 5 Nov 2012, Dave Martin wrote:
> > > > 
> > > > > This patch fixes aes-armv4.S and sha1-armv4-large.S to work
> > > > > natively in Thumb.  This allows ARM/Thumb interworking workarounds
> > > > > to be removed.
> > > > > 
> > > > > I also take the opportunity to convert some explicit assembler
> > > > > directives for exported functions to the standard
> > > > > ENTRY()/ENDPROC().
> > > > > 
> > > > > For the code itself:
> > > > > 
> > > > >   * In sha1_block_data_order, use of TEQ with sp is deprecated in
> > > > >     ARMv7 and not supported in Thumb.  For the branches back to
> > > > >     .L_00_15 and .L_40_59, the TEQ is converted to a CMP, under the
> > > > >     assumption that clobbering the C flag here will not cause
> > > > >     incorrect behaviour.
> > > > > 
> > > > >     For the first branch back to .L_20_39_or_60_79 the C flag is
> > > > >     important, so sp is moved temporarily into another register so
> > > > >     that TEQ can be used for the comparison.
> > > > > 
> > > > >   * In the AES code, most forms of register-indexed addressing with
> > > > >     shifts and rotates are not permitted for loads and stores in
> > > > >     Thumb, so the address calculation is done using a separate
> > > > >     instruction for the Thumb case.
> > > > > 
> > > > >     The resulting code is unlikely to be optimally scheduled, but
> > > > >     it should not have a large impact given the overall size of the
> > > > >     code.  I haven't run any benchmarks.
> > > > > 
> > > > > Signed-off-by: Dave Martin <dave.martin@linaro.org>
> > > > 
> > > > Acked-by: Nicolas Pitre <nico@linaro.org>
> > > > 
> > > > I didn't test it either, only reviewed the patch.  Looks obvious enough.  
> > > > And if something is wrong, then it is very unlikely to be unnoticed in 
> > > > practice.
> > > 
> > > I'd prefer someone tests this before I upload to the patch system.
> > > 
> > > I can do that, but I'm busy so it won't happen quickly...
> > > 
> > > Or do you think I should just go ahead?
> > 
> > If no one has provided test confirmation so far, that's either because 
> > no one cares, or any breakage in functionality would be so obvious that 
> > the incentive for testing before this hits mainline is not there.  In 
> > either cases that should be fine for you to go ahead and let any fixes, 
> > if any, be sent during the -rc period.
> 
> Ok,  I have tested this on little endian Kendin Micrel 8695 and on
> a big endian Intel Xscale 425.  Both worked fine.  I don't have any thumb
> systems,  but the patch can only improve thumb support since there was none
> previously :-)
> 
> Cheers,
> Davidm
> 
> Tested-by: David McCullough <ucdevel@gmail.com>
> Acked-by: David McCullough <ucdevel@gmail.com>

I didn't get time to test this for Thumb, but since it works for ARM and
the change is straightforward, I've sent it to the patch system now.
Apologies for the delay.  (Thanks to Nico for the reminder!)

Cheers
---Dave