On Mon, Oct 24, 2005 at 05:57:52PM +0000, Nicolas Pitre wrote: > > The final length buffer construction in sha1_final() is moved just before > where it is used. This makes for slightly more readable code and actually > more efficient as well since the compiler doesn't have to bother preserving > the bits array across the first call to sha1_update(). In fact it makes for > slightly smaller stack usage on i386 or slightly smaller code size on ARM > (depending on your GCC version that could vary of course).
This change no longer reduces code size/stack usage with the cpu_to_be64 change that Denis initiated. Since this patch actually separates the two sctx->count references across the call to sha1_update, which means that the store to the temp count variable is either going to chew up two saved registers or is going sit on the stack, I've decided to not apply this patch. Thanks, -- Visit Openswan at http://www.openswan.org/ Email: Herbert Xu ~{PmV>HI~} <[EMAIL PROTECTED]> Home Page: http://gondor.apana.org.au/~herbert/ PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt - To unsubscribe from this list: send the line "unsubscribe linux-crypto" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
