.skip 16
RC4:
.prologue
- .fframe 0
.save ar.pfs,r2
- .save ar.lc,r3
- .save pr,prsave
{ .mii; alloc r2=ar.pfs,4,12,0,16
+ .save pr,prsave
mov prsave=pr
ADDP key=0,in0 };;
{ .mib; cmp.eq p6,p0=0,in1 // len==0?
+ .save ar.lc,r3
mov r3=ar.lc
(p6) br.ret.spnt.many b0 };; // emergency exit
* implementations suffer from significant performance
* losses then, e.g. PIII exhibits >2x deterioration,
* and so does Opteron. In order to assure optimal
- * all-round performance, let us [try to] detect P4 at
- * run-time by checking upon HTT bit in CPU capability
+ * all-round performance, we detect P4 at run-time by
+ * checking upon reserved bit 20 in CPU capability
* vector and set up compressed key schedule, which is
* recognized by correspondingly updated assembler
- * module...
+ * module... Bit 20 is set up by OPENSSL_ia32_cpuid.
+ *
* <appro@fy.chalmers.se>
*/
#ifdef OPENSSL_FIPS