On Wed, 2007-04-25 at 13:11 +0530, jimmy wrote:
> Robert Lill wrote:
> > Hello Haubi and all,
> >
> > the explanation sounds reasonable, but the suggested solution doesn't work.
> > I don't understand the assembler mnemonics so I just copied that section
> > finding out, that padlock_cipher_data * cdata doesn't have members
> > 'len' and 'addr'.
> >
> > Any suggestion how to get around this lack of registers or to use a
> > C implementation for PADLOCK_XCRYPT_ASM?
For my $0.02 (forgive me if this is nonsense, I do not fully understand
parameter-definition for inline-assembly yet):
Question is what ".byte 0xf3,0x0f,0xa7,0xc8" (repz xcrypt-ecb) needs to
be set up on the stack besides %ecx, %edx, %ebx, %esi, %edi and %eax as
returnvalue.
Because when I remove definition of "%1" to "%3", it can be compiled
with '-fPIC -O0':
asm volatile (
"pushl %%ebx\n"
" leal 16(%0),%%edx\n"
" leal 32(%0),%%ebx\n"
".byte 0xf3,0x0f,0xa7,0xc8" "\n"
" popl %%ebx"
: "=a"(iv), "=c"(cnt), "=D"(out), "=S"(inp)
- : "0"(cdata), "1"(cnt), "2"(out), "3"(inp), "m"(*cdata)
+ : "0"(cdata), "m"(*cdata)
: "edx", "cc", "memory"
);
> >
> > Robert
>
> sounds like you should use no-asm in configure (unless you insist on
> assembly in the debug lib)
Basically the problem comes from gentoo, and a fix is available there
now: http://thread.gmane.org/gmane.linux.gentoo.alt/2351/focus=2403
But anyway, it should also be compileable with '-fPIC -O0' IMO.
/haubi/
--
Salomon Automation GmbH - Friesachstrasse 15 - A-8114 Friesach bei Graz
Sitz der Gesellschaft: Friesach bei Graz
UID-NR:ATU28654300 - Firmenbuchnummer: 49324 K
Firmenbuchgericht: Landesgericht für Zivilrechtssachen Graz
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List [email protected]
Automated List Manager [EMAIL PROTECTED]
