On Mon, Sep 23, 2013 at 11:00 AM, Tsutomu Oyamada
<[email protected]>wrote:

> Hi, all.
>
> We are using ClamAV for IBM zLinux (s390x architecture)
> So many Warnings were output as follows, when we made a new build
> with using the new release 0.98.
>
> # ./configure --prefix=/usr/lib/clamav --exec-prefix=/usr/lib/clamav
> --bindir=/usr/lib/clamav --sbindir=/usr/lib/clamav --sysconfdir=/etc/clamav
> --libdir=/usr/lib/clamav --datarootdir=/usr/lib/clamav
> --with-dbdir=/usr/lib/clamav --disable-clamav --with-zlib=/usr/local
> --with-libbz2-prefix=/usr/local
> (snip)
> # make
> (snip)
> tomsfastmath/mul/fp_mul_comba_20.c:91: warning: right shift count >= width
> of type
> tomsfastmath/mul/fp_mul_comba_20.c:91: warning: right shift count >= width
> of type
> tomsfastmath/mul/fp_mul_comba_20.c:91: warning: right shift count >= width
> of type
>
> And, the binary did not work properly as a result.
>
> Is there any special settings in a case of make for s390?
> Can I have any advice/suggestion?
>
> Best Regards,
> Oyamada
>
> _______________________________________________
> http://lurker.clamav.net/list/clamav-devel.html
> Please submit your patches to our Bugzilla: http://bugs.clamav.net
>

That line (and other ones like it in fp_mul_comba_20.c) are doing
calculations using an array of elements of type "fp_digit". "fp_digit" is
defined in libclamav/bignum_fast.h.

I don't have access to a s390 to test with and I don't see any notes on
tomsfastmath and s390, so I will need your help to investigate. Can you get
me some values as calculated by bignum_fast.h?

1) Sizes of two types: sizeof(fp_digit), sizeof(unsigned long)
2) Values of these defined macros, if defined: FP_64BIT, DIGIT_BIT,
CHAR_BIT, FP_MAX_SIZE, FP_SIZE, TFM_ASM

Thanks,

Dave R.

-- 
---
Dave Raynor
Sourcefire Vulnerability Research Team
[email protected]
_______________________________________________
http://lurker.clamav.net/list/clamav-devel.html
Please submit your patches to our Bugzilla: http://bugs.clamav.net

Reply via email to