6.0.0 crash at high precision mul?

Torbjörn Granlund tg at gmplib.org
Fri Jun 6 12:01:47 UTC 2014

bodrato at mail.dm.unipi.it writes:

  I keep on thinking it's not a bad idea...
We never brought up the aspect of total allocation size back then.
I'd suggest using something like

if (total_size > 65536)
  xp = TMP_BALLOC (...);
  yp = TMP_BALLOC (...);
  zp = TMP_BALLOC (...);
  xp = TMP_SALLOC (...);
  yp = TMP_SALLOC (...);
  zp = TMP_SALLOC (...);
in relevant places.

  But we this thread was started because of a crash occurring at symbol
  __gmpn_toom8h_mul ... toom8 does not allocate temporaries, and mul should
  allocate it all at once...
But then the crash will happen when unmapped stack is actually used.

Please encrypt, key id 0xC8601622

More information about the gmp-discuss mailing list