Re: [PATCH] Improve mutt's random data generator.

[Kevin J. McCarthy]

On Thu, Apr 23, 2026 at 10:12:25PM -0400, Kurt Hackenberg wrote:

On Fri, Apr 24, 2026 at 09:48 +0800, Kevin J. McCarthy wrote:

Ah, sorry.  Yes, the code is executed.  The whole snippet is:

if ((res == (size_t) -1) || (res < length_requested))
{
  if (res < length_requested)
  {
    length_requested -= res;
    random_bytes += res;
  }
  prng_random_bytes(random_bytes, length_requested);
}

I don't know what prng_random_bytes() is, so don't quite know the intent here. But counting on (res < length_requested) when getrandom() failed seems like pushing C pretty hard.

The man page says:



On error, -1 is returned, and errno is set to indicate the error.



The call to prng_random_bytes() was falling back on the pseudo random
generator.



Thanks for your feedback, Kurt and Alex!

--
Kevin J. McCarthy
GPG Fingerprint: 8975 A9B3 3AA3 7910 385C  5308 ADEF 7684 8031 6BDA

On success, getrandom() returns the number of bytes that were copied to the buffer buf. This may be less than the number of bytes requested via size if either GRND_RANDOM was specified in flags and insufficient entropy was present in the random source or the system call was interrupted by a signal. So that outer "if" checks for one of two possibilities we need to correct for: failure (-1), or less than the number bytes requests via size. However, I'm think I'll take the fallback out, and just loop the call to getrandom() until we fill the buffer up. I'll send a v2 patch today or tomorrow.

signature.asc
Description: PGP signature