Audigy problems.

[Walter]

Matthew Dillon wrote:
:Up to two weeks ago, that version of the snd_emu10k1.ko driver would
:load under DF and run without any (noticable) problems.  I updated my
:copy of the source tree yesterday, and rebuilt world and the kernel.
:Unfortunately, it seems that audigy driver no longer works, but I have
:no idea how to track down the source of the problem.  This is the kernel
:panic andtrace:
    Ok, this is a real simple fix.  Your driver is calling malloc() without
    either the M_WAITOK or M_NOWAIT flags.  It must call malloc() with one
    or the other.  If it is not using either flag then add M_WAITOK to it.
I know I'm being a bit of a butt-in-ski here, but wouldn't
it be the safer, more backwards compatible path to make
malloc() internally use the M_WAITOK logic if neither that
flag or M_NOWAIT was explicitly set?  Otherwise isn't there
always the risk that an incoming SW bundle will break upon
entering DfBSD space? (Please ignore me if I'm wrong...)