time for 1.5.11? (/dev/dsp problems -- Gerd take note)

Christopher Faylor me@cgf.cx
Thu Jun 10 02:46:00 GMT 2004

On Wed, Jun 09, 2004 at 09:18:45PM -0400, Igor Pechtchanski wrote:
>While we're on the topic, here's what seems like another potential bug
>(just from eyeballing the code):
>fhandler_dev_dsp::open (int flags, mode_t mode)
>  open_count++;
>  if (open_count > 1)
>    {
>      set_errno (EBUSY);
>      return 0;
>      ^^^^^^^^^
>    }
>If the device is busy, open_count will never get decremented (it's only
>decremented on close(), and who'd think of closing an fd that failed to
>open?), so nobody will ever be able to open() it afterwards, even if it
>becomes available.  I believe open_count should be decremented just
>before the return above.  Comments?

Just one: Is there any hope for Red Hat actually giving you an
assignment?  Or is that still stalled?

I asked Red Hat's legal dept to look over the cygwin assignment form a
couple of months before I left to see if it could be made more better.
I think Corinna has carried on the fight.  As you all know, however,
Cygwin is far from a core interest of Red Hat's.  Even still the VP I
was talking to was really a great guy and very responsive on other
matters, so there is always hope, I guess.


More information about the Cygwin-developers mailing list