Re: New DRM driver model - gets rid of DRM() macros!

[Keith W. <ke...@tu...>]

Keith Whitwell wrote:
> Christoph Hellwig wrote:
>=20
>> On Wed, Sep 29, 2004 at 02:29:24PM +0100, Keith Whitwell wrote:
>>
>>> Christoph Hellwig wrote:
>>>
>>>
>>>> - drm_flush is a noop.  a NULL ->flush does the same thing, just eas=
ier
>>>> - dito or ->poll
>>>> - dito for ->read
>>>
>>>
>>> Pretty sure you couldn't get away with null for these in 2.4, at leas=
t.
>>
>>
>>
>> Umm, of course you could.  There's only a hanfull instance defining a
>> ->flush at all.  Similarly all file_ops for regular files and many cha=
r
>> devices don't have ->poll.  no ->read is pretty rare but 2.4 ch=E6cks =
it
>> aswell.
>=20
>=20
> I tried it, led to crashes (panics, I guess) & the change had to be=20
> reverted.  On reverting the crashes stopped.  This was for poll and rea=
d:

Thinking about it, it may not have been a problem of crashing, but rather=
 that=20
  the behaviour visible from a program attempting to read (or poll) was=20
different with noop versions of these functions to NULL versions, and tha=
t was=20
causing problems.  This is 18 months ago, so yes, I'm being vague.

The X server does look at this file descriptor, which is where the proble=
m=20
would have arisen, but only the gamma & maybe ffb drivers do anything wit=
h it.

Keith