* drm - first steps towards 64-bit correctness..
@ 2004-07-31 9:13 Dave Airlie
2004-07-31 9:32 ` Eric Anholt
0 siblings, 1 reply; 6+ messages in thread
From: Dave Airlie @ 2004-07-31 9:13 UTC (permalink / raw)
To: linux-kernel; +Cc: dri-devel
Hi,
As a first step towards sorting getting the DRM into shape for
proper use on 32/64-bit systems, I'd like to sort out all the type
definitions in drivers/char/drm/drm.h, this file is also included in
userspace and BSD builds...
After reading the thread "32/64bit issues in ioctl struct passing" on
dri-devel, I'm still not 100% sure what we need to do, I just know we to
do something sooner rather than later!! we are getting more and more
32/64-bit users everyday....
While avoiding breakage of current users is "a good thing" I'm not sure it
overrides "getting it right", at the moment mixed 32/64-bit is broken for
most cards anyways... I'd like to try and not break pure-32 or pure-64 bit
setups alright but I think pure-64 bit might take some collateral damage
:-(..
I've looked across the SuSE patch[1] for 64-bit, but it looks like it will
add complexity and making future maintenance nightmareish...
We do need to sort this out ASAP, and I also would like to say I'm
probably not the best person to do the work, I've no non-32bit hardware to
test this stuff on, I've little 32/64 mixed environment experience,
everytime I think I've grasped the issues I dig a bit further :-), though
I also believe this is the single biggest issue with the DRM currently (as
the maintainer..)
Regards,
Dave.
[1] http://marc.theaimsgroup.com/?l=dri-devel&m=109040871011904&w=2
--
David Airlie, Software Engineer
http://www.skynet.ie/~airlied / airlied at skynet.ie
pam_smb / Linux DECstation / Linux VAX / ILUG person
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: drm - first steps towards 64-bit correctness..
2004-07-31 9:13 drm - first steps towards 64-bit correctness Dave Airlie
@ 2004-07-31 9:32 ` Eric Anholt
2004-07-31 9:54 ` Arjan van de Ven
0 siblings, 1 reply; 6+ messages in thread
From: Eric Anholt @ 2004-07-31 9:32 UTC (permalink / raw)
To: Dave Airlie; +Cc: linux-kernel, DRI
On Sat, 2004-07-31 at 02:13, Dave Airlie wrote:
> Hi,
> As a first step towards sorting getting the DRM into shape for
> proper use on 32/64-bit systems, I'd like to sort out all the type
> definitions in drivers/char/drm/drm.h, this file is also included in
> userspace and BSD builds...
>
> After reading the thread "32/64bit issues in ioctl struct passing" on
> dri-devel, I'm still not 100% sure what we need to do, I just know we to
> do something sooner rather than later!! we are getting more and more
> 32/64-bit users everyday....
> While avoiding breakage of current users is "a good thing" I'm not sure it
> overrides "getting it right", at the moment mixed 32/64-bit is broken for
> most cards anyways... I'd like to try and not break pure-32 or pure-64 bit
> setups alright but I think pure-64 bit might take some collateral damage
> :-(..
>
> I've looked across the SuSE patch[1] for 64-bit, but it looks like it will
> add complexity and making future maintenance nightmareish...
>
> We do need to sort this out ASAP, and I also would like to say I'm
> probably not the best person to do the work, I've no non-32bit hardware to
> test this stuff on, I've little 32/64 mixed environment experience,
> everytime I think I've grasped the issues I dig a bit further :-), though
> I also believe this is the single biggest issue with the DRM currently (as
> the maintainer..)
I've got a 64-bit cleanliness patch for SiS that I'd like to either get
someone else to review (preferable) or find time to re-read myself. I
replace the memory management code that existed with much less code
(using bsd queue macros that I'm more familiar with).
Once the flames die down in X.Org I'll isolate the diff from my very
dirty local drm tree and post it again.
I'm hoping that most of the general DRM fixes will be replacing longs
with fixed-size types that are the same on x86, but I haven't looked at
Egbert's diffs yet, unfortunately. As long as you don't use the linux-y
"u32"-type types, BSD should be happy with the changes.
--
Eric Anholt eta@lclark.edu
http://people.freebsd.org/~anholt/ anholt@FreeBSD.org
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: drm - first steps towards 64-bit correctness..
2004-07-31 9:32 ` Eric Anholt
@ 2004-07-31 9:54 ` Arjan van de Ven
2004-07-31 9:57 ` Eric Anholt
0 siblings, 1 reply; 6+ messages in thread
From: Arjan van de Ven @ 2004-07-31 9:54 UTC (permalink / raw)
To: Eric Anholt; +Cc: Dave Airlie, linux-kernel, DRI
[-- Attachment #1: Type: text/plain, Size: 257 bytes --]
On Sat, 2004-07-31 at 11:32, Eric Anholt wrote:
> As long as you don't use the linux-y
> "u32"-type types, BSD should be happy with the changes.
can you explain why u32 would be outlawed? Surely it's trivial to do a
typedef for u32 on BSD for drm ??
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: drm - first steps towards 64-bit correctness..
2004-07-31 9:54 ` Arjan van de Ven
@ 2004-07-31 9:57 ` Eric Anholt
2004-07-31 10:06 ` William Lee Irwin III
2004-07-31 11:02 ` Dave Airlie
0 siblings, 2 replies; 6+ messages in thread
From: Eric Anholt @ 2004-07-31 9:57 UTC (permalink / raw)
To: arjanv; +Cc: Dave Airlie, linux-kernel, DRI
On Sat, 2004-07-31 at 02:54, Arjan van de Ven wrote:
> On Sat, 2004-07-31 at 11:32, Eric Anholt wrote:
> > As long as you don't use the linux-y
> > "u32"-type types, BSD should be happy with the changes.
>
> can you explain why u32 would be outlawed? Surely it's trivial to do a
> typedef for u32 on BSD for drm ??
If there are nice standard types (uint32_t or u_int32_t, can't remember
which at the moment, I mentioned it in an email some time ago) out there
already that linux has too, why not use those?
--
Eric Anholt eta@lclark.edu
http://people.freebsd.org/~anholt/ anholt@FreeBSD.org
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: drm - first steps towards 64-bit correctness..
2004-07-31 9:57 ` Eric Anholt
@ 2004-07-31 10:06 ` William Lee Irwin III
2004-07-31 11:02 ` Dave Airlie
1 sibling, 0 replies; 6+ messages in thread
From: William Lee Irwin III @ 2004-07-31 10:06 UTC (permalink / raw)
To: Eric Anholt; +Cc: arjanv, Dave Airlie, linux-kernel, DRI
On Sat, 2004-07-31 at 02:54, Arjan van de Ven wrote:
>> can you explain why u32 would be outlawed? Surely it's trivial to do a
>> typedef for u32 on BSD for drm ??
On Sat, Jul 31, 2004 at 02:57:17AM -0700, Eric Anholt wrote:
> If there are nice standard types (uint32_t or u_int32_t, can't remember
> which at the moment, I mentioned it in an email some time ago) out there
> already that linux has too, why not use those?
uint*_t. ISTR something about Linux' usage predating standard type
names for the things.
I have much more serious issues with other naming conventions to get
worked up about this one. In general I don't mind ones that are less
verbose than the standard.
-- wli
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: drm - first steps towards 64-bit correctness..
2004-07-31 9:57 ` Eric Anholt
2004-07-31 10:06 ` William Lee Irwin III
@ 2004-07-31 11:02 ` Dave Airlie
1 sibling, 0 replies; 6+ messages in thread
From: Dave Airlie @ 2004-07-31 11:02 UTC (permalink / raw)
To: Eric Anholt; +Cc: arjanv, linux-kernel, DRI
> > can you explain why u32 would be outlawed? Surely it's trivial to do a
> > typedef for u32 on BSD for drm ??
>
> If there are nice standard types (uint32_t or u_int32_t, can't remember
> which at the moment, I mentioned it in an email some time ago) out there
> already that linux has too, why not use those?
>
Lets get this bit of the discussion over with :-), the kernel has uint*_t
in it in a few places, this is now a standard type so we will use it, if
someone is going to work on the DRM they'll see the surrounding uint32_t
so they'll know what it looks like and I'll make sure none of the others
sneak in....
Dave.
--
David Airlie, Software Engineer
http://www.skynet.ie/~airlied / airlied at skynet.ie
pam_smb / Linux DECstation / Linux VAX / ILUG person
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2004-07-31 11:02 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-07-31 9:13 drm - first steps towards 64-bit correctness Dave Airlie
2004-07-31 9:32 ` Eric Anholt
2004-07-31 9:54 ` Arjan van de Ven
2004-07-31 9:57 ` Eric Anholt
2004-07-31 10:06 ` William Lee Irwin III
2004-07-31 11:02 ` Dave Airlie
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.