xf86: VGA arbiter lock on CloseScreen() still needed ?

[Enrico Weigelt, metux IT consult]

On 29.04.25 15:27, Michel Dänzer wrote:

>> In most cases, that's really simple and straightforward, [...]
> 
> That contradicts my experience working on the X server code for over two decades. 

Sorry, most cases, I've converted so far. (just speaking about what's in
that queue)

Of course there also are tricky cases, I'll have a closer look at those
when it's their time :p

My goal right how is decoupling extensions (which are quite orthogonal
to screen drivers/DIX). Layered drivers (w/ fb, mi, ...) aren't my focus
yet (even though did some cleanup there, too.

> There are many cases where the callbacks must be executed in a specific order 
> for correct operation. 

Yes, sometimes that's the case. That's what I'm currently decoupling.
In most of the cases I've touched so far, it's because the daisy-chaining.

> And the driver's callback can be anywhere in that order.

*drivers*, yes. My current focus are extensions.
But I've already found (and cleaned) several cases of drivers where we
don't need that at all, because the call chain actually is determined at
compile time (eg. when using MI or FB functions).

>> When fully done, it's also now possible hook/unhook at runtime (eg.
>> enable/disable extensions on the fly, etc)
> 
> That's already possible, and done.

Change hooking at runtime ?
How do you do this, when the daisy-chain scattered through lots of
devPrivate's ?



--mtx

-- 
---
All racism is bad. All lives matter.
---
Enrico Weigelt, metux IT consult
Free software and Linux embedded engineering
info at metux.net -- +49-151-27565287