RE: RFC on writel and writel_relaxed

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



> Fair enough. I'd rather people used _relaxed by default, but I have to admit
> that it will probably just result in them getting things wrong...

Certainly requiring the driver writes use explicit barriers should make
them understand when and why they are needed - and then put in the correct ones.

The problem I've had is that I have a good idea which barriers are needed
but find that readl/writel seem to contain a lot of extra ones.
Maybe the are required in some places, but the extra synchronising
instructions could easily have measureable performance effects on
hot paths.

Drivers are likely to contain sequences like:
	read_io
	if (...) return
	write_mem
	...
	write_mem
	barrier
	write_mem
	barrier
	write_io
for things like ring updates.
Where the 'mem' might actually be in io space.
In such sequences not all the synchronising instructions are needed.
I'm not at all sure it is easy to get the right set.

	David




--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html



[Index of Archives]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Photo]     [Yosemite News]     [Yosemite Photos]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux