Re: [PATCH 2/2] mtd: nand: use hrtimer to measure timeout in nand_wait{_ready, }

[Artem Bityutskiy <dedekind1@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 1075 bytes --]

On Tue, 2012-05-22 at 19:10 +0200, Ivan Djelic wrote:
> Current NAND devices require approximately 1-3 ms per block erase, and 100-200 us per page program operation.
> But I think this timeout is only useful to detect and recover from broken hardware: IMHO there is no point in
> trying to "optimize" the timeout delay to 20ms or 400ms depending on which operation is being done (program or erase).
> 
> Why not simplify and just use a single 1s (1000ms) timeout ?


I have many things on my plate now so my answers are are sloppy - I
spent 1 minute or less looking at the code. But I thought that this
timeout is important for the the case when 'chip->device_ready' is NULL.
This is why I wanted to get rid of that case, and then the timeout would
only become about detecting the forever loops. And this is exactly what
I meant by saying that I have further idea on improving that without
hrtimer - just have one single 1 second timeout. We indeed do not have
to be precise for the errors detection.

Oops, SIGBABY.

-- 
Best Regards,
Artem Bityutskiy

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]