On 2018/4/20 17:27, Wolfram Sang wrote:
Hi,
The intention for this revert is that the it's too engineering
solution for a special board but force all platforms to wait
for that long time, especially painful for mmc_power_up for eMMC
when booting.
Careful! It was *introduced* for one specific board. But we can't tell
if other platforms benefit from this since 2009(!).
I agree.
In practise, the modern hardware should have a stable power supply
While I agree, sill notice the "should". Reality is often enough
different.
right, I think "most likely" should be appropriate.
It has been long time proved 2ms is enough when testing variety of
Where is that proof? If we have such a proof, things would be easy, but
I really have strong doubts. We have indications, at best, I'd say.
I just checked the instances(boards) as much as I could got from
different vendors, but I didn't say it stands for every one. :)
This is true if there is an external regulator, but sometimes we use
host specific ways to power the card.
This is also true and to be considered.
Just revert it now and see if it could smokes out something interesting.
Hell, no! "Let's change it to my needs and if it breaks for other
people, they can complain"? That's not how Linux development works, we
should try as hard as possible to not introduce regressions.
Thanks for reminding :)
An option is to use an opt-out method, perhaps via a new host cap?
Opt-in for the shortened delay, I'd suggest. Platforms which know they
need less delay can set it. White listing.
NAK for the original patch.
Actually I never expect to apply a RFC, but always post them for
disscussion for how we could to improve it :). Opt-in for the shoteded
delay looks fine to me. If we are both agree with that, then I could
post a regular patch for it.
Regards,
Wolfram
--
To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
