RE: [PATCH] mmc: renesas_sdhi: increase suspend/resume latency limit

Hi Ulrich-san,

Thank you for your comments!

From: Ulrich Hecht, Sent: Monday, August 2, 2021 8:18 PM

quoted

On 08/02/2021 7:34 AM Yoshihiro Shimoda [off-list ref] wrote:


Hi Wolfram-san, Ulrich-san,

quoted

From: Wolfram Sang, Sent: Saturday, July 31, 2021 12:28 AM

On Fri, May 14, 2021 at 05:53:18PM +0200, Ulrich Hecht wrote:

quoted

The TMIO core sets a very low latency limit (100 us), but when using R-Car
SDHI hosts with SD cards, I have observed typical latencies of around 20-30
ms. This prevents runtime PM from working properly, and the devices remain
on continuously.

This patch sets the default latency limit to 100 ms to avoid that.

Signed-off-by: Ulrich Hecht <redacted>

Adding Shimoda-san to CC.

Shimoda-san: can you kindly run your SDHI tests with this patch applied?

Sure!

However, I have a question about this patch.
Would you know how to measure the latencies?

IIRC I simply put a printk() in default_suspend_ok() that dumps td->suspend_latency_ns and td->resume_latency_ns.

Thanks! I could get the latencies.

quoted

I enabled function trace of rpm and checked the log, but I could not observe
any behavior changes with and without applying this patch.

So you are saying that for you, the clock is suspended as expected when removing the card, even without this patch?
If so, I wonder if there are variations between boards...

I'm sorry, I only observed the trace log.
As Wolfram-san mentioned on other email thread, I should check the clk_summary whether
the clock is enabled or not.

So, I'll test this patch with my test environment tomorrow.

CU
Uli