It might be that something goes wrong during tuning so the MMC core will immediately trigger a retune. In our case it was: - we sent a tuning block - there was an error so we need to send an abort cmd to the eMMC - the abort cmd had a CRC error - retune was set by the MMC core This lead to a vicious circle causing a performance regression of 75%. So, disable retuning while we tune. Let the tuning complete and see then if it worked out or not. Reported-by Yoshihiro Shimoda <yoshihiro.shimoda.uh@xxxxxxxxxxx> Signed-off-by: Wolfram Sang <wsa+renesas@xxxxxxxxxxxxxxxxxxxx> --- Hi Ulf, this patch is marked as RFC because I think this is a generic issue. Lots of things could happen in the driver callback which cause a retune, so I'd think it makes sense to deactivate it globally here. If you think this is a driver specific issue, just let me know. I can provide a small patch to create the issue for SDHI hardware, created by Shimoda-san. We couldn't think of an easy way to reproduce it with the fault injector, sadly. Let me know if you want to see that patch. drivers/mmc/core/core.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c index b039dcff17f8..54f0814f110c 100644 --- a/drivers/mmc/core/core.c +++ b/drivers/mmc/core/core.c @@ -927,6 +927,8 @@ int mmc_execute_tuning(struct mmc_card *card) if (!host->ops->execute_tuning) return 0; + mmc_retune_disable(host); + if (host->cqe_on) host->cqe_ops->cqe_off(host); -- 2.30.2
