Hi Ulf,
On Thu, May 11, 2023 at 12:43 PM Ulf Hansson <ulf.hansson@xxxxxxxxxx> wrote:
[...]
> > Indeed, removing mmc_set_clock() from mmc_select_hs200() also makes my
> > eMMC appear again on top of Linux 6.4-rc1.
> > See the attached diff in case it's not fully clear which
> > mmc_set_clock() call I removed.
>
> Thanks for the update! Removing that call restores mmc_select_hs200()
> to the previous behaviour - so thanks for confirming that this is
> working.
>
> However, to find the proper solution, I think we need to understand
> why we are hanging in the meson-mx-sdhc driver first. Here's a couple
> of follow up questions from me:
>
> 1) Before calling mmc_set_clock() what is the actual clock rate that
> has been set by the meson driver?
>
> 2) Does the call to mmc_set_clock() return or hang? Can we verify that
> the clock gets set correctly?
I used the attached diff to answer these two questions. See the
following log extract (full log is attached):
meson-mx-sdhc c1108e00.mmc: Trying to set MMC clock to 400000Hz
meson-mx-sdhc c1108e00.mmc: Actual MMC clock to 399812Hz
mmc1: mmc_select_hs200 switching to clock from card->ext_csd.hs_max_dtr...
meson-mx-sdhc c1108e00.mmc: Trying to set MMC clock to 52000000Hz
meson-mx-sdhc c1108e00.mmc: Actual MMC clock to 51000000Hz
mmc1: mmc_select_hs200 mmc_set_clock returned
> 3) If 2) seems to work above, we need to figure out why
> mmc_switch_status() is hanging. If there is a problem with the eMMC
> card responding in-correctly, the host driver should return with an
> error code, right?
You're right: it's indeed hanging in mmc_switch_status()
I don't get any interrupts (timeout, CRC error, ...) for it.
Do you have any suggestions what to check next?
Best regards,
Martin
diff --git a/drivers/mmc/core/mmc.c b/drivers/mmc/core/mmc.c
index 89cd48fcec79..ef2986ff4528 100644
--- a/drivers/mmc/core/mmc.c
+++ b/drivers/mmc/core/mmc.c
@@ -1500,7 +1500,11 @@ static int mmc_select_hs200(struct mmc_card *card)
old_timing = host->ios.timing;
old_clock = host->ios.clock;
mmc_set_timing(host, MMC_TIMING_MMC_HS200);
+ pr_err("%s: %s switching to clock from card->ext_csd.hs_max_dtr...\n", mmc_hostname(card->host),
+ __func__);
mmc_set_clock(card->host, card->ext_csd.hs_max_dtr);
+ pr_err("%s: %s mmc_set_clock returned\n", mmc_hostname(card->host),
+ __func__);
/*
* For HS200, CRC errors are not a reliable way to know the
@@ -1508,6 +1512,8 @@ static int mmc_select_hs200(struct mmc_card *card)
* tuning will fail and the result ends up the same.
*/
err = mmc_switch_status(card, false);
+ pr_err("%s: %s status after mmc_set_clock: %d\n", mmc_hostname(card->host),
+ __func__, err);
/*
* mmc_select_timing() assumes timing has not changed if
diff --git a/drivers/mmc/host/meson-mx-sdhc-mmc.c b/drivers/mmc/host/meson-mx-sdhc-mmc.c
index 8799f47edf23..104f0027594f 100644
--- a/drivers/mmc/host/meson-mx-sdhc-mmc.c
+++ b/drivers/mmc/host/meson-mx-sdhc-mmc.c
@@ -272,6 +272,8 @@ static int meson_mx_sdhc_set_clk(struct mmc_host *mmc, struct mmc_ios *ios)
meson_mx_sdhc_disable_clks(mmc);
+ dev_err(mmc_dev(mmc), "Trying to set MMC clock to %uHz\n", ios->clock);
+
if (ios->clock) {
ret = clk_set_rate(host->sd_clk, ios->clock);
if (ret) {
@@ -317,6 +319,8 @@ static int meson_mx_sdhc_set_clk(struct mmc_host *mmc, struct mmc_ios *ios)
mmc->actual_clock = 0;
}
+ dev_err(mmc_dev(mmc), "Actual MMC clock to %uHz\n", mmc->actual_clock);
+
return 0;
}
@@ -562,6 +566,9 @@ static irqreturn_t meson_mx_sdhc_irq(int irq, void *data)
if (!(ictl & ista))
return IRQ_NONE;
+ dev_err(mmc_dev(host->mmc), "CMD%d ISTA: 0x%08x, ICTL: 0x%08x\n",
+ cmd->opcode, ista, ictl);
+
if (ista & MESON_SDHC_ISTA_RXFIFO_FULL ||
ista & MESON_SDHC_ISTA_TXFIFO_EMPTY)
cmd->error = -EIO;
