Thread (12 messages) 12 messages, 2 authors, 2018-07-17

Re: [PATCH RFC 1/2] mmc: sdhci: Allow platform controlled voltage switching

From: Adrian Hunter <adrian.hunter@intel.com>
Date: 2018-07-17 08:43:50
Also in: linux-arm-msm, linux-mmc, lkml

On 17/07/18 11:40, Vijay Viswanath wrote:

On 7/17/2018 1:00 PM, Adrian Hunter wrote:
quoted
On 17/07/18 08:14, Vijay Viswanath wrote:
quoted

On 7/10/2018 4:37 PM, Adrian Hunter wrote:
quoted
On 21/06/18 15:23, Vijay Viswanath wrote:
quoted
Some controllers can have internal mechanism to inform the SW that it
is ready for voltage switching. For such controllers, changing voltage
before the HW is ready can result in various issues.

Add a quirk, which can be used by drivers of such controllers.

Signed-off-by: Vijay Viswanath <redacted>
---
   drivers/mmc/host/sdhci.c | 20 +++++++++++++++-----
   drivers/mmc/host/sdhci.h |  2 ++
   2 files changed, 17 insertions(+), 5 deletions(-)
diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
index 1c828e0..f0346d4 100644
--- a/drivers/mmc/host/sdhci.c
+++ b/drivers/mmc/host/sdhci.c
@@ -1615,7 +1615,8 @@ void sdhci_set_power_noreg(struct sdhci_host *host,
unsigned char mode,
   void sdhci_set_power(struct sdhci_host *host, unsigned char mode,
                unsigned short vdd)
   {
-    if (IS_ERR(host->mmc->supply.vmmc))
+    if (IS_ERR(host->mmc->supply.vmmc) ||
+            (host->quirks2 & SDHCI_QUIRK2_INTERNAL_PWR_CTL))
I think you should provide your own ->set_power() instead of this
will do
quoted
quoted
           sdhci_set_power_noreg(host, mode, vdd);
       else
           sdhci_set_power_reg(host, mode, vdd);
@@ -2009,7 +2010,9 @@ int sdhci_start_signal_voltage_switch(struct
mmc_host *mmc,
           ctrl &= ~SDHCI_CTRL_VDD_180;
           sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
   -        if (!IS_ERR(mmc->supply.vqmmc)) {
+        if (!IS_ERR(mmc->supply.vqmmc) &&
+                !(host->quirks2 &
+                    SDHCI_QUIRK2_INTERNAL_PWR_CTL)) {
And your own ->start_signal_voltage_switch()
sdhci_msm_start_signal_voltage_switch() would be an exact copy of
sdhci_start_signal_voltage_switch()..... will incorporate this if not using
quirk.
quoted
quoted
               ret = mmc_regulator_set_vqmmc(mmc, ios);
               if (ret) {
                   pr_warn("%s: Switching to 3.3V signalling voltage
failed\n",
@@ -2032,7 +2035,8 @@ int sdhci_start_signal_voltage_switch(struct
mmc_host *mmc,
       case MMC_SIGNAL_VOLTAGE_180:
           if (!(host->flags & SDHCI_SIGNALING_180))
               return -EINVAL;
-        if (!IS_ERR(mmc->supply.vqmmc)) {
+        if (!IS_ERR(mmc->supply.vqmmc) &&
+            !(host->quirks2 & SDHCI_QUIRK2_INTERNAL_PWR_CTL)) {
               ret = mmc_regulator_set_vqmmc(mmc, ios);
               if (ret) {
                   pr_warn("%s: Switching to 1.8V signalling voltage
failed\n",
@@ -3485,7 +3489,10 @@ int sdhci_setup_host(struct sdhci_host *host)
        * the host can take the appropriate action if regulators are not
        * available.
        */
-    ret = mmc_regulator_get_supply(mmc);
+    if (!(host->quirks2 & SDHCI_QUIRK2_INTERNAL_PWR_CTL))
Since we expect mmc_regulator_get_supply() to have been called, this could
be:

     if (!mmc->supply.vmmc) {
         ret = mmc_regulator_get_supply(mmc);
         enable_vqmmc = true;
     } else {
         ret = 0;
     }
quoted
quoted
+        ret = mmc_regulator_get_supply(mmc);
+    else
+        ret = 0;
       if (ret)
           return ret;
   @@ -3736,7 +3743,10 @@ int sdhci_setup_host(struct sdhci_host *host)
         /* If vqmmc regulator and no 1.8V signalling, then there's no
UHS */
       if (!IS_ERR(mmc->supply.vqmmc)) {
-        ret = regulator_enable(mmc->supply.vqmmc);
+        if (!(host->quirks2 & SDHCI_QUIRK2_INTERNAL_PWR_CTL))
And this could be:

         if (enable_vqmmc)
             ret = regulator_enable(mmc->supply.vqmmc);
         else
             ret = 0;
  > However, you still need to ensure
regulator_disable(mmc->supply.vqmmc) is
only called if regulator_enable() was called.
I missed this. Will cover it.

Also I missed one more place where we are doing regulator_disable. During
sdhci-msm unbinding, we would end up doing an extra regulator disable
(thanks Evan for pointing it out) in sdhci_remove_host.

To avoid the quirk( or having any flag), it would require copying the code
of sdhci_start_signal_voltage_switch() and sdhci_remove_host() and creating
You do not need to duplicate sdhci_remove_host(), just change it so that it
only  disables what was enabled i.e.

    if (host->vqmmc_enabled)
        regulator_disable(mmc->supply.vqmmc);
Ok, so we will be adding a new flag "vqmmc_enabled" in sdhci_host, ryt ?
Yes
Just wanted to clarify
quoted
quoted
2 new functions in sdhci_msm layer which would do the exact same as above,
with just the regulator parts removed.

This looks messy (considering any future changes to the 2 sdhci API will
need to be copied to their duplicate sdhci_msm API) and a bit overkill to
avoid quirk. At the same time, I don't know how useful such a quirk would be
to other platform drivers.

Please let me know your view/suggestions.
Let's try without the quirk.
quoted
quoted
quoted
+            ret = regulator_enable(mmc->supply.vqmmc);
+        else
+            ret = 0;
           if (!regulator_is_supported_voltage(mmc->supply.vqmmc, 1700000,
                               1950000))
               host->caps1 &= ~(SDHCI_SUPPORT_SDR104 |
diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h
index 23966f8..3b0c97a 100644
--- a/drivers/mmc/host/sdhci.h
+++ b/drivers/mmc/host/sdhci.h
@@ -450,6 +450,8 @@ struct sdhci_host {
    * obtainable timeout.
    */
   #define SDHCI_QUIRK2_DISABLE_HW_TIMEOUT            (1<<17)
+/* Regulator voltage changes are being done from platform layer */
+#define SDHCI_QUIRK2_INTERNAL_PWR_CTL                (1<<18)
So maybe the quirk is not needed.
quoted
         int irq;        /* Device IRQ */
       void __iomem *ioaddr;    /* Mapped address */
Thanks for the review & suggestions!
Vijay
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help