Skip to contentSkip to conversation

tmio_mmc_pio: prevent endless loop in tmio_mmc_set_clock()

PATCH Superseded Sergei Shtylyov 2 participants 3 messages Sep 5, 2014 – Sep 14, 2014

Subsystems (2)

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 tmio_mmc_pio: prevent endless loop in tmio_mmc_set_clock()

I spent a couple of days with the driver just hanging due to me forgetting to specify the external crystal frequency, so that clk_get_rate() returned 0 and thus the loop in tmio_mmc_set_clock() never ended. I don't think that's an acceptable behavior, so I suggest that the minimum frequency is checked for 0 in tmio_mmc_host_probe().

---
The patch is against Ulf Hansson's 'mmc.git' repo's 'fixes' branch.

Changes in version 2:
- fixed grammar in the changelog.

 drivers/mmc/host/tmio_mmc_pio.c |    9 +++++++++
 1 file changed, 9 insertions(+)

Index: mmc/drivers/mmc/host/tmio_mmc_pio.c =================================--- mmc.orig/drivers/mmc/host/tmio_mmc_pio.c +++ mmc/drivers/mmc/host/tmio_mmc_pio.c
@@ -1044,6 +1044,15 @@ int tmio_mmc_host_probe(struct tmio_mmc_
}

/*
+ * Check the sanity of mmc->f_min to prevent tmio_mmc_set_clock() from
+ * looping forever...
+ */
+ if (mmc->f_min = 0) {
+ ret = -EINVAL;
+ goto pm_disable;
+ }
+
+ /*
* There are 4 different scenarios for the card detection:
* 1) an external gpio irq handles the cd (best for power savings)
* 2) internal sdhi irq handles the cd

#1 lore raw
Quoted 10 lines · Sergei Shtylyov

I spent a couple of days with the driver just hanging due to me forgetting to specify the external crystal frequency, so that clk_get_rate() returned 0 and thus the loop in tmio_mmc_set_clock() never ended. I don't think that's an acceptable behavior, so I suggest that the minimum frequency is checked for 0 in tmio_mmc_host_probe().

---
The patch is against Ulf Hansson's 'mmc.git' repo's 'fixes' branch.

Hi Sergei,

Could you please base it towards "next" instead, it doesn't apply. If you think this is important to go stable, I suggest you to add a stable tag for it.

Kind regards
Uffe

Quoted 28 lines

Changes in version 2:
- fixed grammar in the changelog.

 drivers/mmc/host/tmio_mmc_pio.c |    9 +++++++++
 1 file changed, 9 insertions(+)

Index: mmc/drivers/mmc/host/tmio_mmc_pio.c
===================================================================

drivers/mmc/host/tmio_mmc_pio.c +9 -0
@@ -1044,6 +1044,15 @@ int tmio_mmc_host_probe(struct tmio_mmc_
        }
 
        /*
+        * Check the sanity of mmc->f_min to prevent tmio_mmc_set_clock() from
+        * looping forever...
+        */
+       if (mmc->f_min == 0) {
+               ret = -EINVAL;
+               goto pm_disable;
+       }
+
+       /*
         * There are 4 different scenarios for the card detection:
         *  1) an external gpio irq handles the cd (best for power savings)
         *  2) internal sdhi irq handles the cd
#2 lore raw

Hello.

Sorry for delay: I'm on vacations and WiFi didn't work well here...

Quoted 8 lines · Ulf Hansson
Quoted 7 lines

I spent a couple of days with the driver just hanging due to me forgetting to specify the external crystal frequency, so that clk_get_rate() returned 0 and thus the loop in tmio_mmc_set_clock() never ended. I don't think that's an acceptable behavior, so I suggest that the minimum frequency is checked for 0 in tmio_mmc_host_probe().

Quoted 8 lines
Quoted 2 lines

---
The patch is against Ulf Hansson's 'mmc.git' repo's 'fixes' branch.

Hi Sergei,

Could you please base it towards "next" instead, it doesn't apply. If you think this is important to go stable, I suggest you to add a stable tag for it.

It's a bit of a corner case, of course, but the consequences are just too
disastrous: AFAIR, an ARM kernel just silently hangs on boot, with no hints of where. So I'm going to CC: -stable.

Quoted 2 lines

Kind regards
Uffe

WBR, Sergei