mmc: jz4740: Fix race condition in IRQ mask update

commit a04f001 upstream.

A spinlock is held while updating the internal copy of the IRQ mask,
but not while writing it to the actual IMASK register. After the lock
is released, an IRQ can occur before the IMASK register is written.
If handling this IRQ causes the mask to be changed, when the handler
returns back to the middle of the first mask update, a stale value
will be written to the mask register.

If this causes an IRQ to become unmasked that cannot have its status
cleared by writing a 1 to it in the IREG register, e.g. the SDIO IRQ,
then we can end up stuck with the same IRQ repeatedly being fired but
not handled. Normally the MMC IRQ handler attempts to clear any
unexpected IRQs by writing IREG, but for those that cannot be cleared
in this way then the IRQ will just repeatedly fire.

This was resulting in lockups after a while of using Wi-Fi on the
CI20 (GitHub issue #19).

Resolve by holding the spinlock until after the IMASK register has
been updated.

Cc: [email protected]
Link: MIPS/CI20_linux#19
Fixes: 61bfbdb ("MMC: Add support for the controller on JZ4740 SoCs.")
Tested-by: Mathieu Malaterre <[email protected]>
Signed-off-by: Alex Smith <[email protected]>
Signed-off-by: Ulf Hansson <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

Author: Alex Smith

drivers/mmc/host/jz4740_mmc.c

@@ -368,9 +368,9 @@ static void jz4740_mmc_set_irq_enabled(struct jz4740_mmc_host *host,
 		host->irq_mask &= ~irq;
 	else
 		host->irq_mask |= irq;
-	spin_unlock_irqrestore(&host->lock, flags);
 
 	writew(host->irq_mask, host->base + JZ_REG_MMC_IMASK);
+	spin_unlock_irqrestore(&host->lock, flags);
 }
 
 static void jz4740_mmc_clock_enable(struct jz4740_mmc_host *host,