On Mon, May 27, 2019 at 03:08:07PM +0300, Adrian Hunter wrote:
> On 27/05/19 12:37 PM, Brian Masney wrote:
> > On Sun, May 26, 2019 at 03:58:19PM -0400, Brian Masney wrote:
> >> I attached a patch that shows how I was able to determine what had
> >> already claimed the host.
> > On Mon, May 27, 2019 at 10:48:24AM +0300, Adrian Hunter wrote:
> >> This is because SDHCI is using the IRQ thread to process the SDIO card
> >> interrupt (sdio_run_irqs()). When the card driver tries to use the card, it
> >> causes interrupts which deadlocks since c07a48c26519 ("mmc: sdhci: Remove
> >> finish_tasklet") has moved the tasklet processing to the IRQ thread.
> >>
> >> I would expect to be able to use the IRQ thread to complete requests, and it
> >> is desirable to do so because it is lower latency.
> >>
> >> Probably, SDHCI should use sdio_signal_irq() which queues a work item, and
> >> is what other drivers are doing.
> >>
> >> I will investigate some more and send a patch.
>
> Please try the patch below:
>
> From: Adrian Hunter <adrian.hunter@xxxxxxxxx>
> Date: Mon, 27 May 2019 14:45:55 +0300
> Subject: [PATCH] mmc: sdhci: Fix SDIO IRQ thread deadlock
>
> Since commit c07a48c26519 ("mmc: sdhci: Remove finish_tasklet"), the IRQ
> thread might be used to complete requests, but the IRQ thread is also used
> to process SDIO card interrupts. This can cause a deadlock when the SDIO
> processing tries to access the card since that would also require the IRQ
> thread. Change SDHCI to use sdio_signal_irq() to schedule a work item
> instead. That also requires implementing the ->ack_sdio_irq() mmc host op.
>
> Signed-off-by: Adrian Hunter <adrian.hunter@xxxxxxxxx>
> Fixes: c07a48c26519 ("mmc: sdhci: Remove finish_tasklet")
Yes, this fixes the issue for me. You can add my:
Reported-by: Brian Masney <masneyb@xxxxxxxxxxxxx>
Tested-by: Brian Masney <masneyb@xxxxxxxxxxxxx>
Thanks,
Brian
