The function dma_controller_irq() locks up the exact same spinlock we
locked before calling it, which obviously resulted in a deadlock when
CONFIG_SMP was enabled. This flew under the radar as none of the boards
supported by this driver needs SMP.
Fixes: 57aadb46bd63 ("usb: musb: jz4740: Add support for DMA")
Signed-off-by: Paul Cercueil <[email protected]>
---
drivers/usb/musb/jz4740.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/usb/musb/jz4740.c b/drivers/usb/musb/jz4740.c
index 54e7b30acc69..42c1e8bfc4be 100644
--- a/drivers/usb/musb/jz4740.c
+++ b/drivers/usb/musb/jz4740.c
@@ -69,11 +69,11 @@ static irqreturn_t jz4740_musb_interrupt(int irq, void *__hci)
irqreturn_t retval = IRQ_NONE, retval_dma = IRQ_NONE;
struct musb *musb = __hci;
- spin_lock_irqsave(&musb->lock, flags);
-
if (IS_ENABLED(CONFIG_USB_INVENTRA_DMA) && musb->dma_controller)
retval_dma = dma_controller_irq(irq, musb->dma_controller);
+ spin_lock_irqsave(&musb->lock, flags);
+
musb->int_usb = musb_readb(musb->mregs, MUSB_INTRUSB);
musb->int_tx = musb_readw(musb->mregs, MUSB_INTRTX);
musb->int_rx = musb_readw(musb->mregs, MUSB_INTRRX);
--
2.26.2
On Wed, May 20, 2020 at 05:01:11PM +0200, Paul Cercueil wrote:
> The function dma_controller_irq() locks up the exact same spinlock we
> locked before calling it, which obviously resulted in a deadlock when
> CONFIG_SMP was enabled. This flew under the radar as none of the boards
> supported by this driver needs SMP.
>
> Fixes: 57aadb46bd63 ("usb: musb: jz4740: Add support for DMA")
> Signed-off-by: Paul Cercueil <[email protected]>
Queued for v5.8. Thanks.
-Bin.
Hi Bin,
The patch it fixes was introduced in 5.7-rc1, is it possible to queue
it for the next -rc? Otherwise I'll need to Cc it to linux-stable.
-Paul
Le jeu. 21 mai 2020 ? 15:34, Bin Liu <[email protected]> a ?crit :
> On Wed, May 20, 2020 at 05:01:11PM +0200, Paul Cercueil wrote:
>> The function dma_controller_irq() locks up the exact same spinlock
>> we
>> locked before calling it, which obviously resulted in a deadlock
>> when
>> CONFIG_SMP was enabled. This flew under the radar as none of the
>> boards
>> supported by this driver needs SMP.
>>
>> Fixes: 57aadb46bd63 ("usb: musb: jz4740: Add support for DMA")
>> Signed-off-by: Paul Cercueil <[email protected]>
>
> Queued for v5.8. Thanks.
> -Bin.
Hi Greg,
On Wed, May 20, 2020 at 05:01:11PM +0200, Paul Cercueil wrote:
> The function dma_controller_irq() locks up the exact same spinlock we
> locked before calling it, which obviously resulted in a deadlock when
> CONFIG_SMP was enabled. This flew under the radar as none of the boards
> supported by this driver needs SMP.
>
> Fixes: 57aadb46bd63 ("usb: musb: jz4740: Add support for DMA")
> Signed-off-by: Paul Cercueil <[email protected]>
Is it too late if you directly take this patch for v5.7-rc7? The bug is
introduced in v5.7-rc1. If so, here is my Acked-by:
Acked-by: Bin Liu <[email protected]>
Thanks,
-Bin.
> ---
> drivers/usb/musb/jz4740.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/usb/musb/jz4740.c b/drivers/usb/musb/jz4740.c
> index 54e7b30acc69..42c1e8bfc4be 100644
> --- a/drivers/usb/musb/jz4740.c
> +++ b/drivers/usb/musb/jz4740.c
> @@ -69,11 +69,11 @@ static irqreturn_t jz4740_musb_interrupt(int irq, void *__hci)
> irqreturn_t retval = IRQ_NONE, retval_dma = IRQ_NONE;
> struct musb *musb = __hci;
>
> - spin_lock_irqsave(&musb->lock, flags);
> -
> if (IS_ENABLED(CONFIG_USB_INVENTRA_DMA) && musb->dma_controller)
> retval_dma = dma_controller_irq(irq, musb->dma_controller);
>
> + spin_lock_irqsave(&musb->lock, flags);
> +
> musb->int_usb = musb_readb(musb->mregs, MUSB_INTRUSB);
> musb->int_tx = musb_readw(musb->mregs, MUSB_INTRTX);
> musb->int_rx = musb_readw(musb->mregs, MUSB_INTRRX);
> --
> 2.26.2
>
On Thu, May 21, 2020 at 10:39:25PM +0200, Paul Cercueil wrote:
> Hi Bin,
>
> The patch it fixes was introduced in 5.7-rc1, is it possible to queue it for
> the next -rc? Otherwise I'll need to Cc it to linux-stable.
I pinged Grag if he can directly take it for -rc7.
-Bin.
Paul,
On Thu, May 21, 2020 at 04:30:28PM -0500, Bin Liu wrote:
> On Thu, May 21, 2020 at 10:39:25PM +0200, Paul Cercueil wrote:
> > Hi Bin,
> >
> > The patch it fixes was introduced in 5.7-rc1, is it possible to queue it for
> > the next -rc? Otherwise I'll need to Cc it to linux-stable.
>
> I pinged Grag if he can directly take it for -rc7.
BTY, it is not an issue if it is too late for -rc7, I've already added
Cc stable@ when queueing it.
-Bin.