[PATCH] usb: ehci: Prevent missed ehci interrupts with edge-triggered MSI

[PATCH] usb: ehci: Prevent missed ehci interrupts with edge-triggered MSI

[David Jeffery]

When MSI is used by the ehci driver, it can cause interrupts to be lost which
results in ehci only continuing to work due to its polling fallback.  But the
reliance of polling drasticly reduces performance of any I/O through ehci.

Interrupts are lost as ehci's interrupt handler does not safely handle
edge-triggered interrupts.  It fails to ensure all interrupt status bits are
cleared, which works with level-triggered interrupts but not the 
edge-triggered interrupts typical from using MSI.

To fix this problem, check if the driver may have raced with the hardware
setting additional interrupt status bits and clear status until it is in a
stable state.

Signed-off-by: David Jeffery <djeffery@redhat.com>
Tested-by: Laurence Oberman <loberman@redhat.com>
---

This is an alternative to reverting 306c54d0edb6ba94d39877524dddebaad7770cf2
which is the patch that allowed MSI to be used with ehci.

 ehci-hcd.c |   16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c
index 36f5bf6a0752..2283205d4b40 100644
--- a/drivers/usb/host/ehci-hcd.c
+++ b/drivers/usb/host/ehci-hcd.c
@@ -704,14 +704,18 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
 {
 	struct ehci_hcd		*ehci = hcd_to_ehci (hcd);
 	u32			status, masked_status, pcd_status = 0, cmd;
+	u32			current_status;
 	int			bh;
 
 	spin_lock(&ehci->lock);
 
-	status = ehci_readl(ehci, &ehci->regs->status);
+	status = 0;
 
+	current_status = ehci_readl(ehci, &ehci->regs->status);
+restart:
+	status |= current_status;
 	/* e.g. cardbus physical eject */
-	if (status == ~(u32) 0) {
+	if (current_status == ~(u32) 0) {
 		ehci_dbg (ehci, "device removed\n");
 		goto dead;
 	}
@@ -720,7 +724,7 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
 	 * We don't use STS_FLR, but some controllers don't like it to
 	 * remain on, so mask it out along with the other status bits.
 	 */
-	masked_status = status & (INTR_MASK | STS_FLR);
+	masked_status = current_status & (INTR_MASK | STS_FLR);
 
 	/* Shared IRQ? */
 	if (!masked_status || unlikely(ehci->rh_state == EHCI_RH_HALTED)) {
@@ -730,6 +734,12 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
 
 	/* clear (just) interrupts */
 	ehci_writel(ehci, masked_status, &ehci->regs->status);
+
+	/* for edge interrupts, don't race with an interrupt bit being raised */
+	current_status = ehci_readl(ehci, &ehci->regs->status);
+	if (current_status & INTR_MASK)
+		goto restart;
+
 	cmd = ehci_readl(ehci, &ehci->regs->command);
 	bh = 0;
 

Re: [PATCH] usb: ehci: Prevent missed ehci interrupts with edge-triggered MSI

[Andy Shevchenko]

On Wed, Jul 14, 2021 at 10:03:09AM -0400, David Jeffery wrote:

Thanks, I have few minor comments, after addressing them feel free to add
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

> When MSI is used by the ehci driver, it can cause interrupts to be lost which

ehci -> EHCI everywhere?

> results in ehci only continuing to work due to its polling fallback.  But the
> reliance of polling drasticly reduces performance of any I/O through ehci.
> 
> Interrupts are lost as ehci's interrupt handler does not safely handle
> edge-triggered interrupts.  It fails to ensure all interrupt status bits are
> cleared, which works with level-triggered interrupts but not the 

Drop trailing white space here and double spaces in the couple of places above?

> edge-triggered interrupts typical from using MSI.
> 
> To fix this problem, check if the driver may have raced with the hardware
> setting additional interrupt status bits and clear status until it is in a
> stable state.

Fixes tag?

> Signed-off-by: David Jeffery <djeffery@redhat.com>
> Tested-by: Laurence Oberman <loberman@redhat.com>
> ---
> 
> This is an alternative to reverting 306c54d0edb6ba94d39877524dddebaad7770cf2
> which is the patch that allowed MSI to be used with ehci.

Thanks!

> 
>  ehci-hcd.c |   16 +++++++++++++---
>  1 file changed, 13 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c
> index 36f5bf6a0752..2283205d4b40 100644
> --- a/drivers/usb/host/ehci-hcd.c
> +++ b/drivers/usb/host/ehci-hcd.c
> @@ -704,14 +704,18 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
>  {
>  	struct ehci_hcd		*ehci = hcd_to_ehci (hcd);
>  	u32			status, masked_status, pcd_status = 0, cmd;
> +	u32			current_status;

Perhaps

	u32			status, current_status, masked_status, pcd_status = 0;
	u32			cmd;

?

>  	int			bh;
>  
>  	spin_lock(&ehci->lock);
>  
> -	status = ehci_readl(ehci, &ehci->regs->status);
> +	status = 0;
>  
> +	current_status = ehci_readl(ehci, &ehci->regs->status);
> +restart:
> +	status |= current_status;
>  	/* e.g. cardbus physical eject */
> -	if (status == ~(u32) 0) {
> +	if (current_status == ~(u32) 0) {
>  		ehci_dbg (ehci, "device removed\n");
>  		goto dead;
>  	}
> @@ -720,7 +724,7 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
>  	 * We don't use STS_FLR, but some controllers don't like it to
>  	 * remain on, so mask it out along with the other status bits.
>  	 */
> -	masked_status = status & (INTR_MASK | STS_FLR);
> +	masked_status = current_status & (INTR_MASK | STS_FLR);
>  
>  	/* Shared IRQ? */
>  	if (!masked_status || unlikely(ehci->rh_state == EHCI_RH_HALTED)) {
> @@ -730,6 +734,12 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
>  
>  	/* clear (just) interrupts */
>  	ehci_writel(ehci, masked_status, &ehci->regs->status);
> +
> +	/* for edge interrupts, don't race with an interrupt bit being raised */

for -> For

> +	current_status = ehci_readl(ehci, &ehci->regs->status);
> +	if (current_status & INTR_MASK)
> +		goto restart;
> +
>  	cmd = ehci_readl(ehci, &ehci->regs->command);
>  	bh = 0;

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH] usb: ehci: Prevent missed ehci interrupts with edge-triggered MSI

[Alan Stern]

On Wed, Jul 14, 2021 at 10:03:09AM -0400, David Jeffery wrote:
> When MSI is used by the ehci driver, it can cause interrupts to be lost which
> results in ehci only continuing to work due to its polling fallback.  But the
> reliance of polling drasticly reduces performance of any I/O through ehci.
> 
> Interrupts are lost as ehci's interrupt handler does not safely handle
> edge-triggered interrupts.  It fails to ensure all interrupt status bits are
> cleared, which works with level-triggered interrupts but not the 
> edge-triggered interrupts typical from using MSI.
> 
> To fix this problem, check if the driver may have raced with the hardware
> setting additional interrupt status bits and clear status until it is in a
> stable state.
> 
> Signed-off-by: David Jeffery <djeffery@redhat.com>
> Tested-by: Laurence Oberman <loberman@redhat.com>
> ---
> 
> This is an alternative to reverting 306c54d0edb6ba94d39877524dddebaad7770cf2
> which is the patch that allowed MSI to be used with ehci.
> 
>  ehci-hcd.c |   16 +++++++++++++---
>  1 file changed, 13 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c
> index 36f5bf6a0752..2283205d4b40 100644
> --- a/drivers/usb/host/ehci-hcd.c
> +++ b/drivers/usb/host/ehci-hcd.c
> @@ -704,14 +704,18 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
>  {
>  	struct ehci_hcd		*ehci = hcd_to_ehci (hcd);
>  	u32			status, masked_status, pcd_status = 0, cmd;
> +	u32			current_status;
>  	int			bh;
>  
>  	spin_lock(&ehci->lock);
>  
> -	status = ehci_readl(ehci, &ehci->regs->status);
> +	status = 0;
>  
> +	current_status = ehci_readl(ehci, &ehci->regs->status);
> +restart:
> +	status |= current_status;
>  	/* e.g. cardbus physical eject */
> -	if (status == ~(u32) 0) {
> +	if (current_status == ~(u32) 0) {
>  		ehci_dbg (ehci, "device removed\n");
>  		goto dead;
>  	}

Mild stylistic quibble: I generally prefer to have a blank line before a 
/* ... */ comment.  And it doesn't seem reasonable to have a blank line 
between "status = 0" and the current_status assignment, since those are 
similar once-only things before the beginning of the "restart" loop.  
Also, I would move the "status |= current_status" line after the test 
for device removed, since that test doesn't use status.

But obviously none of these things affect the patch's correntness.

> @@ -720,7 +724,7 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
>  	 * We don't use STS_FLR, but some controllers don't like it to
>  	 * remain on, so mask it out along with the other status bits.
>  	 */
> -	masked_status = status & (INTR_MASK | STS_FLR);
> +	masked_status = current_status & (INTR_MASK | STS_FLR);
>  
>  	/* Shared IRQ? */
>  	if (!masked_status || unlikely(ehci->rh_state == EHCI_RH_HALTED)) {
> @@ -730,6 +734,12 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
>  
>  	/* clear (just) interrupts */
>  	ehci_writel(ehci, masked_status, &ehci->regs->status);
> +
> +	/* for edge interrupts, don't race with an interrupt bit being raised */
> +	current_status = ehci_readl(ehci, &ehci->regs->status);
> +	if (current_status & INTR_MASK)
> +		goto restart;
> +
>  	cmd = ehci_readl(ehci, &ehci->regs->command);
>  	bh = 0;

You can choose to submit a new version of the patch with those stylistic 
changes, or if you prefer, just stick with this version.  Either way,

Acked-by: Alan Stern <stern@rowland.harvard.edu>

Does this issue affect any other PCI-based host controller drivers?

Alan Stern

Re: [PATCH] usb: ehci: Prevent missed ehci interrupts with edge-triggered MSI

[David Jeffery]

On Wed, Jul 14, 2021 at 10:27 AM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
>
> On Wed, Jul 14, 2021 at 10:03:09AM -0400, David Jeffery wrote:
>
> Thanks, I have few minor comments, after addressing them feel free to add
> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
>
> > When MSI is used by the ehci driver, it can cause interrupts to be lost which
>
> ehci -> EHCI everywhere?
>

Are you asking for a capitalization change in the text or asking what
all is affected by the bug?

> > diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c
> > index 36f5bf6a0752..2283205d4b40 100644
> > --- a/drivers/usb/host/ehci-hcd.c
> > +++ b/drivers/usb/host/ehci-hcd.c
> > @@ -704,14 +704,18 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
> >  {
> >       struct ehci_hcd         *ehci = hcd_to_ehci (hcd);
> >       u32                     status, masked_status, pcd_status = 0, cmd;
> > +     u32                     current_status;
>
> Perhaps
>
>         u32                     status, current_status, masked_status, pcd_status = 0;
>         u32                     cmd;
>
> ?
>

Is this a style preference?  I can change it and just did it in a way
to minimize line changes.

>
> >       int                     bh;
> >
> >       spin_lock(&ehci->lock);
> >
> > -     status = ehci_readl(ehci, &ehci->regs->status);
> > +     status = 0;
> >
> > +     current_status = ehci_readl(ehci, &ehci->regs->status);
> > +restart:
> > +     status |= current_status;
> >       /* e.g. cardbus physical eject */
> > -     if (status == ~(u32) 0) {
> > +     if (current_status == ~(u32) 0) {
> >               ehci_dbg (ehci, "device removed\n");
> >               goto dead;
> >       }
> > @@ -720,7 +724,7 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
> >        * We don't use STS_FLR, but some controllers don't like it to
> >        * remain on, so mask it out along with the other status bits.
> >        */
> > -     masked_status = status & (INTR_MASK | STS_FLR);
> > +     masked_status = current_status & (INTR_MASK | STS_FLR);
> >
> >       /* Shared IRQ? */
> >       if (!masked_status || unlikely(ehci->rh_state == EHCI_RH_HALTED)) {
> > @@ -730,6 +734,12 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
> >
> >       /* clear (just) interrupts */
> >       ehci_writel(ehci, masked_status, &ehci->regs->status);
> > +
> > +     /* for edge interrupts, don't race with an interrupt bit being raised */
>
> for -> For
>

I'll update it.

Re: [PATCH] usb: ehci: Prevent missed ehci interrupts with edge-triggered MSI

[David Jeffery]

On Wed, Jul 14, 2021 at 10:29 AM Alan Stern <stern@rowland.harvard.edu> wrote:
>
> On Wed, Jul 14, 2021 at 10:03:09AM -0400, David Jeffery wrote:

> >
> > diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c
> > index 36f5bf6a0752..2283205d4b40 100644
> > --- a/drivers/usb/host/ehci-hcd.c
> > +++ b/drivers/usb/host/ehci-hcd.c
> > @@ -704,14 +704,18 @@ static irqreturn_t ehci_irq (struct usb_hcd *hcd)
> >  {
> >       struct ehci_hcd         *ehci = hcd_to_ehci (hcd);
> >       u32                     status, masked_status, pcd_status = 0, cmd;
> > +     u32                     current_status;
> >       int                     bh;
> >
> >       spin_lock(&ehci->lock);
> >
> > -     status = ehci_readl(ehci, &ehci->regs->status);
> > +     status = 0;
> >
> > +     current_status = ehci_readl(ehci, &ehci->regs->status);
> > +restart:
> > +     status |= current_status;
> >       /* e.g. cardbus physical eject */
> > -     if (status == ~(u32) 0) {
> > +     if (current_status == ~(u32) 0) {
> >               ehci_dbg (ehci, "device removed\n");
> >               goto dead;
> >       }
>
> Mild stylistic quibble: I generally prefer to have a blank line before a
> /* ... */ comment.  And it doesn't seem reasonable to have a blank line
> between "status = 0" and the current_status assignment, since those are
> similar once-only things before the beginning of the "restart" loop.
> Also, I would move the "status |= current_status" line after the test
> for device removed, since that test doesn't use status.
>
> But obviously none of these things affect the patch's correntness.

I'll update it with requested changes and resend it.

>
> You can choose to submit a new version of the patch with those stylistic
> changes, or if you prefer, just stick with this version.  Either way,
>
> Acked-by: Alan Stern <stern@rowland.harvard.edu>
>
> Does this issue affect any other PCI-based host controller drivers?
>
> Alan Stern
>

Possibly.  The uhci driver should have the same issue if hardware
exists with it and has MSI for it.  I suspect the ohci driver has a
similar issue, but I'm not familiar enough with its interfaces and
specification to say for sure.

David Jeffery

Re: [PATCH] usb: ehci: Prevent missed ehci interrupts with edge-triggered MSI

[Alan Stern]

On Wed, Jul 14, 2021 at 12:10:26PM -0400, David Jeffery wrote:
> On Wed, Jul 14, 2021 at 10:29 AM Alan Stern <stern@rowland.harvard.edu> wrote:
> >
> > Does this issue affect any other PCI-based host controller drivers?
> >
> > Alan Stern
> >
> 
> Possibly.  The uhci driver should have the same issue if hardware
> exists with it and has MSI for it.  I suspect the ohci driver has a
> similar issue, but I'm not familiar enough with its interfaces and
> specification to say for sure.

Hmmm.  I've never heard of an MSI implementation of either of those, 
although it may be possible for OHCI.  UHCI was used only by Intel and 
one or two others; as far as I know none of them are producing those 
controllers any more.  Nowadays when people want to support USB-1.1 
devices but not USB-3, they use an EHCI controller together with a 
"rate-matching" hub, which avoids the need for any UHCI or OHCI 
controller.

I was wondering more about dwc2 or other non-EHCI USB-2 controller 
drivers.

Alan Stern

Re: [PATCH] usb: ehci: Prevent missed ehci interrupts with edge-triggered MSI

[Andy Shevchenko]

On Wed, Jul 14, 2021 at 6:55 PM David Jeffery <djeffery@redhat.com> wrote:
> On Wed, Jul 14, 2021 at 10:27 AM Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> wrote:
> > On Wed, Jul 14, 2021 at 10:03:09AM -0400, David Jeffery wrote:
> >
> > Thanks, I have few minor comments, after addressing them feel free to add
> > Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> >
> > > When MSI is used by the ehci driver, it can cause interrupts to be lost which
> >
> > ehci -> EHCI everywhere?
> >
>
> Are you asking for a capitalization change in the text or asking what
> all is affected by the bug?

I'm suggesting to capitalize the abbreviation (maybe not in all cases,
but where you refer to the hardware, for the driver it's called
ehci-hcd).

...

> > >       u32                     status, masked_status, pcd_status = 0, cmd;
> > > +     u32                     current_status;
> >
> > Perhaps
> >
> >         u32                     status, current_status, masked_status, pcd_status = 0;
> >         u32                     cmd;
> >
> > ?
>
> Is this a style preference?  I can change it and just did it in a way
> to minimize line changes.

Yes, I believe it's a preference of having something unified
semantically on one line.

-- 
With Best Regards,
Andy Shevchenko