[Qemu-trivial] Bug #757654: UHCI fails to signal stall response patch

[Qemu-trivial] Bug #757654: UHCI fails to signal stall response patch

[Jan Vesely]

Hi,

I'm sending a patch for bug #757654. The bug does not really break
anything it just makes USB error detection harder.
It's a quick fix and might need some polishing but it works (I am
currently using it).

thx,
jan

PS: I guess you need this line:
Signed-off-by: Jan Vesely <jano.vesely@gmail.com>

diff --git a/hw/usb-uhci.c b/hw/usb-uhci.c
index 346db3e..a51d89b 100644
--- a/hw/usb-uhci.c
+++ b/hw/usb-uhci.c
@@ -732,11 +732,21 @@ out:
     case USB_RET_STALL:
         td->ctrl |= TD_CTRL_STALL;
         td->ctrl &= ~TD_CTRL_ACTIVE;
+        s->status |= UHCI_STS_USBERR;
+        *int_mask |= 0x02;
+        if (td->ctrl & TD_CTRL_IOC)
+            *int_mask |= 0x01;
+        uhci_update_irq(s);
         return 1;

     case USB_RET_BABBLE:
         td->ctrl |= TD_CTRL_BABBLE | TD_CTRL_STALL;
         td->ctrl &= ~TD_CTRL_ACTIVE;
+        s->status |= UHCI_STS_USBERR;
+        *int_mask |= 0x02;
+        if (td->ctrl & TD_CTRL_IOC)
+           *int_mask |= 0x01;
+        uhci_update_irq(s);
         /* frame interrupted */
         return -1;

Re: [Qemu-trivial] [Qemu-devel] Bug #757654: UHCI fails to signal stall response patch

[Gerd Hoffmann]

> diff --git a/hw/usb-uhci.c b/hw/usb-uhci.c
> index 346db3e..a51d89b 100644
> --- a/hw/usb-uhci.c
> +++ b/hw/usb-uhci.c
> @@ -732,11 +732,21 @@ out:
>       case USB_RET_STALL:
>           td->ctrl |= TD_CTRL_STALL;
>           td->ctrl&= ~TD_CTRL_ACTIVE;
> +        s->status |= UHCI_STS_USBERR;

Just this line should be enougth.

>       case USB_RET_BABBLE:
>           td->ctrl |= TD_CTRL_BABBLE | TD_CTRL_STALL;
>           td->ctrl&= ~TD_CTRL_ACTIVE;
> +        s->status |= UHCI_STS_USBERR;

Likewise.

Tried that?

cheers,
   Gerd

Re: [Qemu-trivial] [Qemu-devel] Bug #757654: UHCI fails to signal stall response patch

[Jan Vesely]

hi,
using the modified (single line) patch only works half-way, it sets
the value in status register (guess that's what that line does :))
but hw interrupt is not generated.

I tried adding uhci_update_irq and this patch:

diff --git a/hw/usb-uhci.c b/hw/usb-uhci.c
index a65e0b3..1e9c1e7 100644
--- a/hw/usb-uhci.c
+++ b/hw/usb-uhci.c
@@ -702,11 +702,15 @@ out:
     case USB_RET_STALL:
         td->ctrl |= TD_CTRL_STALL;
         td->ctrl &= ~TD_CTRL_ACTIVE;
+        s->status |= UHCI_STS_USBERR;
+        uhci_update_irq(s);
         return 1;

     case USB_RET_BABBLE:
         td->ctrl |= TD_CTRL_BABBLE | TD_CTRL_STALL;
         td->ctrl &= ~TD_CTRL_ACTIVE;
+        s->status |= UHCI_STS_USBERR;
+        uhci_update_irq(s);
         /* frame interrupted */
         return -1;

works.

jan

On Thu, May 5, 2011 at 2:05 PM, Gerd Hoffmann <kraxel@redhat.com> wrote:
>> diff --git a/hw/usb-uhci.c b/hw/usb-uhci.c
>> index 346db3e..a51d89b 100644
>> --- a/hw/usb-uhci.c
>> +++ b/hw/usb-uhci.c
>> @@ -732,11 +732,21 @@ out:
>>      case USB_RET_STALL:
>>          td->ctrl |= TD_CTRL_STALL;
>>          td->ctrl&= ~TD_CTRL_ACTIVE;
>> +        s->status |= UHCI_STS_USBERR;
>
> Just this line should be enougth.
>
>>      case USB_RET_BABBLE:
>>          td->ctrl |= TD_CTRL_BABBLE | TD_CTRL_STALL;
>>          td->ctrl&= ~TD_CTRL_ACTIVE;
>> +        s->status |= UHCI_STS_USBERR;
>
> Likewise.
>
> Tried that?
>
> cheers,
>  Gerd
>
>

Re: [Qemu-trivial] [Qemu-devel] Bug #757654: UHCI fails to signal stall response patch

[Gerd Hoffmann]

On 05/06/11 10:07, Jan Vesely wrote:
> hi,
> using the modified (single line) patch only works half-way, it sets
> the value in status register (guess that's what that line does :))
> but hw interrupt is not generated.
>
> I tried adding uhci_update_irq and this patch:

> +        s->status |= UHCI_STS_USBERR;
> +        uhci_update_irq(s);

> works.

Good.  Can you create a patch with a changelog entry & signed-off 
please?  I'll go queue it up then.

thanks,
   Gerd

Re: [Qemu-trivial] [Qemu-devel] Bug #757654: UHCI fails to signal stall response patch

[Jan Vesely]

UHCI host controller status register indicates error and
an interrupt is triggered on BABBLE and STALL errors.

Signed-off-by: Jan Vesely <jano.vesely@gmail.com>
---
 hw/usb-uhci.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/hw/usb-uhci.c b/hw/usb-uhci.c
index a65e0b3..1e9c1e7 100644
--- a/hw/usb-uhci.c
+++ b/hw/usb-uhci.c
@@ -702,11 +702,15 @@ out:
     case USB_RET_STALL:
         td->ctrl |= TD_CTRL_STALL;
         td->ctrl &= ~TD_CTRL_ACTIVE;
+        s->status |= UHCI_STS_USBERR;
+        uhci_update_irq(s);
         return 1;

     case USB_RET_BABBLE:
         td->ctrl |= TD_CTRL_BABBLE | TD_CTRL_STALL;
         td->ctrl &= ~TD_CTRL_ACTIVE;
+        s->status |= UHCI_STS_USBERR;
+        uhci_update_irq(s);
         /* frame interrupted */
         return -1;

-- 
1.7.3.4

Re: [Qemu-trivial] [Qemu-devel] Bug #757654: UHCI fails to signal stall response patch

[Gerd Hoffmann]

On 05/09/11 12:16, Jan Vesely wrote:
> UHCI host controller status register indicates error and
> an interrupt is triggered on BABBLE and STALL errors.

Queued up.

thanks,
   Gerd

Re: [Qemu-trivial] [Qemu-devel] Bug #757654: UHCI fails to signal stall response patch

[Jan Vesely]

glad I could help

the original bug report is here: https://bugs.launchpad.net/qemu/+bug/757654
should I update it? or will it be updated when the patch reaches master?

j.

On Mon, May 9, 2011 at 1:43 PM, Gerd Hoffmann <kraxel@redhat.com> wrote:
> On 05/09/11 12:16, Jan Vesely wrote:
>>
>> UHCI host controller status register indicates error and
>> an interrupt is triggered on BABBLE and STALL errors.
>
> Queued up.
>
> thanks,
>  Gerd
>
>

Re: [Qemu-trivial] [Qemu-devel] Bug #757654: UHCI fails to signal stall response patch

[Gerd Hoffmann]

On 05/11/11 13:33, Jan Vesely wrote:
> glad I could help
>
> the original bug report is here: https://bugs.launchpad.net/qemu/+bug/757654
> should I update it? or will it be updated when the patch reaches master?

Better wait until the stuff hits master, that is less confusing and you 
can also add the git commit hash then.

cheers,
   Gerd