double_buffer_not_ok compile error

double_buffer_not_ok compile error

[Michael Jones]

Hi Felipe,

In commit 0662481855c389b75a0a54c32870cc90563d80a9, the member
'double_buffer_not_ok' was added to 'struct musb', but only inside
#ifdef CONFIG_USB_GADGET_MUSB_HDRC.  In musb_host.c, this member is used
outside of the same #ifdef, so I get a compile error because I have
CONFIG_USB_GADGET_MUSB_HDRC=n (because I have CONFIG_USB_GADGET_OMAP=y).

here is the error:
drivers/usb/musb/musb_host.c: In function 'musb_rx_reinit':
drivers/usb/musb/musb_host.c:612: error: 'struct musb' has no member
named 'double_buffer_not_ok'

I'm not familiar enough with this to propose a correct patch, but for
example with the patch below, I can compile the kernel.  Maybe the
member definition can just be taken out of the #ifdef.

Can you propose a proper fix for this?

thanks,
Michael Jones

diff --git a/drivers/usb/musb/musb_host.c b/drivers/usb/musb/musb_host.c
index 0f523d7..809965f 100644
--- a/drivers/usb/musb/musb_host.c
+++ b/drivers/usb/musb/musb_host.c
@@ -609,9 +609,11 @@ musb_rx_reinit(struct musb *musb, struct musb_qh *qh, struct musb_hw_ep *ep)
 	/* Set RXMAXP with the FIFO size of the endpoint
 	 * to disable double buffer mode.
 	 */
+#ifdef CONFIG_USB_GADGET_MUSB_HDRC
 	if (musb->double_buffer_not_ok)
 		musb_writew(ep->regs, MUSB_RXMAXP, ep->max_packet_sz_rx);
 	else
+#endif
 		musb_writew(ep->regs, MUSB_RXMAXP,
 				qh->maxpacket | ((qh->hb_mult - 1) << 11));
 
@@ -784,10 +786,12 @@ static void musb_ep_program(struct musb *musb, u8 epnum,
 		/* protocol/endpoint/interval/NAKlimit */
 		if (epnum) {
 			musb_writeb(epio, MUSB_TXTYPE, qh->type_reg);
+#ifdef CONFIG_USB_GADGET_MUSB_HDRC
 			if (musb->double_buffer_not_ok)
 				musb_writew(epio, MUSB_TXMAXP,
 						hw_ep->max_packet_sz_tx);
 			else
+#endif
 				musb_writew(epio, MUSB_TXMAXP,
 						qh->maxpacket |
 						((qh->hb_mult - 1) << 11));


MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

Re: double_buffer_not_ok compile error

[Felipe Balbi]

Hi,

On Mon, Feb 21, 2011 at 01:59:52PM +0100, Michael Jones wrote:
> In commit 0662481855c389b75a0a54c32870cc90563d80a9, the member
> 'double_buffer_not_ok' was added to 'struct musb', but only inside
> #ifdef CONFIG_USB_GADGET_MUSB_HDRC.  In musb_host.c, this member is used
> outside of the same #ifdef, so I get a compile error because I have
> CONFIG_USB_GADGET_MUSB_HDRC=n (because I have CONFIG_USB_GADGET_OMAP=y).
> 
> here is the error:
> drivers/usb/musb/musb_host.c: In function 'musb_rx_reinit':
> drivers/usb/musb/musb_host.c:612: error: 'struct musb' has no member
> named 'double_buffer_not_ok'
> 
> I'm not familiar enough with this to propose a correct patch, but for
> example with the patch below, I can compile the kernel.  Maybe the
> member definition can just be taken out of the #ifdef.
> 
> Can you propose a proper fix for this?

There's already a fix for that queued on Greg's queue ;-)

Thanks though

Here it is:

commit 5990378b393429244559f4750f2ee3a50929b932
Author: Felipe Balbi <balbi@ti.com>
Date:   Fri Feb 11 10:00:02 2011 +0200

    usb: musb: fix build breakage
    
    commit 0662481855c389b75a0a54c32870cc90563d80a9
    (usb: musb: disable double buffering when it's broken),
    introduced a compile error when gadget API is disabled.
    
    Fix it.
    
    Signed-off-by: Felipe Balbi <balbi@ti.com>

diff --git a/drivers/usb/musb/musb_core.h b/drivers/usb/musb/musb_core.h
index d74a811..e6400be 100644
--- a/drivers/usb/musb/musb_core.h
+++ b/drivers/usb/musb/musb_core.h
@@ -488,6 +488,15 @@ struct musb {
        unsigned                set_address:1;
        unsigned                test_mode:1;
        unsigned                softconnect:1;
+
+       u8                      address;
+       u8                      test_mode_nr;
+       u16                     ackpend;                /* ep0 */
+       enum musb_g_ep0_state   ep0_state;
+       struct usb_gadget       g;                      /* the gadget */
+       struct usb_gadget_driver *gadget_driver;        /* its driver */
+#endif
+
        /*
         * FIXME: Remove this flag.
         *
@@ -501,14 +510,6 @@ struct musb {
         */
        unsigned                double_buffer_not_ok:1 __deprecated;
 
-       u8                      address;
-       u8                      test_mode_nr;
-       u16                     ackpend;                /* ep0 */
-       enum musb_g_ep0_state   ep0_state;
-       struct usb_gadget       g;                      /* the gadget */
-       struct usb_gadget_driver *gadget_driver;        /* its driver */
-#endif
-
        struct musb_hdrc_config *config;
 
 #ifdef MUSB_CONFIG_PROC_FS

-- 
balbi

Re: double_buffer_not_ok compile error

[Michael Jones]

Hi Felipe,

Sorry, thanks.  I tried to avoid getting that reply by searching the
mailing list to see whether this was already discovered.  I searched for
"double_buffer_not_ok" and "musb_host" on the mailing list but didn't
see anything newer than the original patch.

Where is the correct place to search for such a recent fix?  Where is
Greg's queue?

-Michael

On 02/21/2011 02:25 PM, Felipe Balbi wrote:
> Hi,
> 
> On Mon, Feb 21, 2011 at 01:59:52PM +0100, Michael Jones wrote:
>> In commit 0662481855c389b75a0a54c32870cc90563d80a9, the member
>> 'double_buffer_not_ok' was added to 'struct musb', but only inside
>> #ifdef CONFIG_USB_GADGET_MUSB_HDRC.  In musb_host.c, this member is used
>> outside of the same #ifdef, so I get a compile error because I have
>> CONFIG_USB_GADGET_MUSB_HDRC=n (because I have CONFIG_USB_GADGET_OMAP=y).
>>
>> here is the error:
>> drivers/usb/musb/musb_host.c: In function 'musb_rx_reinit':
>> drivers/usb/musb/musb_host.c:612: error: 'struct musb' has no member
>> named 'double_buffer_not_ok'
>>
>> I'm not familiar enough with this to propose a correct patch, but for
>> example with the patch below, I can compile the kernel.  Maybe the
>> member definition can just be taken out of the #ifdef.
>>
>> Can you propose a proper fix for this?
> 
> There's already a fix for that queued on Greg's queue ;-)
> 
> Thanks though
> 
> Here it is:
> 
> commit 5990378b393429244559f4750f2ee3a50929b932
> Author: Felipe Balbi <balbi@ti.com>
> Date:   Fri Feb 11 10:00:02 2011 +0200
> 
>     usb: musb: fix build breakage
>     
>     commit 0662481855c389b75a0a54c32870cc90563d80a9
>     (usb: musb: disable double buffering when it's broken),
>     introduced a compile error when gadget API is disabled.
>     
>     Fix it.
>     
>     Signed-off-by: Felipe Balbi <balbi@ti.com>
> 
> diff --git a/drivers/usb/musb/musb_core.h b/drivers/usb/musb/musb_core.h
> index d74a811..e6400be 100644
> --- a/drivers/usb/musb/musb_core.h
> +++ b/drivers/usb/musb/musb_core.h
> @@ -488,6 +488,15 @@ struct musb {
>         unsigned                set_address:1;
>         unsigned                test_mode:1;
>         unsigned                softconnect:1;
> +
> +       u8                      address;
> +       u8                      test_mode_nr;
> +       u16                     ackpend;                /* ep0 */
> +       enum musb_g_ep0_state   ep0_state;
> +       struct usb_gadget       g;                      /* the gadget */
> +       struct usb_gadget_driver *gadget_driver;        /* its driver */
> +#endif
> +
>         /*
>          * FIXME: Remove this flag.
>          *
> @@ -501,14 +510,6 @@ struct musb {
>          */
>         unsigned                double_buffer_not_ok:1 __deprecated;
>  
> -       u8                      address;
> -       u8                      test_mode_nr;
> -       u16                     ackpend;                /* ep0 */
> -       enum musb_g_ep0_state   ep0_state;
> -       struct usb_gadget       g;                      /* the gadget */
> -       struct usb_gadget_driver *gadget_driver;        /* its driver */
> -#endif
> -
>         struct musb_hdrc_config *config;
>  
>  #ifdef MUSB_CONFIG_PROC_FS
> 


MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

Re: double_buffer_not_ok compile error

[Felipe Balbi]

On Mon, Feb 21, 2011 at 02:54:27PM +0100, Michael Jones wrote:
> Hi Felipe,
> 
> Sorry, thanks.  I tried to avoid getting that reply by searching the
> mailing list to see whether this was already discovered.  I searched for
> "double_buffer_not_ok" and "musb_host" on the mailing list but didn't
> see anything newer than the original patch.
> 
> Where is the correct place to search for such a recent fix?  Where is
> Greg's queue?

Yes, Greg's usb-linux or my gitorious musb branch.

-- 
balbi

Re: double_buffer_not_ok compile error

[Markos Chandras]

[-- Attachment #1: Type: text/plain, Size: 1034 bytes --]

On Mon, Feb 21, 2011 at 03:56:13PM +0200, Felipe Balbi wrote:
> On Mon, Feb 21, 2011 at 02:54:27PM +0100, Michael Jones wrote:
> > Hi Felipe,
> > 
> > Sorry, thanks.  I tried to avoid getting that reply by searching the
> > mailing list to see whether this was already discovered.  I searched for
> > "double_buffer_not_ok" and "musb_host" on the mailing list but didn't
> > see anything newer than the original patch.
> > 
> > Where is the correct place to search for such a recent fix?  Where is
> > Greg's queue?
> 
> Yes, Greg's usb-linux or my gitorious musb branch.
> 
> -- 
> balbi
> --
> To unsubscribe from this list: send the line "unsubscribe linux-omap" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Hi, thanks for that. Ran into the same problem whilst trying the
linux-linaro-2.6.38 tree. I hope your patch will make it to this tree
soon

Regards,
-- 
Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]