2006-02-14 15:38:03

by Paolo Ornati

[permalink] [raw]
Subject: [trivial PATCH] "drivers/usb/media/stv680.h": fix jiffies timeout

From: Paolo Ornati <[email protected]>

stv680.c driver calls "usb_control_msg" passing PENCAM_TIMEOUT as
jiffies timout. However PENCAM_TIMEOUT is defined to the fixed value of
1000, this leads to different timeouts with different HZ settings.

Since stv680.c is there since 2.4.18 I don't know if 1000 means 10s or
1s... I've picked the bigger.

---

diff --git a/drivers/usb/media/stv680.h b/drivers/usb/media/stv680.h
index b0551cd..b1a4bf5 100644
--- a/drivers/usb/media/stv680.h
+++ b/drivers/usb/media/stv680.h
@@ -45,7 +45,7 @@
#define USB_CREATIVEGOMINI_VENDOR_ID 0x041e
#define USB_CREATIVEGOMINI_PRODUCT_ID 0x4007

-#define PENCAM_TIMEOUT 1000
+#define PENCAM_TIMEOUT (10*HZ)
/* fmt 4 */
#define STV_VIDEO_PALETTE VIDEO_PALETTE_RGB24


--
Paolo Ornati
Linux 2.6.15.4-suspend2 on x86_64


2006-02-14 16:15:17

by Nishanth Aravamudan

[permalink] [raw]
Subject: Re: [trivial PATCH] "drivers/usb/media/stv680.h": fix jiffies timeout

On 14.02.2006 [16:33:12 +0100], Paolo Ornati wrote:
> From: Paolo Ornati <[email protected]>
>
> stv680.c driver calls "usb_control_msg" passing PENCAM_TIMEOUT as
> jiffies timout. However PENCAM_TIMEOUT is defined to the fixed value of
> 1000, this leads to different timeouts with different HZ settings.
>
> Since stv680.c is there since 2.4.18 I don't know if 1000 means 10s or
> 1s... I've picked the bigger.

NACK. PENCAM_TIMEOUT is a *milliseconds* timeout value.

>From the comment for usb_control_msg:

* @timeout: time in msecs to wait for the message to complete before
* timing out (if 0 the wait is forever)

Milliseconds do not depend on HZ in anyway.

Thanks,
Nish

2006-02-14 17:37:29

by Paolo Ornati

[permalink] [raw]
Subject: Re: [trivial PATCH] "drivers/usb/media/stv680.h": fix jiffies timeout --

On Tue, 14 Feb 2006 08:15:35 -0800
Nishanth Aravamudan <[email protected]> wrote:

>
> NACK. PENCAM_TIMEOUT is a *milliseconds* timeout value.
>
> From the comment for usb_control_msg:
>
> * @timeout: time in msecs to wait for the message to complete before
> * timing out (if 0 the wait is forever)
>
> Milliseconds do not depend on HZ in anyway.
>
> Thanks,
> Nish

Opsss... I was using a source browser ;)

http://lxr.linux.no/source/drivers/usb/core/message.c#L118

It's a bit outdated... (2.6.11)

--
Paolo Ornati
Linux 2.6.15.4-suspend2 on x86_64


--
Paolo Ornati
Linux 2.6.15.4-suspend2 on x86_64