usb-storage: Wait less time for settling when a memory device is connected

usb-storage: Wait less time for settling when a memory device is connected

[Martin Bammer]

Hi,

I was wondering why the usb-storage driver waits for my usb-stick and my
sdhc-card 5s before scanning. IMHO this is only needed for HDs. I've
created a patch for the driver which shortens the timeout to 1s when a
memory device is detected.
But I'm not shure if the detection code is right for all devices. This
maybe has to be tested and discussed.
I've tested it with a usb-stick, a sdhc-card and a usb-harddrive.
This patch is created for kernel 2.6.27.18.

Please CC me for answers to this thread.

--- a/drivers/usb/storage/usb.c 2009-03-03 23:33:37.000000000 +0100
+++ b/drivers/usb/storage/usb.c 2009-03-03 23:32:44.000000000 +0100
@@ -913,10 +913,16 @@ static int usb_stor_scan_thread(void * _
 
        set_freezable();
 
+       /* Shorten delay time for memory sticks/cards */
+       if ((1 == us->pusb_dev->descriptor.iManufacturer)
+           && (2 == us->pusb_dev->descriptor.iProduct)
+           && (3 == us->pusb_dev->descriptor.iSerialNumber)
+           && (delay_use > 0)) delay_use = 1;
+
        /* Wait for the timeout to expire or for a disconnect */
        if (delay_use > 0) {
                printk(KERN_DEBUG "usb-storage: waiting for device "
-                               "to settle before scanning\n");
+                               "to settle before scanning for %d
seconds\n", delay_use);
                wait_event_freezable_timeout(us->delay_wait,
                                test_bit(US_FLIDX_DONT_SCAN,
&us->dflags),
                                delay_use * HZ);

Re: usb-storage: Wait less time for settling when a memory device is connected

[Greg KH]

On Fri, Mar 06, 2009 at 11:34:07AM +0100, Martin Bammer wrote:
> Hi,

First off, usb issues/patches should be sent to the
linux-usb@vger.kernel.org list so they don't get ignored.

> I was wondering why the usb-storage driver waits for my usb-stick and my
> sdhc-card 5s before scanning. IMHO this is only needed for HDs. I've
> created a patch for the driver which shortens the timeout to 1s when a
> memory device is detected.
> But I'm not shure if the detection code is right for all devices. This
> maybe has to be tested and discussed.
> I've tested it with a usb-stick, a sdhc-card and a usb-harddrive.
> This patch is created for kernel 2.6.27.18.
> 
> Please CC me for answers to this thread.
> 
> --- a/drivers/usb/storage/usb.c 2009-03-03 23:33:37.000000000 +0100
> +++ b/drivers/usb/storage/usb.c 2009-03-03 23:32:44.000000000 +0100
> @@ -913,10 +913,16 @@ static int usb_stor_scan_thread(void * _
>  
>         set_freezable();
>  
> +       /* Shorten delay time for memory sticks/cards */
> +       if ((1 == us->pusb_dev->descriptor.iManufacturer)
> +           && (2 == us->pusb_dev->descriptor.iProduct)
> +           && (3 == us->pusb_dev->descriptor.iSerialNumber)
> +           && (delay_use > 0)) delay_use = 1;

Please always run your patches through scripts/checkpatch.pl first.

Why are you testing for the manufacturer and product here?

And what's wrong with just changing the module parameter for delay_use,
that is what it is there for.  No kernel patch should be needed.

thanks,

greg k-h