Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-19 Thread Alan Stern
On Wed, 19 Jan 2005, Thomas Winischhofer wrote: > Thanks, Alan. Do you consider the current (1.0.2) patch ok in the above > mentioned regards? Yes, it looks fine. Alan Stern --- The SF.Net email is sponsored by: Beat the post-holiday blues G

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-19 Thread Thomas Winischhofer
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Alan Stern wrote: | Even though you've already updated the driver, you might like to get some | answers to your questions. | | On Tue, 18 Jan 2005, Thomas Winischhofer wrote: | | |>Is -ECONNRESET only set if the URB is (forcefully) unlinked (by me)? | |

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-19 Thread Alan Stern
Even though you've already updated the driver, you might like to get some answers to your questions. On Tue, 18 Jan 2005, Thomas Winischhofer wrote: > Is -ECONNRESET only set if the URB is (forcefully) unlinked (by me)? The status will be set to -ECONNRESET whenever the URB is terminated by som

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-18 Thread Thomas Winischhofer
Greg KH wrote: On Wed, Jan 19, 2005 at 01:14:07AM +0100, Thomas Winischhofer wrote: OK, this one better? if((retval == 0) && timeout) { if(sisusb->urbstatus[index] & SU_URB_BUSY) { add_wait_queue(&sisusb->wait_q, &wait); set_current_state(TASK_INTERRUPTIBLE);

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-18 Thread Thomas Winischhofer
Alan Stern wrote: +static int sisusb_bulkout_msg(struct sisusb_usb_data *sisusb, int index, unsigned int pipe, + void *data, int len, int *actual_length, int timeout, + unsigned int tflags, dma_addr_t transfer_dma) +{ <...> + /* If OK, and if timeout > 0, wait for completion */ + if((retval ==

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-18 Thread Thomas Winischhofer
Greg KH wrote: On Wed, Jan 19, 2005 at 01:14:07AM +0100, Thomas Winischhofer wrote: OK, this one better? if((retval == 0) && timeout) { if(sisusb->urbstatus[index] & SU_URB_BUSY) { add_wait_queue(&sisusb->wait_q, &wait); set_current_state(TASK_INTERRUPTIBLE);

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-18 Thread Thomas Winischhofer
Alan Stern wrote: Since you are only adding a single .[ch] pair of files, why create a whole new subdirectory? As I told Greg in a private mail before, I am planning to add an fbdev interface ("lite" version of sisfb) later. This will increase the number of files to up to 10 or so. +static voi

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-18 Thread Greg KH
On Wed, Jan 19, 2005 at 01:14:07AM +0100, Thomas Winischhofer wrote: > > OK, this one better? > > if((retval == 0) && timeout) { > if(sisusb->urbstatus[index] & SU_URB_BUSY) { > add_wait_queue(&sisusb->wait_q, &wait); > set_current_state(TASK_INTERRUPTIBLE); >

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-18 Thread Greg KH
On Tue, Jan 18, 2005 at 11:25:14PM +0100, Thomas Winischhofer wrote: > Alan Stern wrote: > >Since you are only adding a single .[ch] pair of files, why create a whole > >new subdirectory? > > > As I told Greg in a private mail before, I am planning to add an fbdev > interface ("lite" version of

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-18 Thread Alan Stern
On Tue, 18 Jan 2005, Thomas Winischhofer wrote: > Attached is a kernel patch which adds a driver for SiS315/Net2280-based > USB2VGA dongles. > > The patch is against 2.6.10 vanilla. The driver is for 2.6 only. > However, since I am far from being an expert on USB, (reasonable) > comments are w

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-18 Thread Thomas Winischhofer
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Oliver Neukum wrote: | + add_wait_queue(&sisusb->wait_q, &wait); | + set_current_state(TASK_INTERRUPTIBLE); | + while(1) { | + if(timeout && (!(i = sisusb_all_free(sisusb { | + timeout = schedule_timeout(t

Re: [linux-usb-devel] SiS USB2VGA kernel patch V1.0

2005-01-18 Thread Oliver Neukum
+ add_wait_queue(&sisusb->wait_q, &wait); + set_current_state(TASK_INTERRUPTIBLE); + while(1) { + if(timeout && (!(i = sisusb_all_free(sisusb { + timeout = schedule_timeout(timeout); If you are sleeping interruptably you must handle signals