On Fri, Mar 15, 2002 at 01:49:45PM +0100, Nemosoft Unv. wrote:
> Hello,
>
> On Monday 11 March 2002 18:38, Greg KH wrote:
> > On Sun, Mar 10, 2002 at 02:42:29AM +0100, Nemosoft Unv. wrote:
> > > The patch is not against a kernel version; rather, it is extracted from
> > > my own CVS repository. However, since the PWC driver is completely
> > > independant, this shouldn't matter. It should apply to both the 2.4 and
> > > 2.5. kernel series.
> >
> > And since it is just from your cvs tree, it does not apply to any kernel
> > tree :(
> > I can mess with the patch to get it to apply to the 2.4 tree, but there
> > are lots of rejections in the 2.5 kernel.
>
> You know, I would really apreciate it if people would send me the necessary
> patches as well (the E-mail address is in every pwc file), in stead of
> sending me on a goose-chase for patches or having to manually verify the
> changes over 3 kernel versions...
Ah, but this is one of the _advantages_ of open source, other people
fixing things in your code when they change things in other parts of the
tree. In looking at just the history of the pwc-if.c file in the 2.5
tree I see changes made to the file by the following people, with the
following reasons:
[EMAIL PROTECTED] v4l: pwc webcam update
[EMAIL PROTECTED] remove uvirt_to_kva(), use vmalloc_to_page()
[EMAIL PROTECTED] Fix up vmalloc page mapping breakage
[EMAIL PROTECTED] update the driver to the new, simpler
vmalloc_to_page() interface
[EMAIL PROTECTED] remove uvirt_to_kva() and replace it with athe
vmalloc_to_page_interface()
[EMAIL PROTECTED] added mem_flags to submit_urb()
[EMAIL PROTECTED] usb_submit_urb() changes
Feel free to request your driver to be pulled from the kernel tree if
you don't like changes like the above being done to the code :)
Personally, I'm happy when others take an interest in fixing problems in
"my" drivers, when they change the interfacing code. The filesystem
changes recently done by Al Viro is a great example of this (he changed
the way superblocks and other stuff is handled in the core, and modified
the usbfs code to work properly, preventing that code from breaking.)
Also realize that the v4l interfaces are changing right now in 2.5, so
you will end up with two different drivers over time.
Because of this, your 2.5.5 patch you attached does not apply to the
most recent kernel tree, sorry. I get the following errors:
[greg@soap gregkh-2.5]$ patch -p1 -g1 --dry-run < ../../pwc-2.5.5.diff
drivers/usb/pwc-ctrl.c 1.9: 1205 lines
patching file drivers/usb/pwc-ctrl.c
Hunk #9 FAILED at 1136.
Hunk #10 FAILED at 1380.
Hunk #11 FAILED at 1393.
Hunk #12 FAILED at 1434.
4 out of 12 hunks FAILED -- saving rejects to file drivers/usb/pwc-ctrl.c.rej
drivers/usb/pwc-if.c 1.18: 2013 lines
patching file drivers/usb/pwc-if.c
Hunk #6 FAILED at 127.
Hunk #7 succeeded at 169 (offset 4 lines).
Hunk #8 succeeded at 179 with fuzz 2.
Hunk #9 FAILED at 219.
Hunk #10 succeeded at 348 (offset -5 lines).
Hunk #12 succeeded at 606 (offset -5 lines).
Hunk #14 succeeded at 867 (offset -5 lines).
Hunk #16 succeeded at 966 (offset -4 lines).
Hunk #17 FAILED at 1601.
Hunk #18 succeeded at 1565 (offset -50 lines).
Hunk #19 succeeded at 1643 (offset -4 lines).
Hunk #20 succeeded at 1618 (offset -50 lines).
Hunk #21 succeeded at 1716 (offset -4 lines).
Hunk #22 succeeded at 1681 (offset -50 lines).
Hunk #23 succeeded at 1744 (offset -4 lines).
Hunk #24 succeeded at 1715 (offset -50 lines).
Hunk #25 succeeded at 1772 (offset -4 lines).
Hunk #26 succeeded at 1735 (offset -50 lines).
Hunk #27 succeeded at 1833 (offset -4 lines).
Hunk #28 succeeded at 1795 (offset -50 lines).
Hunk #29 succeeded at 2005 (offset -4 lines).
Hunk #30 succeeded at 2016 (offset -50 lines).
3 out of 30 hunks FAILED -- saving rejects to file drivers/usb/pwc-if.c.rej
drivers/usb/pwc-ioctl.h 1.4: 123 lines
patching file drivers/usb/pwc-ioctl.h
drivers/usb/pwc-misc.c 1.3: 104 lines
patching file drivers/usb/pwc-misc.c
drivers/usb/pwc-uncompress.c 1.3: 190 lines
patching file drivers/usb/pwc-uncompress.c
drivers/usb/pwc-uncompress.h 1.2: 77 lines
patching file drivers/usb/pwc-uncompress.h
drivers/usb/pwc.h 1.11: 267 lines
patching file drivers/usb/pwc.h
> > Can you please resend this against the latest kernel trees?
>
> I did, but it's ugly. I had to put in a few #if LINUX_VERSION because of
> all the changes in memory mamagement (and I don't want to maintain two
> separate source trees). The patch against 2.4.18 works (verified by me),
> but the 2.5.5 branch is unsure; it would not boot for me. That kernel is
> fishy anyhow: NFS was broken (wouldn't compile for me), and during 'make
> modules_install' depmod complained about exactly those memory management
> functions that were changed in 2.5.[3-5]. So I'm sceptical.
I've applied the 2.4 patch to my tree, and will push that to Marcelo
soon.
thanks,
greg k-h
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
