Hi Laurent & all

Regarding Logitech-problems, I have just been trying out a number of UVC
revisions tonight. As you suggested.

First, I must admit, that my current kernel somehow has advanced from
the below mentioned "2.6.21-1.3228.fc7" to a 2.6.22 version; I must have
accepted one of the "updates" within FC7 without paying attention.

With this kernel 2.6.22, I am unable to reproduce any Logitech webcams
working "for periods of multiple minuttes at a time". I have tried
revisions 120, 119, 118, 115, 114, 112, 111, 110, 92 (I also tried to
check out revisions 117, 116, 113 and 91, but the subversion client
apparently just... hanged). I have kept the revisions in separate
directories. I have done a lot of "modprobe -r uvcvideo", "make", "make
install", "modprobe uvcvideo" to shift between the revisions. There are
minor differences, bot common for all these revisions are, that both my
Logitech Fusion and my Logitech Ultra Vision fails within five seconds.
I do not see any complete, stable images as seen on screen with
"luvcview". The images have either large rectangles with noise or many
small rectangles with noise; the images look really bad.

The somewhat better results as mentioned below from around Friday 20.
July and with the 2.6.21 kernel I can not reproduce. I wonder if this
kernel is still floating around upon my system, so I can make grub boot
with it??

Together with the two Logitech webcams I have attached a Creative Optia
AF. The video-stream from this Optia AF webcam is pretty stable
throughout all the revisions tried tonight. However, the stream does
tend to produce a single bad image with an interval of about 5-10
seconds. From some point in the middle of the frame, the rest of the
frame does not look right. Except from this, the video-stream from the
Creative webcam is really, really good.

I have also tried playing around with the Logitech webcams and image
sizes - just to see if I could trigger some better or worse response.
The one Logitech would not display 960x720, while trying to access the
second with this resolution made my computer lock up completely! No more
keyboard or mouse. Pressing the power-button on the chasses did produce
some kind of fc-shutdown dialog, but still with no keyboard. I have
mouse and keyboard attached through USB, not PS/2. This was with
revision 120.

As of right now, I can not reproduce any of my Logitech webcams working
"for periods of multiple minuttes". But they did, once.

Below is a listing of the kernel version (I am running Fedora 7 - and my
installation should be completely up-to-date) together with a listing of
the USB devices and the hieracy. The three webcams have at all times
been attached through a USB hub.

Regards
Morten Sabroe Mortensen

   -----
[EMAIL PROTECTED] ~]# uname -a
Linux localhost.localdomain 2.6.22.1-41.fc7 #1 SMP Fri Jul 27 18:10:34
EDT 
2007 i686 i686 i386 GNU/Linux
   -----
[EMAIL PROTECTED] ~]# lsusb
Bus 004 Device 009: ID 046d:c01e Logitech, Inc. MX518 Optical Mouse
Bus 004 Device 007: ID 10d5:5552 Uni Class Technology Co., Ltd
Bus 004 Device 006: ID 046d:08c9 Logitech, Inc.
Bus 004 Device 005: ID 041e:4058 Creative Technology, Ltd
Bus 004 Device 004: ID 046d:08ca Logitech, Inc.
Bus 004 Device 003: ID 04cc:1520 Philips Semiconductors
Bus 004 Device 002: ID 0409:0058 NEC Corp. HighSpeed Hub
Bus 004 Device 001: ID 0000:0000
Bus 001 Device 001: ID 0000:0000
Bus 003 Device 001: ID 0000:0000
Bus 002 Device 001: ID 0000:0000
[EMAIL PROTECTED] ~]# lsusb -t
Bus#  4
`-Dev#   1 Vendor 0x0000 Product 0x0000
  |-Dev#   2 Vendor 0x0409 Product 0x0058
  | |-Dev#   4 Vendor 0x046d Product 0x08ca <- Logitech webcam!
  | |-Dev#   5 Vendor 0x041e Product 0x4058 <- Create Optia AF!
  | `-Dev#   6 Vendor 0x046d Product 0x08c9 <- Logitech webcam!
  `-Dev#   3 Vendor 0x04cc Product 0x1520
    |-Dev#   7 Vendor 0x10d5 Product 0x5552
    `-Dev#   9 Vendor 0x046d Product 0xc01e
Bus#  3
`-Dev#   1 Vendor 0x0000 Product 0x0000
Bus#  2
`-Dev#   1 Vendor 0x0000 Product 0x0000
Bus#  1
`-Dev#   1 Vendor 0x0000 Product 0x0000
[EMAIL PROTECTED] ~]#
   -----

 

-----Oprindelig meddelelse-----
Fra: Morten Mortensen 
Sendt: 13. august 2007 10:44
Til: 'Laurent Pinchart'; [email protected]
Emne: SV: [Linux-uvc-devel] Patch for USB problem with Logitech webcams?


Yes, I am pretty certain, that I used the same kernel for both tests. My
Fedora 7 system did have its kernel upgraded at some point some weeks
back from... a 2.6.19 to a 2.6.21, I belive out from memory. But this
kernel-upgrade was weeks before I changed UVC drivers, I think.

Yes, I will try to find out which SVN revision made a difference. I can
try it out within a day or two, I expect. 

But right now, I can tell you for certain, that when the UVC driver
worked "for periods of multiple minuttes at a time", I wrote to mr.
Xhaard asking about the luvcview output with cc to this list about this
very problem. This was Friday 20. July - I see from my earlier posting,
that my kernel was the current "2.6.21-1.3228.fc7". So no kernel changes
was made in between. An instance of the UVC driver working "for periods
of a few seconds only" is the one which I checked out from the
Subversion repository last night about 12 hours ago. From my point of
view, something has changed in between. I have not reconfigured the
physical USB setup or anything like that, but I have been playing around
with other (USB) drivers (I say this because you mention, that the
problem is somehow related to timing; maybe something else than changes
in the UVC source code triggered something).

I will *try* to narrow this down relative to SVN revisions.

Regards
Morten Sabroe Mortensen
 

-----Oprindelig meddelelse-----
Fra: Laurent Pinchart [mailto:[EMAIL PROTECTED] 
Sendt: 13. august 2007 10:04
Til: [email protected]
Cc: Morten Mortensen
Emne: Re: [Linux-uvc-devel] Patch for USB problem with Logitech webcams?

Hi Morten,

On Monday 13 August 2007 08:25, Morten Mortensen wrote:
> - I own a couple of these "fine" Logitech webcams. Something *is*
changing
> in the UVC-driver, because the driver I had until yesterday - a couple
of
> weeks old - could run the Ultra Vision and the Fusion for a period of
half
> a minute and up to three, four or even five minuttes.
>
> The latest UVC-driver runs for one to three seconds, so something has
been
> changed, which influences this by an order of magnitude. These
scenarios
> are absolutely identical for my instances of the Ultra Vision and the
> Fusion.

Have you used the same kernel for both tests ? Could you please try to
find 
out which SVN revision introduced the problem ?

> One wonders how Logitech circumvents the problem in their Windows
drivers.

They don't. The Windows USB stack seems to be slower than the Linux one,
and 
it just doesn't trigger the problem. It's a matter of timings, and
Logitech 
never noticed the problem. From the limited amount of knowledge I have
about 
this bug, it seems there is a hardware problem which has been
worked-around 
by a firmware hack in recent webcams. Both the hardware and the firmware
are 
provided by the IC vendor, so Logitech isn't the only one to blame.

> One wonders *when* manufacturers implement reflashability of their
> monitors, webcams, etc.

The two issues here are support and cost. They don't want to support
firmware 
reflashing, as that would mean replacing dead webcams. And flash costs
more 
than ROM.

> Please keep up the good work.

Thanks. I'll do my best :-)

Regards,

Laurent Pinchart
_______________________________________________
Linux-uvc-devel mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/linux-uvc-devel

Reply via email to