On Tue, 2006-07-11 at 22:31 +0400, Sergei Organov wrote:
Andy Gay [EMAIL PROTECTED] writes:
Adapted from an earlier patch by Greg KH [EMAIL PROTECTED].
That patch added multiple read urbs and larger transfer buffers to allow
data transfers at full EvDO speed.
Below are two more problems
On Mon, 2006-07-03 at 00:00 -0700, Jeremy Fitzhardinge wrote:
Andy Gay wrote:
BTW - Jeremy suggested that the number of EPs to configure should be
determined from the device ID. Makes sense to me, but then many users
may have no use for the additional EPs. Alternatively, Greg suggested
On Mon, 2006-07-03 at 08:43 -0700, Ken Brush wrote:
On 6/29/06, Andy Gay [EMAIL PROTECTED] wrote:
Adapted from an earlier patch by Greg KH [EMAIL PROTECTED].
That patch added multiple read urbs and larger transfer buffers to allow
data transfers at full EvDO speed.
This version
On Mon, 2006-07-03 at 09:28 -0700, Jeremy Fitzhardinge wrote:
Andy Gay wrote:
I think if the hardware has the EPs, they should be exposed by the
driver. You can tweak usermode as to whether they get device nodes,
what they're called, etc.
I tend to agree. I'm thinking for now I
On Mon, 2006-07-03 at 10:00 -0700, Greg KH wrote:
Yes, this driver is already split into 2 different ones (look in the
recent -mm releases). Sierra wants to have their devices be in their
own driver, as the chip is a little different from the other ones. This
means that those devices are now
On Mon, 2006-07-03 at 11:16 -0700, Greg KH wrote:
On Mon, Jul 03, 2006 at 01:55:28PM -0400, Andy Gay wrote:
On Mon, 2006-07-03 at 10:00 -0700, Greg KH wrote:
Yes, this driver is already split into 2 different ones (look in the
recent -mm releases). Sierra wants to have their devices
On Sun, 2006-07-02 at 11:48 -0700, Roland Dreier wrote:
this works well on my kyocera kpc650 -- throughput is up to about 1
mbit/sec vs. ~250 kbit/sec with the stock airprime driver.
-
Thanks for the feedback.
I'm working on fixing the concerns Andrew Morton expressed regarding
memory leaks in
On Fri, 2006-06-30 at 00:10 -0700, Andrew Morton wrote:
...
+static void airprime_read_bulk_callback(struct urb *urb, struct pt_regs
*regs)
+{
...
+ tty = port-tty;
+ if (tty urb-actual_length) {
+ tty_buffer_request_room(tty, urb-actual_length);
+
On Fri, 2006-06-30 at 01:52 -0700, Pete Zaitcev wrote:
The idea to allocate a URB for every little user write bothers me as
well. It was a dirty code thrown together quickly by someone who could
not be bothered to use a circular buffer and two URBs. It was fine
for the visor.c, but the
On Fri, 2006-06-30 at 13:04 -0700, Roland Dreier wrote:
+ /* something happened, so free up the memory for this urb /*
an obvious glitch here at the end of the line...
Oops. Sorry 'bout that.. that comment had some more stuff that no longer
applies, I edited it just before I
On Thu, 2006-06-29 at 19:40 -0700, Roland Dreier wrote:
or:
- send a patch against 2.6.17 that is my changes + your fixes to
actually make it work.
My patch was just a throw it out there and see what works or not, as I
don't even have the device to test it with.
I would
(0x0c88, 0x17da) for the Kyocera Wireless KPC650/Passport
was added but is not yet tested.
Signed-off-by: Andy Gay [EMAIL PROTECTED]
---
commit 3d1346863aac4b3c016acb409a3b9e6651af8f7a
tree f4359718b8550ce0d95b57ba1b5b0d902bf2ada8
parent 501b7c77de3e90519e95fd99e923bf9a29cd120d
author andy [EMAIL
I have adapted the modified Airprime driver that Greg posted a few weeks
ago to add support for these 2 modules.
That driver works for these modules if the USB IDs are added, and fixes
the throughput problems in the earlier driver. I had to make some
changes though -
- there's a memory leak
On Wed, 2006-06-28 at 17:11 -0700, Jeremy Fitzhardinge wrote:
Andy Gay wrote:
- these modules present 3 bulk EPs, the 2nd 3rd can be used for
control status monitoring while data transfer is in progress on the
1st EP. This is useful (and necessary for my application) so we need
14 matches
Mail list logo