).
Greg, how about doing an additional retry in the HCD? It will make the whole
USB more reliable.
best regards
Wolfgang Mües
---
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine
not be possible if the device
has no intact copy of the data at retry time.
And there is another benefit of doing it in HCD: all device drivers will get
the desired behaviour at once. Make the life easy for device driver writers!
best regards
Wolfgang Mües
an idea what is going on?
best regards
Wolfgang Mües
---
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive? Does it
help you create better code? SHARE THE LOVE, and help us help
YOU! Click
. auerswald
driver because I wait for the isdn support beeing in place.
best regards
Wolfgang Mües
---
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive? Does it
help you create better code
of failure or retry.
But I won't wait until 2.7
best regards
Wolfgang Mües
---
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills. Sign up for IBM's
Free Linux Tutorials. Learn everything
, urb_unlink should work on all urbs in a dependable way.
No exception.
best regards
Wolfgang Mües
---
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills. Sign up for IBM's
Free Linux Tutorials
Hello Dave,
On Monday 08 December 2003 21:15, David Brownell wrote:
Sure, but I can't think of that as anything other than a
set of long-standing driver bugs.
If few drivers are wrong: shame on the driver writer.
If most drivers are wrong: shame on the API designer.
best regards
Wolfgang
are called.
In general, a driver can acquire the serialize semaphore whenever it wants
to block one of the functions above temporarily.
Will there be function calls / macros for this semaphore handling?
So that you can change implementation without breaking interface?
best regards
Wolfgang Mües
Hello Greg,
this patch adds a new device which uses the same protocol:
This patch is against 2.4.24, but I think it should apply to 2.6 also.
best regards
Wolfgang Mües
-
diff -urN linux-2.4.24/drivers/usb/auermain.c
Alan,
On Monday 19 January 2004 05:07, Alan Stern wrote:
I still have a problem with marking devices as soon as we know they are
unplugged. Even using David's scheme where this is done using a separate
unprotected flag in struct usb_device, I can't see any way to make it work
without
Hello Alan,
On Thursday 10 February 2005 16:41, Alan Stern wrote:
We've discussed in the past that the notion of unlinking a single URB
is unnatural. When a driver has multiple URBs queued for an endpoint
it will always want to unlink all of them together; it will never
want to unlink only
Hello Alan!
On Thursday 10 February 2005 22:12, Alan Stern wrote:
On Thu, 10 Feb 2005, Wolfgang Mües wrote:
My driver (auerswald) tries so send more than one URB to ep0 at a
time! But - I have learned that this does not work reliable,
especially with the UHCI hardware. This is mostly
Hello Marco,
On Wednesday 11 May 2005 08:35, Marco Schramel wrote:
we use linux 2.4.25. Our device has 2 endpoint (bulk in/bulk out). I
am writing a device driver which serves the bulk endpoints
periodically. Exactly i start a timer. If this timer expires the
timer service will send a write
Hello,
I am in the process of selecting an USB chip with device and/or host
function for en embedded linux system (the CPU has no builtin USB).
Device and host function is not needed in parallel, so a switchable
device or an OTG device will be fine.
Any recommendations?
best regards
Wolfgang
Hello Olav,
On Thursday 26 May 2005 08:10, Olav Kongas wrote:
I am in the process of selecting an USB chip with device and/or
host function for en embedded linux system (the CPU has no builtin
USB).
Would have been good to know more about your platform and
needs (can you do DMA, do you
Hello Olav,
On Thursday 26 May 2005 08:10, Olav Kongas wrote:
Philips:
isp1362, full speed, host driver available (I think both
with and without DMA versions, see list archives)
H I have read in the errata that this chip is unable to transfer
an odd number of bytes... and that
Wolfgang Mües
___
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
http://lists.sourceforge.net/lists/listinfo/linux-usb-devel
is hand-crafted.
And no link to a syntax explanation...
Many thanks for your comments, Dave.
Wolfgang Mües
___
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
http://lists.sourceforge.net/lists/listinfo/linux-usb-devel
. And maybe I get some nice code reviews, because I'm
*not* an experienced kernel developer
have fun!
Wolfgang Mües
// small library to form a chain of urbs.
// By Wolfgang Mües ([EMAIL PROTECTED])
#include gpl.licence// as usual
// urb chain element
struct auerchain
On Tuesday, 12. June 2001 18:48, Dmitri wrote:
Quoting Wolfgang Mües [EMAIL PROTECTED]:
On Monday, 11. June 2001 23:24, Brad Hards wrote:
timeout is in jiffies, and is the time waited from URB submission to
getting a -ETIMEOUT if things go wrong. Normally HZ is a good thing to
put
On Wednesday, 13. June 2001 23:12, Pete Zaitcev wrote:
From: Wolfgang Mües [EMAIL PROTECTED]
Can you explain timeout is in jiffies to a user who don't know
anything about jiffies and ticks ?
Wolfgang is trying his best to be difficult, so I am not
replying to his trolling.
Pete,
I do
On Thursday, 14. June 2001 21:01, Alan Cox wrote:
I do this with intention: *you* as an old-fashioned linux
programmer - and many other here on the net - is knowing
very well what a jiffie is.
1/HZ of a second. Where HZ is a preprocessor constant dependant on
platform, cpu and other
On Saturday, 16. June 2001 18:50, Georg Acher wrote:
On Sat, Jun 16, 2001 at 02:48:39PM +0200, Wolfgang Mües wrote:
In dabusb.c, I have seen some statical allocated URBs inside the driver
data structures. I assume that these are not the only ones without
initialized lock
Where
On Friday, 6. July 2001 15:41, Wolfgang Mües wrote:
may I ask for some help with my current problem?
[Problem with usb_control_msg]
I have found the source of the problem .. it was a bad device (software)!
I have got a new device - and all is up and running!
best regards
Wolfgang Mües
Hello,
I have a question regarding the right design of
an USB character driver:
at open(), the device descriptor is allocated from the heap.
Inside the descriptor is the mutex to protect against user space
race conditions. The pointer to the device descriptor is stored
in file-private_data.
On Friday, 2. November 2001 23:59, Dmitri wrote:
If you have initialization done in probe() then it is automatically
wrong, at very least because you are not running in freely blockable
thread.
If I *have* to do some initialisation in probe(), what should I do?
(Sending about 5-10 control
On Wednesday, 14. November 2001 20:18, Pete Zaitcev wrote:
I did not see the original message from Dmitry. I disagree,
probe() only happens on the context of the insmod or its
children, or in init, or khubd. It certainly can call schedule().
Yes. But I have problems in calling functions like
incremented, but then it
gets the next address. (8 in this case)
Me too!
best regards
Wolfgang Mües
--
Durch den Tod sagt Dir die Natur,
daß Du langsamer fahren sollst...
___
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https
On Friday, 16. November 2001 09:36, Brad Hards wrote:
[USB device no accepting new address]
Err, some detail?
What kernel, what architecture, what device, what behaviour with the
retries set to 2, what behaviour with the retries set to 3?
Sorry Brad,
I have to correct myself. I got the
On Monday, 24. December 2001 14:11, Georg Acher wrote:
On Sun, Dec 23, 2001 at 05:49:17PM -0800, Greg KH wrote:
Now to try to get rid of those *_t and p* typedefs in usb.h...
p* can be removed (I don't like it either), but why the _t? It's a clear
marking that this is a defined type and it
On Monday, 24. December 2001 22:36, Greg KH wrote:
Good point. I'll leave the urb_t typedef alone for now as it's used by
quite a bit of code. The p* typedefs are easy to remove, and I'll do
that.
This will break my code =8-(
I find the p*_t typedefs handy because they relax me from
the
On Tuesday, 8. January 2002 15:16, Alan Cox wrote:
For those paying attention, this driver uses a new USB minor number, and
some new ioctls. Here's a patch that Wolfgang sent me that documents
these changes. I'll be sending it on to the proper people, but wanted
to post it here for
Hello Georg!
On Tuesday, 22. January 2002 14:17, Georg Acher wrote:
this patch for usb-uhci fixes the the following problems:
- one shot interrupts work now, there was a wrong data toggle...
- correct timeout check with time_after-macros
- memory leak in synchronous unlinking of interrupt
Hello David!
So, the INTR-OUT problem. For example, maxpacket size is 64,
period is maybe 8, but the driver needs to send 100 bytes. That
adds up to 64 bytes in one frame, then 36 bytes after a delay of
8 frames ... then wanting to stop/unlink. This is one problem that
the LEGO folk
Hello David,
API CHANGE #1: Let INTR completion handlers change
the transfer_buffer_length. A value of -1 (it's not
a size_t!!) would mean don't transfer anything. Any
other value, up to the maximum allowed by the endpoint
descriptor (which is what would control the bandwidth
allocation
Hello Greg,
Well I messed up in giving the auerswald driver the minor numbes 80-95.
Turns out those minors are used by the dc2xx driver :)
AND the hiddev driver uses minor 96-111, which was not documented
_anywhere_ (in dc2xx's defence, the linux-usb.org site did show this.
now it is
Hello all,
I am using Kernel 2.4.10 as Target. I have tried to run my driver in SMP
mode, but the probe() function is locking the whole system, on the first
sleep.
Is there a bug in 2.4.10 which forbidds sleeping in probe ?
Not that I know of. Does a more recent kernel not show this
Hello,
Your driver should do like drivers/usb/usb.c (in this order):
#ifdef CONFIG_USB_DEBUG
#define DEBUG
#else
#undef DEBUG
#endif
#include linux/usb.h
using DEBUG is *not* a good idea!
I have to change
#define DEBUG
into
+define DEBUG 1
because there are some kernel
Thomas Sailer wrote:
Yes. But still we should probably have control queueing for
the uhci's (possibly at the HCD level?)
Yes!
You will also need it if you have to send control requests from interrupt
context.
best regards
Wolfgang
Hello,
On Monday 30 September 2002 16:24, Alan Stern wrote:
Actually, it is possible to receive EPIPE for a control endpoint. But it
shouldn't happen very often. In fact, with normal use we shouldn't see it
at all.
Oh yes, I have seen it. EPIPE on control endpoints. Mostly with the
Hello David,
On Sunday 13 October 2002 22:04, David Brownell wrote:
When you say it does not work, do you mean you've been sitting on a
bug report all this while? Because I don't recall *EVER* hearing about
any problems in the bandwidth reservation code. Not one.
Bandwidth reservation is
Hello Pavel,
On Sunday 15 December 2002 17:34, Pavel Machek wrote:
C++ -style comments should not be used in kernel code.
Can you point me to this rule?
It is not mentioned in Documentation/CodingStyle,
and was adopted by ANSI in 1999.
best regards
Wolfgang
have to guarantee that usb_submit_urb() returns before
the completion handler is called?
Why should you want to guarantee this?
best regards
Wolfgang Mües
---
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld
and racefree.
Don't introduce exceptions simply because you are lazy. It is a pain.
I wish the usb core driver writers would take their heads around and look at
the API from the device drivers view.
Just me 2 cent
best regards
Wolfgang Mües
nothing to me.
best regards
Wolfgang Mües
---
This SF.NET email is sponsored by: Take your first step towards giving
your online business a competitive advantage. Test-drive a Thawte SSL
certificate - our easy online guide will show you how
it on an URB that is not yet linked is a driver bug.
Calling it on an URB that is no longer linked is not an error,
it's unavoidable.
How do you differ between the two cases? They may be the same.
best regards
Wolfgang Mües
---
This SF.NET
?
If it is common to get meaningless errors from functions, device driver
writers will come to ignore errors ... maybe all errors.
IMHO it is good design that cleanup functions don't have an return errorcode.
best regards
Wolfgang Mües
hardware doesn't know better.
So you are better off doing some retries in software if you get
EILSEQ from UHCI.
best regards
Wolfgang Mües
---
This SF.Net email sponsored by Black Hat Briefings Training.
Attend Black Hat Briefings Training, Las
- that set_configuration is not usable
inside probe(). I have looked at the hotplug scripts, but I have found
no example for doing this. The usb hotplug scripts seems not to know
about different configurations?
best regards
Wolfgang Mües
the
set_configuration() before probe() is called?
best regards
Wolfgang Mües
---
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http
needed?
From usb.h, I know that urb-timeout is a valid input parameter for all
types of urbs, regardless of synchronous or asynchronous submitting.
So when the timeout expires, the hcd has to call the completion function
anyway?
best regards
Wolfgang Mües
Hello Tilman,
On Thursday 07 July 2005 02:44, Tilman Schmidt wrote:
What's the proper way to handle the case of a control write
URB completing with status == -EPIPE?
-EPIPE on control endpoint is a protocoll stall. The USB device
has not understand what you are saying.
There are two common
On Monday 24 October 2005 17:04, Philip Hindman wrote:
I'll assume that if usb_submit_urb() fails, my complete routine
wasn't called.
Yes.
Philip, I don't like this also...
In my driver, if usb_submit_urb() fails, I set urb-status to contain
the error code, and the call the completion
see
http://www.newhousenews.com/archive/francis042904.html
best regards
Wolfgang Mües
---
This SF.Net email is sponsored by: Oracle 10g
Get certified on the hottest thing ever to hit the market... Oracle 10g.
Take an Oracle 10g class now
Greg,
the appended patch adds some sanity checking to the buffer handling code
and fixes an kernel oops at disconnect or reconnect time because of an
endless urb resubmit.
Please apply to 2.4.
best regards
Wolfgang Mües
diff -urN linux-2.4.26/drivers/usb/auerbuf.c
linux-2.4.26-auerusb-1.2.7
to go from here?
best regards
Wolfgang Mües
---
This SF.Net email is sponsored by: Oracle 10g
Get certified on the hottest thing ever to hit the market... Oracle 10g.
Take an Oracle 10g class now, and we'll give you the exam FREE.
http
.
If the device really does have requirements for relative timing of
interrupt and control messages
No.
best regards
Wolfgang Mües
---
This SF.Net email is sponsored by: Oracle 10g
Get certified on the hottest thing ever to hit the market... Oracle
is not present in any Suse System I use normally for development.
Can someone give me hint what may cause this problem?
best regards
Wolfgang Mües
---
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps Web
urbs in the queue are really unlinked??
Can this actually happen?
best regards
Wolfgang Mües
---
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies
Hello Pete,
On Monday 07 November 2005 03:31, Pete Zaitcev wrote:
The difficulty is, in 2.6.14, both ub and usb-storage clear halts
when the device stalls (in ub_probe and usb_stor_Bulk_max_lun). It
may be incorrect, because it's a stall on the control pipe, not a
bulk pipe. But I have no
Hello Javier,
On Monday 21 November 2005 21:32, Javier Campos wrote:
Attempting to use the gadget layer for ethernet with a windows host
gives me some problems. I am using the gadget backport in 2.4 on a
Texas Instrument 2420 omap board.
The problem is that if the USB cable is connected
On Friday 27 January 2006 08:38, Ulrich Windl wrote:
(Note: For SuSE's 10.0 kernel 2.6.13-15.7-default I have the problem
with ehci_hcd on a USB 2.0 device: The devices also seem to not
accept the address, and when I unload the ehci_hcd module, the
device successfully connects as USB1. Bug
Hello Pete,
On Sunday 19 February 2006 07:49, Pete Zaitcev wrote:
I see... This is a device which dislikes the stall clear when called
after a stall on the control pipe.
Just for the record... the device is doing a protocoll stall of the
control pipe... and in reaction to that, you are doing
Hello Paul,
On Saturday 27 May 2006 11:27, Paul Giblock wrote:
Another problem is the way the radio seeks. To increase the
frequency I must send two messages, in sequence, with a delay in
between them. Then, I must wait for the next report sent from the
device.
Why don't you use kernel
On Friday 13 October 2006 19:20, Open Source wrote:
Alan -- yes, I understand the ability to increase throughput
by transfering more bytes and I am definitely able to see
better overall throughput when increasing the number
of bytes per transaction. However, I needs to still have
good
Hi,
On Friday 13 October 2006 21:31, Open Source wrote:
Thanks for the input. However, I am not understanding
exactly why kernel mode is treated any differently than
user mode for this sort of thing. I am looking at the code
in ehci-q.c and ehci-hcd.c.
It seems like the unlinking of
On Saturday 14 October 2006 01:02, Open Source wrote:
Either the problem is in the ehci code or in devio.c. In
devio.c the user space process calls an ioctl to reap the
urb (blocking until it is complete). The asynchronous
callback for the urb is called when the urb is unlinked
and that
On Saturday 14 October 2006 18:40, Open Source wrote:
I'm not so sure about that. Once the interrupt is finished being
serviced, the scheduler should immediately wake up
and service any blocked processes rather than sitting
idle for one CONFIG_HZ period (4 ms when CONFIG_HZ=250).
At least
Till,
On Saturday 04 November 2006 10:24 am, Till Harbaum wrote:
My device does all its communication via its control endpoint. Is
this already a bad idea?
Yes, it is. I have to maintain a driver which is doing control transfers
during normal operation, and it is a major cause of problems:
-
Hello,
On Freitag, 2. Februar 2007, Till Harbaum / Lists wrote:
i have a similar card. I somehow figured out that this card in fact
has 1k sectors, not 512 bytes
No. The card has 512 bytes per sector (this is fixed for all SD cards),
but the CAPACITY of the card is reported in 1K sectors.
Tobias,
On Dienstag, 27. Februar 2007, Tobias Terrahe wrote:
I am using Suse 10.0 with Kernel 2.6.13-15.13-default. I have a new
Cam Labtec webcam pro and downloaded the driver spca5xx from
http://mxhaard.free.fr/download.html.
If I try to install the driver I alway get the error (as root):
Pete,
On Dienstag, 24. Juli 2007, Pete Zaitcev wrote:
Please keep in mind that all this discussion about short reads has
nothing to do with your real problem (e.g. the device which used to
reply with a stall stopped to do so, or it continues to reply with a
stall token, but EHCI fails to
Hi Jason,
On Samstag, 4. August 2007, Jason LeBrun wrote:
I'm using a Kingston 2.0GB SD card with a unbranded card reader
(model number UCR-61). When I insert the card, it's detected as
device using 512-byte sector sizes
This is correct. All SD and SDHC cards are using 512 Byte sector size.
73 matches
Mail list logo