rt2x00: remove USB-ID 148f:2573 from rt2500usb driver
PATCH 1/1 rt2x00: remove USB-ID 148f:2573 from rt2500usb driver
The drivers rt2500usb and rt73usb both contain the usb id 148f:2573 while only the latter is suitable for devices with that id.
As a consequence, both drivers will be loaded and system log shows messages like "rt2500usb_init_eeprom: Error - Invalid RT chipset detected".
If you use the above search term, you will find more than 1500 hits with a well known search engine, confirming that rt2500usb is always the wrong driver for devices with usb id 148f:2573 and quite often as a side effect confusion arose about this -non fatal but misleading- error message in the respective bug reports or threads.
This trivial patch removes the id 148f:2573 from rt2500usb driver.
---
diff -Naur a/drivers/net/wireless/rt2x00/rt2500usb.c b/drivers/net/wireless/rt2x00/rt2500usb.c
--- a/drivers/net/wireless/rt2x00/rt2500usb.c 2010-05-03 16:03:29.572101896
+0200
+++ b/drivers/net/wireless/rt2x00/rt2500usb.c 2010-05-03 16:03:51.902854565
+0200
@@ -1879,7 +1879,6 @@
/* Ralink */
{ USB_DEVICE(0x148f, 0x1706), USB_DEVICE_DATA(&rt2500usb_ops) }, { USB_DEVICE(0x148f, 0x2570), USB_DEVICE_DATA(&rt2500usb_ops) },
- { USB_DEVICE(0x148f, 0x2573), USB_DEVICE_DATA(&rt2500usb_ops) },
{ USB_DEVICE(0x148f, 0x9020), USB_DEVICE_DATA(&rt2500usb_ops) }, /* Sagem */
{ USB_DEVICE(0x079b, 0x004b), USB_DEVICE_DATA(&rt2500usb_ops) },
Quoted 11 lines · On Mon, May 03, 2010 at 04:24:45PM +0200, [email protected] wrote:
The drivers rt2500usb and rt73usb both contain the usb id 148f:2573 while only the latter is suitable for devices with that id.
As a consequence, both drivers will be loaded and system log shows messages like "
rt2500usb_init_eeprom: Error - Invalid RT chipset detected".If you use the above search term, you will find more than 1500 hits with a well known search engine, confirming that rt2500usb is always the wrong driver for devices with usb id 148f:2573 and quite often as a side effect confusion arose about this -non fatal but misleading- error message in the respective bug reports or threads.
I also get more than 100 hits from (probably the same) well-known search engine for "rt73usb_init_eeprom: Error - Invalid RT chipset detected". I'm sure those aren't all for the same USB ID, and perhaps none of them are. Then again, not all of the 1500+ that you cite are for the same USB ID either.
So I'm curious, how can you be sure that 148f:2573 is _always_ wrong for rt2500usb? I do acknowledge that the "73" part makes it suspicious...
John
Signature · 2 lines
John W. Linville Someday the world will need a hero, and you [email protected] might be all we have. Be ready.
Ivo Van Doorn
NAKQuoted 13 lines · On 5/3/10, [email protected] <[email protected]> wrote:
The drivers rt2500usb and rt73usb both contain the usb id 148f:2573 while only
the latter is suitable for devices with that id.As a consequence, both drivers will be loaded and system log shows messages like "
rt2500usb_init_eeprom: Error - Invalid RT chipset detected".If you use the above search term, you will find more than 1500 hits with a well known search engine, confirming that rt2500usb is always the wrong driver for devices with usb id 148f:2573 and quite often as a side effect confusion arose about this -non fatal but misleading- error message in the respective bug reports or threads.
Quoted 3 lines
This trivial patch removes the id 148f:2573 from rt2500usb driver.
Signed-off-by Axel Koellhofer <[email protected]>
NACK
The ID is confusing indeed, but unfortunately we can't remove this ID, since in the early days people did report they had to use the rt2500usb driver for these devices. So the majority of devices with this ID is indeed rt73usb, but
not for all.
Ivo
Am Monday 03 May 2010 16:58:16 schrieben Sie:
Quoted 5 lines
I also get more than 100 hits from (probably the same) well-known search engine for "
rt73usb_init_eeprom: Error - Invalid RT chipset detected". I'm sure those aren't all for the same USB ID, and perhaps none of them are. Then again, not all of the 1500+ that you cite are for the same USB ID either.
Well, of course I can't be 100% sure, I had this patch "in storage" for a long time and already applied it so some packages I build, but hesitated for quite a while if I should submit it.
Additionally, the search string was not too good, it would have been better to cite this one here
148f:2573 rt2500usb_init_eeprom: Error
(about 500 hits)
Quoted 4 lines
So I'm curious, how can you be sure that 148f:2573 is _always_ wrong for rt2500usb? I do acknowledge that the "73" part makes it suspicious...
As I am active in several fora and read quite a few more, I stumbled over this annoyance (as I said it is not a fatal error) for several times in the past few months/years and in all cases I know rt73usb was the working driver.
But of course the problem in such cases is, you can't prove it, you can only add more and more examples for "not being wrong" while finding one example where actually a device with 148f:2573 working with rt2500usb und not working with rt73usb would prove me wrong once and for all.
Best wishes,
Axel K
Quoted 6 lines · Ivo Van Doorn
Quoted 1 line
This trivial patch removes the id 148f:2573 from rt2500usb driver.
The ID is confusing indeed, but unfortunately we can't remove this ID, since in the early days people did report they had to use the rt2500usb driver for these devices. So the majority of devices with this ID is indeed rt73usb, but not for all.
Since this seems to come up every couple of weeks, and none of us will always be around to remember and answer it, how about capturing that in a comment in the code? :)
johannes
Quoted 9 lines · Johannes Berg
Quoted 7 lines · Ivo Van Doorn
Quoted 1 line
This trivial patch removes the id 148f:2573 from rt2500usb driver.
The ID is confusing indeed, but unfortunately we can't remove this ID, since in the early days people did report they had to use the rt2500usb driver
for these devices. So the majority of devices with this ID is indeed rt73usb, but not for all.
Quoted 3 lines
Since this seems to come up every couple of weeks, and none of us will always be around to remember and answer it, how about capturing that in a comment in the code? :)
Good point. :)
I;ll send a patch later today.
Ivo
Quoted 9 lines · On Mon, May 03, 2010 at 05:20:36PM +0200, [email protected] wrote:
Am Monday 03 May 2010 16:58:16 schrieben Sie:
Quoted 2 lines
So I'm curious, how can you be sure that 148f:2573 is _always_ wrong for rt2500usb? I do acknowledge that the "73" part makes it suspicious...
But of course the problem in such cases is, you can't prove it, you can only add more and more examples for "not being wrong" while finding one example where actually a device with 148f:2573 working with rt2500usb und not working with rt73usb would prove me wrong once and for all.
http://osdir.com/ml/linux-wireless/2009-07/msg00267.html
This seems to be such an example?
John
Signature · 2 lines
John W. Linville Someday the world will need a hero, and you [email protected] might be all we have. Be ready.
Am Monday 03 May 2010 17:47:46 schrieben Sie:
Quoted 4 lines
http://osdir.com/ml/linux-wireless/2009-07/msg00267.html
This seems to be such an example?
Yeah, I also found that one before, but the message about missing firmware (which is the most common user's error with rt73usb) and lack of feedback from the person opening the thread if rt2500usb got the device working made it "inconclusive" to me.
But maybe this is the main reason one finds so many threads about problems with this device id and rt73usb being the correct driver.
If it works out of the box (no matter which driver), nobody starts a thread, if it doesn't work, in most cases it's missing firmware and if rt2500usb works instead of rt73usb, you can't miss that as it does not need installation of extra firmware.
I CCed Ivo instead of first sending him this mail only personally, because although I knew he would be the most likely person to know if my idea was good or bad (which was exactly what happened), I expected that if there really are a few devices with 148f:2573 needing rt2500usb instead of rt73usb, the more likely somebody else actually will know about such a (rare) exception.
Maybe that was not too clever, sorry for this rather useless thread.
Greetings,
Axel K