Bug 121785 - eciadsl support on 2.6.x
Summary: eciadsl support on 2.6.x
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel   
(Show other bugs)
Version: rawhide
Hardware: All
OS: Linux
Target Milestone: ---
Assignee: Arjan van de Ven
QA Contact: Brian Brock
Keywords: FutureFeature
Depends On:
TreeView+ depends on / blocked
Reported: 2004-04-27 20:36 UTC by Andrea Santilli
Modified: 2007-11-30 22:10 UTC (History)
0 users

Fixed In Version:
Doc Type: Enhancement
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2004-06-15 00:50:24 UTC
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)

Description Andrea Santilli 2004-04-27 20:36:10 UTC
Description of problem:
I and many people use an usermode usb adsl modem driver for
ECI-chipset based modems. This driver needs kernel 2.6 to be patched (
http://eciadsl.flashtux.org/download/beta/2.6.x-usb.patch ) to work.
It would be a pity to force people who use that driver to patch and
recompile the kernel

Version-Release number of selected component (if applicable):
kernel >= 2.6.0

How reproducible:

Actual results:
Patch not included in distribution kernel rpm.

Expected results:
Patch included

Comment 1 Arjan van de Ven 2004-04-28 09:26:48 UTC
do you know why this patch hasn't been submitted for inclusion in the
kernel.org kernel ?

Comment 2 Andrea Santilli 2004-04-28 12:32:35 UTC
No sorry, I have no idea... :-(
Another problem is that the driver maintainer never hacked 2.6 code
and i can't reach the patch maintainer

Comment 3 Andrea Santilli 2004-04-28 21:41:20 UTC
There's a new:
an eciadsl driver developer proposed this patch:
that guy says that interval of iso urb doesn't get initialized

I have to say that both patches do not resolve a problem: if you
detach the modem before the driver starter exits, the comp gets locked.
It looks like a deadlock in interrupt system...

Comment 4 Andrea Santilli 2004-06-09 01:12:11 UTC
Hello! There are good news!
As you can see from http://eciadsl.flashtux.org/index.php?lang=en&id=171
the bug of linux 2.6 about ISO URB  has been resolved in 2.6.7-rc2
(havent tested it yet, I'm still downloading 2.6.7-rc3)...
Please, whatever they did to the kernel, backport it! :)

Comment 5 Arjan van de Ven 2004-06-09 10:02:09 UTC
please try the 424 kernel from http://people.redhat.com/arjanv/2.6/

Comment 6 Andrea Santilli 2004-06-09 14:17:04 UTC
Well done! The modem works great! Thank you very much!

But there is still something weird (apart the wrong disk geometry
guessing, but this is another bug)...

Well, I made a "service" that starts the modem at boot and I probably
made a mistake while giving it the stopping priority so the kernel
sends some "Badness" message... This doesnt happen if I stop the
driver before I reboot/halt the machine. I know there's something
wrong in the script I built but I guess this shouldn't happen anyway.
Look this happens _NOT_ only on your kernel but also on patched
<2.6.7-rc2 kernels (vanilla and not) and on unpatched >= 2.6.7-rc3
kernels. The only difference is that on your 424 kernel the Badness
weirdly isn't reported on /var/log/messages*.

This is the output:

Sending all processes the TERM signal... Badness in local_bh_enable at
[<0212407a>] local_bh_enable+0x39/0x5c
[<42f0499e>] ppp_sync_push+0xd2/0x149 [ppp_synctty]
[<42f044aa>] ppp_sync_wakeup+0x1d/0x35 [ppp_synctty]
[<021d549f>] do_tty_hangup+0x12d/0x34c
[<021d652d>] release_dev+0x1c0/0x53f
[<0222e096>] usb_unbind_interface+0x41/0x50
[<021f4a8d>] device_release_driver+0x3c/0x46
[<0222e27e>] usb_driver_release_interface+0x2c/0x40
[<02235b32>] releaseintf+0x76/0x8f
[<021d6bc6>] tty_release+0x29/0x4e
[<02152baf>] __fput+0x3f/0xe3
[<0215189c>] filp_close+0x59/0x5f
[<02122006>] put_files_struct+0x57/0xaa
[<02122b45>] do_exit+0x211/0x390
[<02122dc0>] sys_exit_group+0x0/0xd
[<0212986b>] get_signal_to_deliver+0x34c/0x372
[<02107208>] do_signal+0x4e/0xbb
[<02160dc3>] file_ioctl+0x167/0x17b
[<02161010>] sys_ioctl+0x239/0x243
[<0210729d>] do_notify_resume+0x28/0x37
                                                             [  OK  ]

Well that's all...
Anyway thanks a lot! :)

Comment 7 Dave Jones 2004-06-15 00:50:24 UTC
the ppp trace should now also be fixed. if not, please open a seperate
bug for that. thanks.

Note You need to log in before you can comment on or make changes to this bug.