[Home]

Summary:DAHLIN-00021: [patch] Ease Ident wheel usage for users
Reporter:Kinsey Moore (opticron)Labels:
Date Opened:2008-07-15 08:40:20Date Closed:2009-03-11 09:51:07
Priority:MajorRegression?No
Status:Closed/CompleteComponents:wct4xxp
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:( 0) dahdi.base.c.diff-guidelines
Description:This patch allows non-strict ident wheel usage.  The first card doesn't have to be zero and skips in the ordering are allowed.  I've seen customers and more experienced people scratch their heads for hours wondering why a card won't load.

****** ADDITIONAL INFORMATION ******

Patches for dahdi and zaptel 1.2 should be up soon as well, though I'm not sure it will be accepdted into 1.2.
Comments:By: Kinsey Moore (opticron) 2008-07-15 08:51:24

hold off, i broke my patch, it's missing a section (use zap14.base.c.diff-fixed instead)


By: Kinsey Moore (opticron) 2008-07-15 15:25:22

zaptel 1.4 svn and dahdi trunk rev 4590 patches are now available, the dahdi patch had to be modified slightly because the temporary variable was being used in intervening code
By: Tzafrir Cohen (tzafrir) 2008-07-27 04:43:46

First off, thanks for your patch.

Moving to DAHDI, as I'm not sure new features should get into Zaptel.

I'm not exactly sure what the patch is about (what is this wheel?). But sorting the list of cards in a single pass does not look right. Think of 3 cards with reverse order.

I also haven't tested that the patch applies. I suspect it will need slight adjustments due to the minor changes in the printk lines.
By: Kinsey Moore (opticron) 2008-08-07 09:41:10

I'd probably agree with your statement about zaptel.  The sort is only a slice.  It gets run once for each probe ensuring that all cards are sorted properly upon insertion into the list.

To clarify, the ident wheel is the little hex wheel on Digium's dual and quad span T1 cards that identifies what order they load in so as to prevent random pci probe order from disrupting configs.


By: Tzafrir Cohen (tzafrir) 2008-10-15 13:13:57

Looks useful. But I don't know those cards well enough.

A small matter of style: please give priority to printk messages. I think that both of them should be KERN_NOTICE.

Also, please try to keep lines at most at 80 columns.
By: Kinsey Moore (opticron) 2008-10-16 08:09:41

I can't seem to delete the old patches.  I fixed the printk calls and knocked the line lengths down some.  There are two there I can't really do anything about.  The new patch is dahdi.base.c.diff-guidelines and is based on r5096.
By: Leif Madsen (lmadsen) 2009-02-02 14:30:00.000-0600

I've reassigned this to sruffell in the hopes he can decide how to move this issue forward. Thanks!
By: Digium Subversion (svnbot) 2009-03-11 09:51:05

Repository: dahdi
Revision: 6125

U   linux/trunk/drivers/dahdi/wct4xxp/base.c

------------------------------------------------------------------------
r6125 | sruffell | 2009-03-11 09:51:05 -0500 (Wed, 11 Mar 2009) | 9 lines

Relax ident wheel requirements.

Do not require the first card to be set to 0 and allow skips in the ident wheel
numbers.  The ident wheel allows a user to determine the order that cards
register there spans with DAHDI.

(closes issue DAHLIN-21)
Reported by: opticron
Patch by: opticron
------------------------------------------------------------------------

http://svn.digium.com/view/dahdi?view=rev&revision=6125