gsmd patches

andrzej zaborowski balrogg at gmail.com
Wed Sep 19 16:52:45 CEST 2007


Hi there,

On 17/09/2007, Jim Huang <jserv at openmoko.org> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Jim Huang 提到:
> >> On Thu, Aug 02, 2007 at 01:49:10PM +0800, Jim Huang wrote:
> >> The patch-set is put on here:
> >>      http://people.openmoko.org/jserv/gsm/patches/
>
> hi list,
>

I took the current gsmd SVN and applied these patches plus whatever I
had in my local tree and from there I tried to get it to work reliably
on my Neo. I got a version that seems to work reliably and I don't see
any kind of timing issues anymore (at least not in gsmd - in the
kernel, yes). My comments:

> Here are new patches:
> 024_sms-text-in-bracket.patch

Works.

> 025_sms-status-report.patch

Works, although I think a libgsmd client who's sending an SMS should
have a choice whether it wants a status report or not. And I don't
like using "len - 1" as an index, it may be unsafe.

> 026_multiline-fixlet.patch

I skipped this patch. What is it for? Looking at the code it looks
like it would break +CMGL and other responses. Without it everything
is fine.

> 027_phonebook-find-and-read-range-support.patch
> 028_shell-phonebook-find-and-read-range-support.patch

Applied but not tested. Note that it uses malloc instead of talloc in one place.

> 029_avoid-gsmd-stuck.patch
> 030_atcmd-debugging.patch

I skipped these two patches, instead I use the patch from
http://bugzilla.openmoko.org/cgi-bin/bugzilla/attachment.cgi?id=239&action=view
which already takes care of partial reads/writes. I rediffed this
patch on top of 024-028 and attached the current version to this
email. In addition I have two more changes in my tree:

0002-Flush... makes gsmd discard all of the pending commands in queue
when it receives the "AT-Command Interpreter ready" line and restart
the initialisation from a clean state.

0003-Correctly... fixes the reading from usock in gsmd, it fixes bug
#766 for me.
Note that there are some more places in gsmd and libgsmd where read()
or write() calls depend on timing. I will try to find all of them and
send a patch to cover this.

BTW why is the initial alive timer removed from current SVN? I think
it was a good thing. Should I send a patch to re-add it?
Regards
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Introduce-ports.patch
Type: text/x-patch
Size: 18289 bytes
Desc: not available
Url : http://lists.openmoko.org/pipermail/gsmd-devel/attachments/20070919/6dddf66f/0001-Introduce-ports-0001.bin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0002-Flush-all-pending-commands-before-restarting-the-mod.patch
Type: text/x-patch
Size: 2268 bytes
Desc: not available
Url : http://lists.openmoko.org/pipermail/gsmd-devel/attachments/20070919/6dddf66f/0002-Flush-all-pending-commands-before-restarting-the-mod-0001.bin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0003-Correctly-segment-incoming-usock-data-into-packets.patch
Type: text/x-patch
Size: 2633 bytes
Desc: not available
Url : http://lists.openmoko.org/pipermail/gsmd-devel/attachments/20070919/6dddf66f/0003-Correctly-segment-incoming-usock-data-into-packets-0001.bin


More information about the gsmd-devel mailing list