From: Pavel Roskin <proski@gnu.org>
To: Johannes Berg <johannes@sipsolutions.net>
Cc: linux-wireless@vger.kernel.org
Subject: Re: [PATCH 18/35] Only cancel correct timeouts for Auth and Assoc replies
Date: Sat, 01 Sep 2007 06:54:51 -0400 [thread overview]
Message-ID: <1188644091.4146.12.camel@gx> (raw)
In-Reply-To: <1188638228.7585.124.camel@johannes.berg>
On Sat, 2007-09-01 at 11:17 +0200, Johannes Berg wrote:
> On Sat, 2007-09-01 at 00:35 -0400, Pavel Roskin wrote:
> > Don't cancel all delayed works indiscriminately. In case of
> > at76_rx_mgmt_auth(), cancel the timeout regardless of the status of the
> > reply, as long as the reply is for us.
>
> Out of curiosity, why are you cleaning up code that needs to be deleted
> anyhow? The way I see it, the driver ought to be be ported to mac80211,
> no? It seems to ship with an in-driver softmac.
I understand that. But at the time I did it, I wasn't even sure that
the mac80211 conversion is possible. I wanted to have a solid base
before starting the risky conversion.
The driver used to have a big event handler that was split into several
works and a big timer handler that was split into several timers. In
some cases, the code would cancel the "big timer" for some errors. The
first step was replace it with the code to cancel all timers. The
second step was to remove some unlikely or wrong cases, when the
specific timer cannot be running or shouldn't be canceled. I'm not
going to do any more cleanup of this kind.
What happened is that the driver in wireless-dev/at76 is a version 57
commits below the top of my repository, and the mac80211 port was
started at the point 35 commits above the version in wireless-dev/at76.
It would have been easier if everybody just had started with my current
code, but since it didn't happen, I'm trying to fix it. The first step
is to go the the point where Kalle started mac80211 conversion. The
rest doesn't contain any major fixes, although some changes are pretty
intrusive. I'll see what to do with them once I have a chance to play
with the mac80211 conversion patch.
The patches I have sent were not meant to be reviewed in this list,
since I expected the latest version to be imported as is. But I think
it's still better to send the patches separately rather than a single
patch. If nothing else, the separate patches could receive some useful
comments.
--
Regards,
Pavel Roskin
next prev parent reply other threads:[~2007-09-01 10:54 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-09-01 4:34 [PATCH 00/35] Update at76_usb to the start of mac80211 port Pavel Roskin
2007-09-01 4:34 ` [PATCH 01/35] Protect at76_get_op_mode() and at76_get_mib() against short reads Pavel Roskin
2007-09-01 4:34 ` [PATCH 02/35] Use existing macros to find bulk in and bulk out endpoints Pavel Roskin
2007-09-01 4:34 ` [PATCH 03/35] Rewrite at76_alloc_urbs() in a more linear fashion Pavel Roskin
2007-09-01 4:34 ` [PATCH 04/35] Avoid overuse of NULL Pavel Roskin
2007-09-01 4:34 ` [PATCH 05/35] Add myself to the author list Pavel Roskin
2007-09-01 4:34 ` [PATCH 06/35] Move (de)initialization functions closer to the end of file Pavel Roskin
2007-09-01 4:34 ` [PATCH 07/35] Don't use shift on numeric constants in usb_control_msg() arguments Pavel Roskin
2007-09-01 4:35 ` [PATCH 08/35] Merge at76_download_external_fw() into at76_load_external_fw() Pavel Roskin
2007-09-01 4:35 ` [PATCH 09/35] Simplify at76_usbdfu_download() Pavel Roskin
2007-09-01 4:35 ` [PATCH 10/35] Add minimal support to 505AMX Pavel Roskin
2007-09-01 4:35 ` [PATCH 11/35] Simplify logic in at76_get_reg_domain() Pavel Roskin
2007-09-01 4:35 ` [PATCH 12/35] Fix hex2str() and mac2str() to avoid buffer overlap Pavel Roskin
2007-09-01 4:35 ` [PATCH 13/35] Rename some long functions and fields Pavel Roskin
2007-09-01 4:35 ` [PATCH 14/35] Fix incorrect queue management in at76_tx_mgmt() Pavel Roskin
2007-09-01 4:35 ` [PATCH 15/35] Introduce at76_quiesce(), use it to stop network activity Pavel Roskin
2007-09-01 4:35 ` [PATCH 16/35] Don't disable and enable tasklets, it doesn't work as expected Pavel Roskin
2007-09-01 4:35 ` [PATCH 17/35] Start beacon timeout task when connected Pavel Roskin
2007-09-01 4:35 ` [PATCH 18/35] Only cancel correct timeouts for Auth and Assoc replies Pavel Roskin
2007-09-01 9:17 ` Johannes Berg
2007-09-01 10:54 ` Pavel Roskin [this message]
2007-09-01 11:04 ` Johannes Berg
2007-09-01 13:16 ` John W. Linville
2007-09-01 20:09 ` Pavel Roskin
2007-09-01 4:36 ` [PATCH 19/35] Improve output of the regdomain id Pavel Roskin
2007-09-01 4:36 ` [PATCH 20/35] Protect at76_iw_handler_set_scan() with mutex Pavel Roskin
2007-09-01 4:36 ` [PATCH 21/35] Eliminate vendor IDs Pavel Roskin
2007-09-01 4:36 ` [PATCH 22/35] Only retry resubmitting rx_urb once Pavel Roskin
2007-09-01 4:36 ` [PATCH 23/35] Simplify at76_get_mib_mdomain() Pavel Roskin
2007-09-01 4:36 ` [PATCH 24/35] Do implicit scanning only with current ESSID Pavel Roskin
2007-09-01 4:36 ` [PATCH 25/35] Don't dump mib_mdomain while scanning, it's done on device startup Pavel Roskin
2007-09-01 4:36 ` [PATCH 26/35] Improve dump of MAC_ADDR Pavel Roskin
2007-09-01 4:36 ` [PATCH 27/35] Remove unneeded braces, found by checkpatch.pl Pavel Roskin
2007-09-01 4:36 ` [PATCH 28/35] Convert dbg() to at76_dbg() or remove it Pavel Roskin
2007-09-01 4:37 ` [PATCH 29/35] Eliminate at76_dbg_dumpbuf() in favor of hex2str() Pavel Roskin
2007-09-01 4:37 ` [PATCH 30/35] Eliminate pr_debug() in favor of at76_dbg() Pavel Roskin
2007-09-01 4:37 ` [PATCH 31/35] Simplify logic in at76_is_hidden_ssid() Pavel Roskin
2007-09-01 4:37 ` [PATCH 32/35] Massive cleanup of dump functions Pavel Roskin
2007-09-01 4:37 ` [PATCH 33/35] Remove international roaming support Pavel Roskin
2007-09-01 4:37 ` [PATCH 34/35] Don't do additional MIB dumps if DEBUG is defined Pavel Roskin
2007-09-01 4:37 ` [PATCH 35/35] Replace scan_runs with scan_need_any Pavel Roskin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1188644091.4146.12.camel@gx \
--to=proski@gnu.org \
--cc=johannes@sipsolutions.net \
--cc=linux-wireless@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).