netdev - [PATCH 0/1] alx driver

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <1370899609-13954-1-git-send-email-johannes@sipsolutions.net>
Date:	Mon, 10 Jun 2013 23:26:48 +0200
From:	Johannes Berg <johannes@...solutions.net>
To:	netdev@...r.kernel.org
Cc:	mcgrof@...not-panic.com, kvalo@...rom.com, adrian.chadd@...il.com
Subject: [PATCH 0/1] alx driver

So ... I was stupid enough to not research drivers for this device
before buying a board that included one, I figured in 2013 it wasn't
really possible for such a simple gigabit ethernet device to not
have a driver in Linux.

Boy was I wrong.

The vendor driver was previously posted on netdev, and subsequently
somewhat cleaned up, but ... it still had useless hand-rolled locks
(with test_and_set_bit/msleep!), horrible macros, and was generally
quite a mess.

I do have to admit that I'm not very familiar with all the netdev
APIs, since wireless devices are so slow that NAPI etc. hardly
matters. In a way, this is also me learning about it since we'll be
betting faster wireless too :-)

I hope I did an OK job at cleaning it up/rewriting it, but while I
thought about the locking for a while, or rather the lack thereof,
I'm not entirely sure that it's completely safe. It seemed to me
that since we have TX/RX all done from NAPI it won't actually have
to be reentrant/locked.

Also, I kept the FIELD macros, which seemed wrong at first but the
code didn't get nicer w/o them. I did fix them for endianness and
converted some to inlines; I don't get why they actually had endian
annotations and even turned on the sparse checks by default only to
then ignore all the errors it was giving. I'll happily kill these
if anyone thinks that would be better.

Anyway, as I said, I'm not very good at ethernet drivers, so let me
know what I need to fix.

Maybe the responsible QCA folks can actually add back the features
that I liberally pruned from the driver.

johannes
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html