linux-kernel - RE: [PATCH 25/26] [REVISED] atl1: add NAPI support

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

Date:	Wed, 2 Jan 2008 11:56:36 +0900
From:	"Joonwoo Park" <joonwpark81@...il.com>
To:	"'Jay Cliburn'" <jacliburn@...lsouth.net>
Cc:	<jeff@...zik.org>, <csnook@...hat.com>,
	<linux-kernel@...r.kernel.org>, <atl1-devel@...ts.sourceforge.net>,
	<netdev@...r.kernel.org>, <shemminger@...ux-foundation.org>
Subject: RE: [PATCH 25/26] [REVISED] atl1: add NAPI support

Hi Jay,

+	if ((work_done < budget) || !netif_running(poll_dev)) {
+quit_polling:
+		netif_rx_complete(poll_dev, napi);
+
+		if (!test_bit(__ATL1_DOWN, &adapter->flags))
+			atlx_irq_enable(adapter);
+	}

Not enough :)
If netif_running() is false, it can make problem.
The problem occurs calling netif_rx_complete with work_done == budget.
If do that, net_rx_action would do poll list double deletion.

Since we had reached a consensus on fixing it without each drivers modifications, there is no best solution for that problem for
now. I'm expecting Dave or others work for net-core.
(http://lkml.org/lkml/2007/12/20/600)

IMHO, atl1_clean should wait for work_done != budget even though netif_running is false at this time.
At least, It would not make oops.

Thanks,
Joonwoo

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/