From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S272921AbTHFW1l (ORCPT ); Wed, 6 Aug 2003 18:27:41 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S272923AbTHFW1l (ORCPT ); Wed, 6 Aug 2003 18:27:41 -0400 Received: from caramon.arm.linux.org.uk ([212.18.232.186]:51468 "EHLO caramon.arm.linux.org.uk") by vger.kernel.org with ESMTP id S272921AbTHFW1i (ORCPT ); Wed, 6 Aug 2003 18:27:38 -0400 Date: Wed, 6 Aug 2003 23:27:35 +0100 From: Russell King To: Linux Kernel List , akpm@osdl.org Subject: Cardbus suspend/resume of 3ccfe575bt card fails Message-ID: <20030806232735.I16116@flint.arm.linux.org.uk> Mail-Followup-To: Linux Kernel List , akpm@osdl.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i X-Message-Flag: Your copy of Microsoft Outlook is vulnerable to viruses. See www.mutt.org for more details. Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi akpm, Yes, I should've reported this sooner. 2.4 and 2.6 kernels exhibit the same problem - upon resume, the card indicates that it has a 10base link to the hub, but the hub indicates it has a 100base link. No traffic is possible. Unplugging and replugging the card restores the link correctly (100base). The following is from 2.6.0-test2: 0000:05:00.0: 3Com PCI 3CCFE575BT Cyclone CardBus at 0x1800. Vers LK1.1.19 Before suspend, mii-diag reports: Basic registers of MII PHY #0: 3000 282d 0300 e54b 00a1 45e1 0001 0000. The autonegotiated capability is 00a0. The autonegotiated media type is 100baseTx. Basic mode control register 0x3000: Auto-negotiation enabled. You have link beat, and everything is working OK. Your link partner advertised 45e1: Flow-control 100baseTx-FD 100baseTx 10baseT-FD 10baseT, w/ 802.3X flow control. End of basic transceiver information. After resume, it reports: Basic registers of MII PHY #0: ffff ffff ffff ffff ffff ffff ffff ffff. No MII transceiver present!. I suspect this may be caused by the device not being restored correctly: 05:00.0 Class 0200: 10b7:5157 (rev 01) Subsystem: 10b7:5b57 Control: I/O- Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- Status: Cap+ 66Mhz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- SERR-