Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932766AbWHOM3R (ORCPT ); Tue, 15 Aug 2006 08:29:17 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932765AbWHOM3R (ORCPT ); Tue, 15 Aug 2006 08:29:17 -0400 Received: from outgoing3.smtp.agnat.pl ([193.239.44.85]:56535 "EHLO outgoing3.smtp.agnat.pl") by vger.kernel.org with ESMTP id S932756AbWHOM3Q convert rfc822-to-8bit (ORCPT ); Tue, 15 Aug 2006 08:29:16 -0400 From: Arkadiusz Miskiewicz Organization: SelfOrganizing To: linux-scsi@vger.kernel.org Subject: Re: qlogic 2312 problems on 2.6.16.22, 2.6.18rc4 Date: Tue, 15 Aug 2006 14:29:09 +0200 User-Agent: KMail/1.9.4 Cc: linux-kernel@vger.kernel.org References: <200608140946.50411.arekm@pld-linux.org> <200608141437.05269.arekm@pld-linux.org> In-Reply-To: <200608141437.05269.arekm@pld-linux.org> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-2" Content-Transfer-Encoding: 8BIT Content-Disposition: inline Message-Id: <200608151429.09082.arekm@pld-linux.org> X-Authenticated-Id: arekm Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 6471 Lines: 152 On Monday 14 August 2006 14:37, Arkadiusz Miskiewicz wrote: > On Monday 14 August 2006 09:46, Arkadiusz Miskiewicz wrote: > > Hi, > > > > I was using QLA2312 FC card on 32-bit machine with 6GB ram > > without problems. Recently I've switched to opteron dual core machine > > also with 6GB ram and I'm having serious problem with access to FC array. > > > > When I switch back to 32-bit machine the problem disappears. Some qla2312 > > problems with 64bit machines? > > I've tested latest git (same as 2.6.18rc4 I guess) using latest ql2300 > firmware from qlogic site. I have simply no idea what's going here :-( Everything works fine on 32-bit dual xeon machine with card inserted into PCI slot. Here in new dual core opteron machine the card sits in PCI-X slot. It's Thunder K8SRE S2891 mainboard, Transport GT24 B2891 1U barebone, Tyan M2075 riser card, 2 x Opteron 270, 6GB ram. Note that booting with 32-bit kernel (which works fine on Xeon system) doesn't cure the problem on Opteron system. Booting 64bit 2.6.18rc4 kernel with mem=3G also doesn't fix anything. /t is on tmpfs, /dev/sda2 in on FC array. Reading the same data several times and I get different md5sum results each time, see below. How I can track where corruption occurs? [root@pldmachine /t]# dd if=/dev/sda2 of=test-from-sda2-2 bs=1M count=1024 1024+0 records in 1024+0 records out [root@pldmachine /t]# md5sum test-from-sda2-2 90b831662a5d3bd2501028eee3c00264 test-from-sda2-2 [root@pldmachine /t]# rm test-from-sda2-2 [root@pldmachine /t]# dd if=/dev/sda2 of=test-from-sda2-2 bs=1M count=1024 1024+0 records in 1024+0 records out [root@pldmachine /t]# md5sum test-from-sda2-2 c06dadfc0726aa07cccdfbe7804e646f test-from-sda2-2 [root@pldmachine /t]# rm test-from-sda2-2 [root@pldmachine /t]# dd if=/dev/sda2 of=test-from-sda2-2 bs=1M count=1024 1024+0 records in 1024+0 records out [root@pldmachine /t]# md5sum test-from-sda2-2 06d61f81245bd242e9420277cfca65e4 test-from-sda2-2 dd'ing over tmpfs only is fine of course: [root@pldmachine /t]# md5sum tmpfs-test ddaaa7eb8da403f6079c55b1391ee1e3 tmpfs-test [root@pldmachine /t]# rm tmpfs-test [root@pldmachine /t]# dd if=test of=tmpfs-test bs=1M count=1024 1024+0 records in 1024+0 records out [root@pldmachine /t]# md5sum tmpfs-test ddaaa7eb8da403f6079c55b1391ee1e3 tmpfs-test [root@pldmachine /t]# rm tmpfs-test [root@pldmachine /t]# dd if=test of=tmpfs-test bs=1M count=1024 1024+0 records in 1024+0 records out [root@pldmachine /t]# md5sum tmpfs-test ddaaa7eb8da403f6079c55b1391ee1e3 tmpfs-test QLogic Fibre Channel HBA Driver PCI: Enabling device 0000:09:08.0 (0150 -> 0153) ACPI: PCI Interrupt 0000:09:08.0[A] -> GSI 24 (level, low) -> IRQ 209 qla2xxx 0000:09:08.0: Found an ISP2312, irq 209, iobase 0xffffc20000004000 qla2xxx 0000:09:08.0: Configuring PCI space... qla2xxx 0000:09:08.0: Configure NVRAM parameters... qla2xxx 0000:09:08.0: Verifying loaded RISC code... scsi(1): **** Load RISC code **** scsi(1): Verifying Checksum of loaded RISC code. scsi(1): Checksum OK, start firmware. qla2xxx 0000:09:08.0: Allocated (412 KB) for firmware dump... scsi(1): Issue init firmware. scsi(1): Asynchronous LIP RESET (f7f7). qla2xxx 0000:09:08.0: LIP reset occured (f7f7). qla2xxx 0000:09:08.0: Waiting for LIP to complete... scsi(1): Asynchronous P2P MODE received. scsi(1): Asynchronous LOOP UP (2 Gbps). qla2xxx 0000:09:08.0: LOOP UP detected (2 Gbps). scsi(1): Asynchronous PORT UPDATE. scsi(1): Port database changed ffff 0006 0000. scsi(1): F/W Ready - OK scsi(1): fw_state=3 curr time=10002c93c. qla2xxx 0000:09:08.0: Topology - (F_Port), Host Loop address 0xffff scsi(1): Configure loop -- dpc flags =0x4080040 scsi(1): RSCN queue entry[0] = [00/000000]. scsi(1): device_resync: rscn overflow. scsi(1): RFT_ID exiting normally. scsi(1): RFF_ID exiting normally. scsi(1): RNN_ID exiting normally. scsi(1): RSNN_NN exiting normally. scsi(1): GID_PT entry - nn 500508b300917e50 pn 500508b300917e51 portid=010000. scsi(1): GID_PT entry - nn 200000e08b8051d6 pn 210000e08b8051d6 portid=010100. scsi(1): GID_PT entry - nn 200000e08b950428 pn 210000e08b950428 portid=010200. scsi(1): device wrap (010200) scsi(1): Trying Fabric Login w/loop id 0x0081 for port 010000. scsi(1): Trying Fabric Login w/loop id 0x0082 for port 010100. scsi(1): LOOP READY DEBUG: detect hba 1 at address = ffff81017d9344b0 scsi1 : qla2xxx qla2xxx 0000:09:08.0: QLogic Fibre Channel HBA Driver: 8.01.05-k3-debug QLogic QLA2340 - ISP2312: PCI-X (133 MHz) @ 0000:09:08.0 hdma+, host#=1, fw=3.03.20 IPX Vendor: COMPAQ Model: MSA1000 Rev: 4.96 Type: RAID ANSI SCSI revision: 04 Vendor: COMPAQ Model: MSA1000 VOLUME Rev: 4.96 Type: Direct-Access ANSI SCSI revision: 04 SCSI device sda: 513790830 512-byte hdwr sectors (263061 MB) sda: Write Protect is off sda: Mode Sense: 83 00 00 08 SCSI device sda: drive cache: write back SCSI device sda: 513790830 512-byte hdwr sectors (263061 MB) sda: Write Protect is off sda: Mode Sense: 83 00 00 08 SCSI device sda: drive cache: write back sda: sda1 sda2 sda3 sd 1:0:0:2: Attached scsi disk sda Vendor: COMPAQ Model: MSA1000 VOLUME Rev: 4.96 Type: Direct-Access ANSI SCSI revision: 04 SCSI device sdb: 1381590 512-byte hdwr sectors (707 MB) sdb: Write Protect is off sdb: Mode Sense: 83 00 00 08 SCSI device sdb: drive cache: write back SCSI device sdb: 1381590 512-byte hdwr sectors (707 MB) sdb: Write Protect is off sdb: Mode Sense: 83 00 00 08 SCSI device sdb: drive cache: write back sdb: sdb1 sdb2 sdb: p2 exceeds device capacity sd 1:0:0:4: Attached scsi disk sdb scsi(1): Asynchronous PORT UPDATE ignored 0000/0006/be3e. scsi(1): Asynchronous PORT UPDATE ignored 0000/0007/be3e. scsi(1): Asynchronous PORT UPDATE ignored 0000/0004/be3e. btw. there is a "8131 Errata 56 PCLK" option in BIOS which is related to AMD-8131 Errata 56. This option is needed for ,, four-function PCI-X cards''. What's is four-function PCI-X card ? Card which is visible in lspci as 4 devices? (This option also doesn't fix the problem unfortunately). -- Arkadiusz Mi?kiewicz PLD/Linux Team arekm / maven.pl http://ftp.pld-linux.org/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/