Kugutsumen
2007-Oct-10 08:50 UTC
DMA critical problems under Xen dom0 (b66) with Areca 1110 SATA Raid Card
Hi, The original thread was posted in the device drivers and zfs discussion. http://www.opensolaris.org/jive/thread.jspa?messageID=162462𧪞 However the problem seems to only occur under the xen dom0: I boot the kernel without the hypervisor, the problem completely disappear. Upgrading the raid card firmware didn''t help, turning off write-back didn''t help (actually made it worse.) This is a 4 core amd 64 machine with 8 gig of ram. I remember there was an old discussion on xen users related to dma on machine with big memory configuration. Here is the description of the problem. Just as I create a ZFS pool and copy the root partition to it.... the performance seems to be really good then suddenly the system hangs all my sesssions and displays on the console: Oct 10 00:23:28 sunrise arcmsr: WARNING: arcmsr0: dma map got ''no resources'' Oct 10 00:23:28 sunrise arcmsr: WARNING: arcmsr0: dma allocate fail Oct 10 00:23:28 sunrise arcmsr: WARNING: arcmsr0: dma allocate fail free scsi hba AR1100 was running in JBOD mode and write-back caching under Solaris Express Xen Drop b66 07/07. Never had problem this setup under linux under heavy load. ARECA SATA-SAS RAID Host Adapter Driver(i386) 1.20.00.13,REV=2006.08.14 Updated to latest firmware 1.43-70417 ... same problem.. WARNING: arcmsr0: dma map got ''no resources'' WARNING: arcmsr0: dma allocate fail WARNING: arcmsr0: dma allocate fail free scsi hba pkt WARNING: arcmsr0: dma map got ''no resources'' WARNING: arcmsr0: dma allocate fail WARNING: a SunOS sunrise 5.11 xen-nv66-2007-06-24 i86pc i386 i86pc prtconf -pv System Configuration: Sun Microsystems i86pc Memory size: 8192 Megabytes System Peripherals (PROM Nodes): Node 0x000001 bios-boot-device: ''81'' stdout: 00000000 name: ''i86pc'' Node 0x000002 existing: 00d5d000.00000000.02368001.00000000 name: ''ramdisk'' Node 0x000003 bus-type: ''isa'' device_type: ''isa'' name: ''isa'' Node 0x000004 device_type: ''pci'' reg: 00000000.00000000.00000000 #size-cells: 00000002 #address-cells: 00000003 name: ''pci'' Node 0x000005 reg: 00000800.00000000.00000000.00000000.00000000 compatible: ''pci1166,36.0'' + ''pci1166,36'' + ''pciclass,060400'' + ''pciclass,0604'' model: ''PCI-PCI bridge'' ranges: 82000000.00000000.ff400000.82000000.00000000.ff400000.00000000.00100000.c2000000.00000000.f6300000.c2000000.00000000.f6300000.00000000.00800000 bus-range: 00000001.00000003 #size-cells: 00000002 #address-cells: 00000003 device_type: ''pci'' power-consumption: 00000001.00000001 devsel-speed: 00000000 unit-address: ''1'' class-code: 00060400 revision-id: 00000000 vendor-id: 00001166 device-id: 00000036 pcix-capid-pointer: 00000090 name: ''pci1166,36'' Node 0x000018 slot-names: 00000008.746f6c53.00000031 reg: 00016800.00000000.00000000.00000000.00000000 compatible: ''pci1166,104.c0'' + ''pci1166,104'' + ''pciclass,060400'' + ''pciclass,0604'' model: ''PCI-PCI bridge'' ranges: 82000000.00000000.ff400000.82000000.00000000.ff400000.00000000.00100000.c2000000.00000000.f6300000.c2000000.00000000.f6300000.00000000.00800000 bus-range: 00000002.00000003 #size-cells: 00000002 #address-cells: 00000003 device_type: ''pci'' power-consumption: 00000001.00000001 66mhz-capable: devsel-speed: 00000001 unit-address: ''d'' class-code: 00060400 revision-id: 000000c0 vendor-id: 00001166 device-id: 00000104 pcix-capid-pointer: 00000090 name: ''pci1166,104'' Node 0x000019 reg: 00021800.00000000.00000000.00000000.00000000 compatible: ''pci8086,335.a'' + ''pci8086,335'' + ''pciclass,060400'' + ''pciclass,0604'' model: ''PCI-PCI bridge'' ranges: 82000000.00000000.ff400000.82000000.00000000.ff400000.00000000.00100000.c2000000.00000000.f6300000.c2000000.00000000.f6300000.00000000.00800000 bus-range: 00000003.00000003 #size-cells: 00000002 #address-cells: 00000003 device_type: ''pci'' power-consumption: 00000001.00000001 66mhz-capable: fast-back-to-back: devsel-speed: 00000001 unit-address: ''3'' class-code: 00060400 revision-id: 0000000a vendor-id: 00008086 device-id: 00000335 pcix-capid-pointer: 000000f0 name: ''pci8086,335'' Node 0x00001a assigned-addresses: 82037010.00000000.ff4ff000.00000000.00001000.c2037018.00000000.f6400000.00000000.00400000 reg: 00037000.00000000.00000000.00000000.00000000.02037010.00000000.00000000.00000000.00001000.42037018.00000000.00000000.00000000.00400000 compatible: ''pci17d3,1110.17d3.1110.0'' + ''pci17d3,1110.17d3.1110'' + ''pci17d3,1110'' + ''pci17d3,1110.0'' + ''pci17d3,1110'' + ''pciclass,010400'' + ''pciclass,0104'' model: ''RAID controller'' power-consumption: 00000001.00000001 66mhz-capable: devsel-speed: 00000001 interrupts: 00000001 max-latency: 00000000 min-grant: 00000080 subsystem-vendor-id: 000017d3 subsystem-id: 00001110 unit-address: ''e'' class-code: 00010400 revision-id: 00000000 vendor-id: 000017d3 device-id: 00001110 pcix-capid-pointer: 000000e0 name: ''pci17d3,1110'' Node 0x000006 reg: 00001000.00000000.00000000.00000000.00000000 compatible: ''pci1166,205.1166.201.0'' + ''pci1166,205.1166.201'' + ''pci1166,201'' + ''pci1166,205.0'' + ''pci1166,205'' + ''pciclass,060000'' + ''pciclass,0600'' model: ''Host bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000001 max-latency: 00000000 min-grant: 00000000 subsystem-vendor-id: 00001166 subsystem-id: 00000201 unit-address: ''2'' class-code: 00060000 revision-id: 00000000 vendor-id: 00001166 device-id: 00000205 name: ''pci1166,201'' Node 0x000007 #size-cells: 00000000 #address-cells: 00000001 device_type: ''pci-ide'' assigned-addresses: 81001110.00000000.000001f0.00000000.00000008.81001114.00000000.000003f6.00000000.00000001.81001118.00000000.00000170.00000000.00000008.8100111c.00000000.00000376.00000000.00000001.81001120.00000000.0000ffa0.00000000.00000010 reg: 00001100.00000000.00000000.00000000.00000000.81001110.00000000.000001f0.00000000.00000008.81001114.00000000.000003f6.00000000.00000001.81001118.00000000.00000170.00000000.00000008.8100111c.00000000.00000376.00000000.00000001.01001120.00000000.00000000.00000000.00000010 compatible: ''pci1166,214.1166.214.0'' + ''pci1166,214.1166.214'' + ''pci1166,214'' + ''pci1166,214.0'' + ''pci1166,214'' + ''pciclass,01018a'' + ''pciclass,0101'' model: ''IDE controller'' power-consumption: 00000001.00000001 devsel-speed: 00000001 max-latency: 00000000 min-grant: 00000000 subsystem-vendor-id: 00001166 subsystem-id: 00000214 unit-address: ''2,1'' class-code: 0001018a revision-id: 00000000 vendor-id: 00001166 device-id: 00000214 name: ''pci-ide'' Node 0x000008 reg: 00000000 name: ''ide'' Node 0x000009 reg: 00000001 name: ''ide'' Node 0x00000a reg: 00001200.00000000.00000000.00000000.00000000 compatible: ''pci1166,234.1166.230.0'' + ''pci1166,234.1166.230'' + ''pci1166,230'' + ''pci1166,234.0'' + ''pci1166,234'' + ''pciclass,060100'' + ''pciclass,0601'' model: ''ISA bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000001 max-latency: 00000000 min-grant: 00000000 subsystem-vendor-id: 00001166 subsystem-id: 00000230 unit-address: ''2,2'' class-code: 00060100 revision-id: 00000000 vendor-id: 00001166 device-id: 00000234 name: ''pci1166,230'' Node 0x00000b assigned-addresses: 82001810.00000000.ff6b4000.00000000.00001000.81001814.00000000.0000e000.00000000.00000100 reg: 00001800.00000000.00000000.00000000.00000000.02001810.00000000.00000000.00000000.00001000.01001814.00000000.00000000.00000000.00000100 compatible: ''pci1166,223.1166.223.1'' + ''pci1166,223.1166.223'' + ''pci1166,223'' + ''pci1166,223.1'' + ''pci1166,223'' + ''pciclass,0c0310'' + ''pciclass,0c03'' model: ''Universal Serial Bus OHCI compliant'' power-consumption: 00000001.00000001 66mhz-capable: fast-back-to-back: devsel-speed: 00000001 interrupts: 00000001 max-latency: 00000000 min-grant: 00000000 subsystem-vendor-id: 00001166 subsystem-id: 00000223 unit-address: ''3'' class-code: 000c0310 revision-id: 00000001 vendor-id: 00001166 device-id: 00000223 name: ''pci1166,223'' Node 0x00000c assigned-addresses: 82001910.00000000.ff6b5000.00000000.00001000.81001914.00000000.0000e400.00000000.00000100 reg: 00001900.00000000.00000000.00000000.00000000.02001910.00000000.00000000.00000000.00001000.01001914.00000000.00000000.00000000.00000100 compatible: ''pci1166,223.1166.223.1'' + ''pci1166,223.1166.223'' + ''pci1166,223'' + ''pci1166,223.1'' + ''pci1166,223'' + ''pciclass,0c0310'' + ''pciclass,0c03'' model: ''Universal Serial Bus OHCI compliant'' power-consumption: 00000001.00000001 66mhz-capable: fast-back-to-back: devsel-speed: 00000001 interrupts: 00000001 max-latency: 00000000 min-grant: 00000000 subsystem-vendor-id: 00001166 subsystem-id: 00000223 unit-address: ''3,1'' class-code: 000c0310 revision-id: 00000001 vendor-id: 00001166 device-id: 00000223 name: ''pci1166,223'' Node 0x00000d assigned-addresses: 82001a10.00000000.ff6b6000.00000000.00001000.81001a14.00000000.0000e800.00000000.00000100 reg: 00001a00.00000000.00000000.00000000.00000000.02001a10.00000000.00000000.00000000.00001000.01001a14.00000000.00000000.00000000.00000100 compatible: ''pci1166,223.1166.223.1'' + ''pci1166,223.1166.223'' + ''pci1166,223'' + ''pci1166,223.1'' + ''pci1166,223'' + ''pciclass,0c0320'' + ''pciclass,0c03'' model: ''Universal Serial Bus EHCI compliant'' power-consumption: 00000001.00000001 66mhz-capable: fast-back-to-back: devsel-speed: 00000001 interrupts: 00000001 max-latency: 00000000 min-grant: 00000000 subsystem-vendor-id: 00001166 subsystem-id: 00000223 unit-address: ''3,2'' class-code: 000c0320 revision-id: 00000001 vendor-id: 00001166 device-id: 00000223 name: ''pci1166,223'' Node 0x00000e assigned-addresses: 82002010.00000000.ff680000.00000000.00020000.82002014.00000000.ff660000.00000000.00020000.81002018.00000000.0000dc00.00000000.00000040 reg: 00002000.00000000.00000000.00000000.00000000.02002010.00000000.00000000.00000000.00020000.02002014.00000000.00000000.00000000.00020000.01002018.00000000.00000000.00000000.00000040 compatible: ''pci8086,1076.10f1.3870.5'' + ''pci8086,1076.10f1.3870'' + ''pci10f1,3870'' + ''pci8086,1076.5'' + ''pci8086,1076'' + ''pciclass,020000'' + ''pciclass,0200'' model: ''Ethernet controller'' power-consumption: 00000001.00000001 66mhz-capable: devsel-speed: 00000001 interrupts: 00000001 max-latency: 00000000 min-grant: 000000ff subsystem-vendor-id: 000010f1 subsystem-id: 00003870 unit-address: ''4'' class-code: 00020000 revision-id: 00000005 vendor-id: 00008086 device-id: 00001076 pcix-capid-pointer: 000000e4 name: ''pci10f1,3870'' Node 0x00000f video-adapter-type: ''svga'' display-type: ''color'' assigned-addresses: c2003010.00000000.f8000000.00000000.04000000.82003014.00000000.ff6c0000.00000000.00040000.81003018.00000000.0000ec00.00000000.00000080.a1003000.00000000.000003b0.00000000.0000000c.a1003000.00000000.000003c0.00000000.00000020.82003000.00000000.000a0000.00000000.00020000 reg: 00003000.00000000.00000000.00000000.00000000.42003010.00000000.00000000.00000000.04000000.02003014.00000000.00000000.00000000.00040000.01003018.00000000.00000000.00000000.00000080.a1003000.00000000.000003b0.00000000.0000000c.a1003000.00000000.000003c0.00000000.00000020.82003000.00000000.000a0000.00000000.00020000 compatible: ''pci18ca,20.18ca.20.0'' + ''pci18ca,20.18ca.20'' + ''pci18ca,20'' + ''pci18ca,20.0'' + ''pci18ca,20'' + ''pciclass,030000'' + ''pciclass,0300'' model: ''VGA compatible controller'' power-consumption: 00000001.00000001 66mhz-capable: devsel-speed: 00000001 max-latency: 00000000 min-grant: 00000000 subsystem-vendor-id: 000018ca subsystem-id: 00000020 device_type: ''display'' unit-address: ''6'' class-code: 00030000 revision-id: 00000000 vendor-id: 000018ca device-id: 00000020 name: ''display'' Node 0x000010 reg: 0000c000.00000000.00000000.00000000.00000000 compatible: ''pci1022,1100.0'' + ''pci1022,1100'' + ''pciclass,060000'' + ''pciclass,0600'' model: ''Host bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000000 max-latency: 00000000 min-grant: 00000000 unit-address: ''18'' class-code: 00060000 revision-id: 00000000 vendor-id: 00001022 device-id: 00001100 name: ''pci1022,1100'' Node 0x000011 reg: 0000c100.00000000.00000000.00000000.00000000 compatible: ''pci1022,1101.0'' + ''pci1022,1101'' + ''pciclass,060000'' + ''pciclass,0600'' model: ''Host bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000000 max-latency: 00000000 min-grant: 00000000 unit-address: ''18,1'' class-code: 00060000 revision-id: 00000000 vendor-id: 00001022 device-id: 00001101 name: ''pci1022,1101'' Node 0x000012 reg: 0000c200.00000000.00000000.00000000.00000000 compatible: ''pci1022,1102.0'' + ''pci1022,1102'' + ''pciclass,060000'' + ''pciclass,0600'' model: ''Host bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000000 max-latency: 00000000 min-grant: 00000000 unit-address: ''18,2'' class-code: 00060000 revision-id: 00000000 vendor-id: 00001022 device-id: 00001102 name: ''pci1022,1102'' Node 0x000013 reg: 0000c300.00000000.00000000.00000000.00000000 compatible: ''pci1022,1103.0'' + ''pci1022,1103'' + ''pciclass,060000'' + ''pciclass,0600'' model: ''Host bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000000 max-latency: 00000000 min-grant: 00000000 unit-address: ''18,3'' class-code: 00060000 revision-id: 00000000 vendor-id: 00001022 device-id: 00001103 name: ''pci1022,1103'' Node 0x000014 reg: 0000c800.00000000.00000000.00000000.00000000 compatible: ''pci1022,1100.0'' + ''pci1022,1100'' + ''pciclass,060000'' + ''pciclass,0600'' model: ''Host bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000000 max-latency: 00000000 min-grant: 00000000 unit-address: ''19'' class-code: 00060000 revision-id: 00000000 vendor-id: 00001022 device-id: 00001100 name: ''pci1022,1100'' Node 0x000015 reg: 0000c900.00000000.00000000.00000000.00000000 compatible: ''pci1022,1101.0'' + ''pci1022,1101'' + ''pciclass,060000'' + ''pciclass,0600'' model: ''Host bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000000 max-latency: 00000000 min-grant: 00000000 unit-address: ''19,1'' class-code: 00060000 revision-id: 00000000 vendor-id: 00001022 device-id: 00001101 name: ''pci1022,1101'' Node 0x000016 reg: 0000ca00.00000000.00000000.00000000.00000000 compatible: ''pci1022,1102.0'' + ''pci1022,1102'' + ''pciclass,060000'' + ''pciclass,0600'' model: ''Host bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000000 max-latency: 00000000 min-grant: 00000000 unit-address: ''19,2'' class-code: 00060000 revision-id: 00000000 vendor-id: 00001022 device-id: 00001102 name: ''pci1022,1102'' Node 0x000017 reg: 0000cb00.00000000.00000000.00000000.00000000 compatible: ''pci1022,1103.0'' + ''pci1022,1103'' + ''pciclass,060000'' + ''pciclass,0600'' model: ''Host bridge'' power-consumption: 00000001.00000001 devsel-speed: 00000000 max-latency: 00000000 min-grant: 00000000 unit-address: ''19,3'' class-code: 00060000 revision-id: 00000000 vendor-id: 00001022 device-id: 00001103 name: ''pci1022,1103'' This message posted from opensolaris.org
Kugutsumen
2007-Oct-10 11:37 UTC
Re: DMA critical problems under Xen dom0 (b66) with Areca 1110 SATA Raid Ca
Could this be related to this old problem: http://www.nabble.com/high-memory-dma-update:-up-against-a-wall-t140777.html I am going to try to limit the memory to 4 gig on the xen dom0 and see what happened. This message posted from opensolaris.org
Mark Johnson
2007-Oct-10 11:57 UTC
Re: DMA critical problems under Xen dom0 (b66) with Areca 1110 SATA Raid Ca
Kugutsumen wrote:> Could this be related to this old problem: > > http://www.nabble.com/high-memory-dma-update:-up-against-a-wall-t140777.html > > I am going to try to limit the memory to 4 gig on the xen dom0 and see what happened.You''ll actually want to limit it to < 4G since some of that the 4G needs to be used for device registers, etc. e.g. You can limit it to 3G with kernel$ /boot/$ISADIR/xen.gz mem=3G com1=9600,8n1 console=com1 Although, I think we should be able to do a better job allocating low memory... So if this solves your problem, we''ll need to look at it some more... MRJ -- Mark Johnson <mark.johnson@sun.com> Sun Microsystems, Inc. (781) 442-0869
Kugutsumen
2007-Oct-10 12:36 UTC
Re: dom0 dma issue on large memory system with Areca 1100 SATA [solved]
Solved!! :D I limit the dom0 memory to 1024M, reserving 7 gig for my domU''s. kernel$ /boot/$ISADIR/xen.gz dom0_mem=1024M com1=57600,8n1 console=com1 The areca driver doesn''t seem to support dma above the first 32bit region. I will send a new support ticket to areca, I hope they can update their drivers; they should also fix the install script so you can install them without patching it on nevada. There has been a lot of discussion in the past around this 4G limit issue: http://lists.xensource.com/archives/html/xen-devel/2005-06/msg00121.html This message posted from opensolaris.org