Bug#580901: core i3: suspend to disk break multithreading
On Thu, Sep 06, 2012 at 05:02:50PM +0200, Bill Allombert wrote: On Thu, Sep 06, 2012 at 01:44:52AM -0700, Jonathan Nieder wrote: Hi Bill, In May, 2010, Bill Allombert wrote: I have a Sony Vaio laptop with a core i3 processor, dual core and hyperthreaded to 4 virtual CPUs. When compiling with 'make -j4', top show the 4 CPUs used. After suspend-to-disk however, performance degrade to the equivalent of a single core system Thanks for a clear report. This is said to be fixed by v2.6.32.22~118 (x86, tsc, sched: Recompute cyc2ns_offset's during resume from sleep states, 2010-08-19), which was part of 2.6.32-24. Can you confirm? I will see if I can do that. The first kernel version which worked correctly on this laptop (including other issues I reported) was I checked this version: Package: linux-image-2.6.32-5-amd64 Version: 2.6.32-45 and it indeed fix this problem (but not all, alas). Thanks for your support, -- Bill. ballo...@debian.org Imagine a large red swirl here. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120907145734.GB2870@yellowpig
Bug#580901: core i3: suspend to disk break multithreading
On Fri, Sep 07, 2012 at 08:08:10AM -0700, Jonathan Nieder wrote: Version: 2.6.32-45 Bill Allombert wrote: I checked this version: Package: linux-image-2.6.32-5-amd64 Version: 2.6.32-45 and it indeed fix this problem (but not all, alas). Thanks! Do you remember bug numbers or words I could search for for the other problems? Well, I reported #600845 which was fixed a long time ago. I do not think I reported the problem with the internal speaker, because I was not even sure I had an internal speaker until I found a working kernel. I also miss the VGA console (the framebuffer console is much slower, just try 'time sl'). Cheers, -- Bill. ballo...@debian.org Imagine a large red swirl here. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120907195755.GA16920@yellowpig
Bug#580901: core i3: suspend to disk break multithreading
Bill Allombert wrote: Well, I reported #600845 which was fixed a long time ago. I do not think I reported the problem with the internal speaker, because I was not even sure I had an internal speaker until I found a working kernel. I also miss the VGA console (the framebuffer console is much slower, just try 'time sl'). Ok, thanks. I sometimes miss the speed of the VGA console, too, but I don't think a bugreport would help with it unless someone has an approach in mind for getting back the speed. Internal speaker problems, on the other hand, are probably easier to fix. If you file a report with details, I'll be happy to look into it. Regards, Jonathan -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120907232954.ga...@mannheim-rule.att.net
Bug#580901: core i3: suspend to disk break multithreading
Hi Bill, In May, 2010, Bill Allombert wrote: I have a Sony Vaio laptop with a core i3 processor, dual core and hyperthreaded to 4 virtual CPUs. When compiling with 'make -j4', top show the 4 CPUs used. After suspend-to-disk however, performance degrade to the equivalent of a single core system Thanks for a clear report. This is said to be fixed by v2.6.32.22~118 (x86, tsc, sched: Recompute cyc2ns_offset's during resume from sleep states, 2010-08-19), which was part of 2.6.32-24. Can you confirm? Sincerely, Jonathan -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120906084452.GA697@mannheim-rule.local
Bug#580901: core i3: suspend to disk break multithreading
On Thu, Sep 06, 2012 at 01:44:52AM -0700, Jonathan Nieder wrote: Hi Bill, In May, 2010, Bill Allombert wrote: I have a Sony Vaio laptop with a core i3 processor, dual core and hyperthreaded to 4 virtual CPUs. When compiling with 'make -j4', top show the 4 CPUs used. After suspend-to-disk however, performance degrade to the equivalent of a single core system Thanks for a clear report. This is said to be fixed by v2.6.32.22~118 (x86, tsc, sched: Recompute cyc2ns_offset's during resume from sleep states, 2010-08-19), which was part of 2.6.32-24. Can you confirm? I will see if I can do that. The first kernel version which worked correctly on this laptop (including other issues I reported) was Linux version 2.6.37-trunk-amd64 (Debian 2.6.37-1~experimental.1) (b...@decadent.org.uk) (gcc version 4.4.5 (Debian 4.4.5-10) ) #1 SMP Thu Jan 6 14:13:28 UTC 2011 Cheers, -- Bill. ballo...@debian.org Imagine a large red swirl here. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120906150250.GH28485@yellowpig
Bug#580901: core i3: suspend to disk break multithreading
Package: linux-2.6 Version: 2.6.32-9 Severity: normal Dear Debian kernel team, I have a Sony Vaio laptop with a core i3 processor, dual core and hyperthreaded to 4 virtual CPUs. When compiling with 'make -j4', top show the 4 CPUs used. After suspend-to-disk however, performance degrade to the equivalent of a single core system: top still show 4 cpus, but only one CPU ever get used and compilation time is twice longer (real time). The same happen with programs using multiple POSIX threads. I found this report which seems similar but with a core i7: http://superuser.com/questions/136291/core-i7-on-linux-loses-its-multithreading-capability-after-suspend Note: I have disabled KMS because it causes segfault after suspend-to-disk (this is bug 534422). This is a very bare-bone system, X is not running. Cheers, Bill. -- Package-specific info: ** Version: Linux version 2.6.32-3-amd64 (Debian 2.6.32-9) (m...@debian.org) (gcc version 4.3.4 (Debian 4.3.4-8) ) #1 SMP Wed Feb 24 18:07:42 UTC 2010 ** Command line: BOOT_IMAGE=/boot/vmlinuz-2.6.32-3-amd64 root=UUID=cb8e6b0f-b188-4e36-a732-63694c2af4ca ro vga=normal nomodeset quiet ** Not tainted ** Kernel log: [3.775208] Bluetooth: Generic Bluetooth USB driver ver 0.6 [3.775531] usbcore: registered new interface driver btusb [3.777348] Monitor-Mwait will be used to enter C-1 state [3.795821] Linux video capture interface: v2.00 [3.814678] Monitor-Mwait will be used to enter C-2 state [3.814796] processor LNXCPU:00: registered as cooling_device0 [3.815761] ACPI: SSDT b3690a98 00303 (v01 Sony VAIO 20100107 INTL 20061109) [3.816531] ACPI: SSDT b368ed98 00119 (v01 Sony VAIO 20100107 INTL 20061109) [3.876520] processor LNXCPU:01: registered as cooling_device1 [3.896722] uvcvideo: Found UVC 1.00 device unnamed (05ca:18b7) [3.897346] ACPI: AC Adapter [AC] (on-line) [3.898876] input: UVC Camera (05ca:18b7) as /devices/pci:00/:00:1a.0/usb1/1-1/1-1.2/1-1.2:1.0/input/input6 [3.898987] usbcore: registered new interface driver uvcvideo [3.899030] USB Video Class driver (v0.1.0) [3.920889] ACPI: Battery Slot [BAT1] (battery absent) [3.957380] processor LNXCPU:02: registered as cooling_device2 [4.072640] processor LNXCPU:03: registered as cooling_device3 [4.094364] ACPI: WMI: Mapper loaded [4.097400] sony-laptop: Sony Notebook Control Driver v0.6. [4.149822] input: Sony Vaio Keys as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:0f/SNY5001:00/input/input7 [4.150181] input: Sony Vaio Jogdial as /devices/virtual/input/input8 [4.150467] sony-laptop: brightness ignored, must be controlled by ACPI video driver [4.171318] iwlagn: Intel(R) Wireless WiFi Link AGN driver for Linux, 1.3.27ks [4.171321] iwlagn: Copyright(c) 2003-2009 Intel Corporation [4.171391] iwlagn :02:00.0: PCI INT A - GSI 16 (level, low) - IRQ 16 [4.171400] iwlagn :02:00.0: setting latency timer to 64 [4.171458] iwlagn :02:00.0: Detected Intel Wireless WiFi Link 6000 Series 2x2 AGN REV=0x74 [4.196368] iwlagn :02:00.0: Tunable channels: 13 802.11bg, 24 802.11a channels [4.196422] alloc irq_desc for 35 on node -1 [4.196425] alloc kstat_irqs on node -1 [4.196446] iwlagn :02:00.0: irq 35 for MSI/MSI-X [4.256105] phy0: Selected rate control algorithm 'iwl-agn-rs' [4.268538] alloc irq_desc for 22 on node -1 [4.268541] alloc kstat_irqs on node -1 [4.268551] HDA Intel :00:1b.0: PCI INT A - GSI 22 (level, low) - IRQ 22 [4.268623] HDA Intel :00:1b.0: setting latency timer to 64 [4.567946] EXT3 FS on sda1, internal journal [4.643992] loop: module loaded [4.769324] Synaptics Touchpad, model: 1, fw: 7.4, id: 0x1c0b1, caps: 0xd04731/0xa4 [4.786244] Adding 11718648k swap on /dev/mapper/Debian-swap. Priority:-1 extents:1 across:11718648k [4.816656] input: SynPS/2 Synaptics TouchPad as /devices/platform/i8042/serio4/input/input9 [4.905589] kjournald starting. Commit interval 5 seconds [4.905708] EXT3 FS on dm-5, internal journal [4.905718] EXT3-fs: mounted filesystem with ordered data mode. [4.907473] kjournald starting. Commit interval 5 seconds [4.907678] EXT3 FS on dm-4, internal journal [4.907686] EXT3-fs: mounted filesystem with ordered data mode. [4.909288] kjournald starting. Commit interval 5 seconds [4.909431] EXT3 FS on dm-1, internal journal [4.909438] EXT3-fs: mounted filesystem with ordered data mode. [4.911040] kjournald starting. Commit interval 5 seconds [4.911193] EXT3 FS on dm-2, internal journal [4.911200] EXT3-fs: mounted filesystem with ordered data mode. [4.913038] kjournald starting. Commit interval 5 seconds [4.913179] EXT3 FS on dm-3, internal journal [4.913187] EXT3-fs: mounted filesystem with ordered data mode. [5.097935] nf_conntrack version 0.5.0 (16384 buckets, 65536 max) [