From: Kasprzyk, N. <NKa...@im...> - 2008-05-21 11:48:39
|
Hello Henry. Thanks for your quick answer. I needed some more time and did tests with and without the proposed "lpj" kernel parameter. The results: lpj=3000000 =========== Cooperative Linux Daemon, 0.8.0 Daemon compiled on Sun May 11 22:49:57 2008 PID: 1636 colinux: booting Linux version 2.6.22.18-co-0.8.0 (hn@hn-dt) (gcc version 4.2.1 (SUSE Linux)) #1 PREEMPT Sun May 11 22:48:31 UTC 2008 128MB LOWMEM available. initrd enabled: start: 0xc7f9a000 size: 0x00065881 Entering add_active_range(0, 0, 32768) 0 entries of 256 used Zone PFN ranges: DMA 0 -> 0 Normal 0 -> 32768 early_node_map[1] active PFN ranges 0: 0 -> 32768 On node 0 totalpages: 32768 DMA zone: 0 pages used for memmap Normal zone: 256 pages used for memmap Normal zone: 32512 pages, LIFO batch:7 Built 1 zonelists. Total pages: 32512 Kernel command line: root=/dev/hda1 lpj=3000000 Enabling fast FPU save and restore... done. Enabling unmasked SIMD FPU exception support... done. Initializing CPU#0 Setting proxy interrupt vectors PID hash table entries: 512 (order: 9, 2048 bytes) Console: colour CoCON 80x25 Dentry cache hash table entries: 16384 (order: 4, 65536 bytes) Inode-cache hash table entries: 8192 (order: 3, 32768 bytes) conet-slirp-daemon: running Memory: 125268k/131072k available (2058k kernel code, 0k reserved, 585k data, 14 0k init, 0k highmem) virtual kernel memory layout: fixmap : 0xffffc000 - 0xfffff000 ( 12 kB) vmalloc : 0xc8800000 - 0xffffa000 ( 887 MB) lowmem : 0xc0000000 - 0xc8000000 ( 128 MB) .init : 0xc0398000 - 0xc03bb000 ( 140 kB) .data : 0xc0302ac8 - 0xc03951e4 ( 585 kB) .text : 0xc0100000 - 0xc0302ac8 (2058 kB) Calibrating delay loop (skipped)... 600.00 BogoMIPS preset Security Framework v1.0.0 initialized SELinux: Disabled at boot. Capability LSM initialized Mount-cache hash table entries: 512 colinux: Linux VM terminated colinux: BUG at ...412.svn/linux-2.6.22.18-source/kernel/posix-cpu-timers.c:1295 lpj=1000 ======== Slows down the Virtual PC in a way that you believe that the Virtual PC does not work anymore. The VirtualPC process takes 100% of the performance of one of my processors. Fortunately, I have two of them. ;) The last line in the concole is: Inode-cache hash table entries: 8192 (order: 3, 32768 bytes) lpj=5000000 =========== Same behavior as with lpj=1000. Lpj=3000000 again ================= Now, it stops at line .text : 0xc0100000 - 0xc0302ac8 (2058 kB) Without lpj =========== I tried one CoLinux version from 05/11/2008. That is the trace: colinux: booting Linux version 2.6.22.18-co-0.8.0 (hn@hn-dt) (gcc version 4.2.1 (SUSE Linux)) #1 PREEMPT Sun May 11 22:48:31 UTC 2008 128MB LOWMEM available. initrd enabled: start: 0xc7f9a000 size: 0x00065881 Entering add_active_range(0, 0, 32768) 0 entries of 256 used Zone PFN ranges: DMA 0 -> 0 Normal 0 -> 32768 early_node_map[1] active PFN ranges 0: 0 -> 32768 On node 0 totalpages: 32768 DMA zone: 0 pages used for memmap Normal zone: 256 pages used for memmap Normal zone: 32512 pages, LIFO batch:7 Built 1 zonelists. Total pages: 32512 Kernel command line: root=/dev/hda1 Enabling fast FPU save and restore... done. Enabling unmasked SIMD FPU exception support... done. Initializing CPU#0 Setting proxy interrupt vectors PID hash table entries: 512 (order: 9, 2048 bytes) Console: colour CoCON 80x25 Dentry cache hash table entries: 16384 (order: 4, 65536 bytes) Inode-cache hash table entries: 8192 (order: 3, 32768 bytes) Memory: 125268k/131072k available (2058k kernel code, 0k reserved, 585k data, 14 0k init, 0k highmem) virtual kernel memory layout: fixmap : 0xffffc000 - 0xfffff000 ( 12 kB) vmalloc : 0xc8800000 - 0xffffa000 ( 887 MB) lowmem : 0xc0000000 - 0xc8000000 ( 128 MB) .init : 0xc0398000 - 0xc03bb000 ( 140 kB) .data : 0xc0302ac8 - 0xc03951e4 ( 585 kB) .text : 0xc0100000 - 0xc0302ac8 (2058 kB) WARNING: at /home/hn/colinux/build/devel-gcc412.svn/linux-2.6.22.18-source/kerne l/softirq.c:122 _local_bh_enable() [<c0103b7a>] show_trace_log_lvl+0x1a/0x30 [<c0103cb2>] show_trace+0x12/0x20 [<c0104ae5>] dump_stack+0x15/0x20 [<c0115e7a>] _local_bh_enable+0x9a/0xb0 [<c0115f1a>] __do_softirq+0x8a/0xa0 [<c0115f65>] do_softirq+0x35/0x40 [<c01161d3>] irq_exit+0x43/0x50 [<c013d174>] co_callback+0x174/0x380 [<c010a5a1>] proxy_interrupt_handler+0x51/0x70 [<c0103873>] common_interrupt+0x23/0x30 [<c010a4df>] co_register_timer+0xf/0x20 [<c0398b60>] start_kernel+0x1c0/0x2b0 [<00000000>] _stext+0x3fefff50/0x20 [<c010a1de>] co_start_arch+0xfe/0x170 [<c0109ec5>] co_arch_start_kernel+0x0/0xb ======================= Calibrating delay loop... colinux: Linux VM terminated colinux: BUG at ...412.svn/linux-2.6.22.18-source/kernel/posix-cpu-timers.c:1295 conet-slirp-daemon: monitor open failed conet-slirp-daemon: exitcode 8a66b801 At the end of every of the "lpj" tests I had to restart VirtualPC because it hangs. Do you have another proposal? Nico > -----Original Message----- > From: Henry Nestler [mailto:Hen...@Ar...] > Sent: Tuesday, May 20, 2008 2:16 AM > To: Kasprzyk, Nico > Cc: col...@li... > Subject: Re: [coLinux-devel] CoLinux and Virtual PC > > Kasprzyk, Nico wrote: > > Some days before, I tried to run CoLinux 0.8.0 in one > Virtual PC Windows > > XP session. Short summary: It did not work. It is not > possible for me to > > judge whether CoLinux may be corrected to run in Virtual PC > or if this > > is one real serious problem. > > > > Thus, i ask the list: Do you think that it will be possible to run > > CoLinux with Virtual PC? > > > > > > Best regards, > > > > Nico > > > > > > An older version did not run with error: > > ======================================== > > > > Cooperative Linux Daemon, 0.8.0 > > Daemon compiled on Sun Jan 20 19:01:26 2008 > > > > PID: 2316 > > colinux: booting > > Linux version 2.6.22-co-0.8.0 (hn@hn-dt) (gcc version 4.2.1 (SUSE > > Linux)) #1 PRE > > EMPT Sun Jan 20 18:59:42 CET 2008 > > 128MB LOWMEM available. > > initrd enabled: start: 0xc7f9a000 size: 0x00065881 > > [...] > > Calibrating delay loop... WARNING: at > > > /home/hn/colinux/build/devel-gcc412/linux-2.6.22-source/kernel > /softirq.c:122 _local_bh_enable() > > This ia a line with: > WARN_ON_ONCE(!irqs_disabled()); > > > [<c0103b7a>] show_trace_log_lvl+0x1a/0x30 > > [<c0103cb2>] show_trace+0x12/0x20 > > [<c0104ae5>] dump_stack+0x15/0x20 > > [<c0115cfa>] _local_bh_enable+0x9a/0xb0 > > [<c0115d9a>] __do_softirq+0x8a/0xa0 > > [<c0115de5>] do_softirq+0x35/0x40 > > [<c0116053>] irq_exit+0x43/0x50 > > [<c013d154>] co_callback+0x164/0x2f0 > > [<c010a681>] proxy_interrupt_handler+0x41/0x50 > > [<c0103873>] common_interrupt+0x23/0x30 > > [<c0396b65>] start_kernel+0x1c5/0x2b0 > > [<00000000>] _stext+0x3fefff50/0x20 > > [<c010a30e>] co_start_arch+0xfe/0x170 > > [<c0109e95>] co_arch_start_kernel+0x0/0xb > > ======================= > > colinux: Linux VM terminated > > colinux: BUG at > > ...vel-gcc412/linux-2.6.22-source/kernel/posix-cpu-timers.c:1295 > > This is the line > BUG_ON(!irqs_disabled()); > > Think, we can fix it. Our timer callback leaks somethere with enabled > interrupts. I'm afraid. > > As workarround set loops_per_jiffy manualy as kernelparameter and skip > the calibration: "lpj=3000000" (see kernel source > Documentation/kernel-parameters.txt:910) > > We currently have kernel 2.6.22.18 and some changes on the proxy > interrupt handler. Please give the latest snapshot a try. If that also > fails, then send please the new stack trace again. > > > A newer version stopped working with: > > ===================================== > > > > Cooperative Linux Daemon, 0.8.0 > > Daemon compiled on Tue May 13 04:06:17 2008 > > > > daemon: exit code 83279001 > > daemon: error - CO_RC_ERROR_ERROR, line 484, file > > src/colinux/user/daemon.c (25) > > The driver is not loaded correctly. > First, please run this and send me your result: > * cd C:\colinux > * colinux-daemon --status-driver > > To fix it, please do follow steps > * cd C:\colinux > * colinux-daemon --remove-driver > * colinux-daemon --install-driver > * colinux-daemon --status-driver > > If that not helps, then reboot windows between remove and install. > > -- > Henry N. > |