muneda's diary(2007-11-13)

2007-11-13 晴れ [長年日記]

_1 ネットワーク不調

理由はわからんがPPPoEが失敗。なので外につながらん。

このエントリは携帯から書いてるんだがクソめんどくさいな。特に記号類が。はよ元の環境を復旧させたい。

追記(20071114): 翌日の朝もダメだったのだが，仕事から帰ってきたら復旧してた．その間自宅内に関しては何も変更していないので，外の問題だった様子．

_2 CPU #1 not responding - cannot use it.

何か遅いと思ったら，CPUが1つ見えない．

% cat /proc/cpuinfo | grep processor
processor       : 0

Pentium4 + HTなので2つ見えるはずなのに．メッセージを調べる．

% dmesg| grep -i cpu
[    0.000000] Initializing CPU#0
[   15.807083] SLUB: Genslabs=22, HWalign=64, Order=0-1, MinObjects=4, CPUs=2, Nodes=1
[   15.887134] CPU: After generic identify, caps: bfebfbff 00000000 00000000 00000000 00004400 00000000 00000000
[   15.887147] CPU: Trace cache: 12K uops, L1 D cache: 8K
[   15.887150] CPU: L2 cache: 512K
[   15.887152] CPU: Physical Processor ID: 0
[   15.887155] CPU: After all inits, caps: bfebfbff 00000000 00000000 0000b080 00004400 00000000 00000000
[   16.222765] CPU0: Intel(R) Pentium(R) 4 CPU 2.60GHz stepping 09
[   21.196489] CPU #1 not responding - cannot use it. ☆
[   21.340520] Brought up 1 CPUs
[   21.955283] Switched to high resolution mode on CPU 0
[   23.975778] ACPI: Processor [CPU0] (supports 8 throttling states)
[   23.975804] tick-braodcast: ignoring broadcast for offline CPU #1
[   23.975810] ACPI: Processor [CPU1] (supports 8 throttling states)

☆が非常にあやしい．

% ls /proc/acpi/processor
CPU0/  CPU1/
% diff -u /proc/acpi/processor/*/info
--- /proc/acpi/processor/CPU0/info      2007-11-13 13:50:08.417348138 +0900
+++ /proc/acpi/processor/CPU1/info      2007-11-13 13:50:08.417348138 +0900
@@ -1,5 +1,5 @@
-processor id:            0
-acpi id:                 0
+processor id:            1
+acpi id:                 1
 bus mastering control:   no
 power management:        no
 throttling control:      yes
% diff -u /proc/acpi/processor/*/limit
% diff -u /proc/acpi/processor/*/power
% diff -u /proc/acpi/processor/*/throttling

物理的には2つ認識してそう．

% ls /sys/devices/system/cpu
cpu0/  sched_smt_power_savings

こちらは1つしか見えない．

% ll /sys/devices/acpi_system:00/ACPI0007:00 | grep driver
lrwxrwxrwx 1 root root    0 2007-11-13 13:53 driver -> ../../../bus/acpi/drivers/processor/
% ll /sys/devices/acpi_system:00/ACPI0007:01 | grep driver
lrwxrwxrwx 1 root root    0 2007-11-13 13:53 driver -> ../../../bus/acpi/drivers/processor/

一応 cpu としては 2 つ認識している模様．先ほどのメッセージを吐いているのは

arch/x86/kernel/smpboot_32.c

 963  static void __init smp_boot_cpus(unsigned int max_cpus)
 964  {
...
1055          /*
1056           * Scan the CPU present map and fire up the other CPUs via do_boot_cpu
1057           *
1058           * In clustered apic mode, phys_cpu_present_map is a constructed thus:
1059           * bits 0-3 are quad0, 4-7 are quad1, etc. A perverse twist on the
1060           * clustered apic ID.
1061           */
1062          Dprintk("CPU present map: %lx\n", physids_coerce(phys_cpu_present_map));
1063
1064          kicked = 1;
1065          for (bit = 0; kicked < NR_CPUS && bit < MAX_APICS; bit++) {
1066                  apicid = cpu_present_to_apicid(bit);
1067                  /*
1068                   * Don't even attempt to start the boot CPU!
1069                   */
1070                  if ((apicid == boot_cpu_apicid) || (apicid == BAD_APICID))
1071                          continue;
1072
1073                  if (!check_apicid_present(bit))
1074                          continue;
1075                  if (max_cpus <= cpucount+1)
1076                          continue;
1077
1078                  if (((cpu = alloc_cpu_id()) <= 0) || do_boot_cpu(apicid, cpu))
1079                          printk("CPU #%d not responding - cannot use it.\n",
1080                                                                  apicid);
1081                  else
1082                          ++kicked;
1083          }

printkメッセージが出ていることから，直前の関数のどちらかが失敗したんやろうけど，reboot後は出なくなったのでとりあえず放置．