2.6.19 Microcode Update causes a ten second wait
Hello World
With the 2.6.19 kernels I've tried (2.6.19-rc1-git7 and 2.6.19-rc4-git10), enabling the Intel Microcode Update Driver causes the kernel to hang for more than ten seconds when I boot. The last thing it shows on the screen is `TCP reno registered' and then it just stops like its a kernel panic. But it isn't, after about ten seconds the text flies up the screen again and the system boots normally. Disabling support for the microcode update makes the problem go away.
Andy
Quoted 8 lines · Andrew Benton
Hello World
With the 2.6.19 kernels I've tried (2.6.19-rc1-git7 and 2.6.19-rc4-git10), enabling the Intel Microcode Update Driver causes the kernel to hang for more than ten seconds when I boot. The last thing it shows on the screen is `TCP reno registered' and then it just stops like its a kernel panic. But it isn't, after about ten seconds the text flies up the screen again and the system boots normally. Disabling support for the microcode update makes the problem go away.
you're lucky, for me it hangs forever until I add this patch:
arch/i386/kernel/microcode.c renamed from arch/i386/kernel/microcode.c.org +1 -1
@@ -577,7 +577,7 @@ static void microcode_init_cpu(int cpu)set_cpus_allowed (current,cpumask_of_cpu (cpu));mutex_lock (µcode_mutex);collect_cpu_info (cpu); -if (uci->valid) +if (uci->valid && system_state==SYSTEM_RUNNING )cpu_request_microcode (cpu);mutex_unlock (µcode_mutex);set_cpus_allowed (current, old);
Signature · 2 lines
if you want to mail me at work (you don't), use arjan (at) linux.intel.com Test the interaction between Linux and your BIOS via http://www.linuxfirmwarekit.org
Quoted 6 lines · Arjan van de Ven
you're lucky, for me it hangs forever until I add this patch:
arch/i386/kernel/microcode.c static void microcode_init_cpu(int cpu) · line 583
Quoted 8 lines
arch/i386/kernel/microcode.c renamed from arch/i386/kernel/microcode.c.org +1
-1@@ -577,7 +577,7 @@ static void microcode_init_cpu(int cpu)set_cpus_allowed (current,cpumask_of_cpu (cpu));mutex_lock (µcode_mutex);collect_cpu_info (cpu); -if (uci->valid) +if (uci->valid && system_state==SYSTEM_RUNNING )cpu_request_microcode (cpu);mutex_unlock (µcode_mutex);set_cpus_allowed (current, old);
Thanks, that patch fixes the problem for me.
Andy