| Message ID | 20151124153305.GD8644@n2100.arm.linux.org.uk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
>> Not sure I understand logic behind this. With the current code, >> resulting cpu_possible_mask depends on CONFIG_SMP_ON_UP: >> - if it is set, cpu_possible_mask contains (0 1), as initialized in >> arm_dt_init_cpu_maps() >> - if it is not set, cpu_possible_mask contains (0), since >> imx_smp_init_cpus() removes 1 from there. > > Right, adding debug to arch/arm/kernel/setup.c, just before the > "if (is_smp())" shows: > > is_smp() 0 possible 3 present 1 online 1 > > which is totally wrong: if is_smp() is false, we should not be setting > up any possible CPUs. See a patch below to fix that. > > However, this doesn't matter much, because the code in setup.c won't > initialise the SMP operations struct ... But cpu start code is not the only place in the kernel that uses cpu_present_mask. Are you sure that running with invalid cpu_present_mask has no side effects? > Here's the patch to fix the DT code, which should not be setting > present CPUs when is_smp() is false. I see that this fixes the issue as well. But I still don't understand rationale behind all these is_smp() checks. This makes init sequence different with and without CONFIG_SMP_ON_UP. Isn't kernel intended to run ok without CONFIG_SMP_ON_UP? And if yes - then why not run the same init sequence in both cases?
28.11.2015 14:13, Nikita Yushchenko ?????: >>> Not sure I understand logic behind this. With the current code, >>> resulting cpu_possible_mask depends on CONFIG_SMP_ON_UP: >>> - if it is set, cpu_possible_mask contains (0 1), as initialized in >>> arm_dt_init_cpu_maps() >>> - if it is not set, cpu_possible_mask contains (0), since >>> imx_smp_init_cpus() removes 1 from there. >> >> Right, adding debug to arch/arm/kernel/setup.c, just before the >> "if (is_smp())" shows: >> >> is_smp() 0 possible 3 present 1 online 1 >> >> which is totally wrong: if is_smp() is false, we should not be setting >> up any possible CPUs. See a patch below to fix that. >> >> However, this doesn't matter much, because the code in setup.c won't >> initialise the SMP operations struct ... > > But cpu start code is not the only place in the kernel that uses cpu_present_mask. > > Are you sure that running with invalid cpu_present_mask has no side effects? At least LTP suite does not like it: while running /opt/ltp/runtest/cpuhotplug, we see things like the above <<<test_start>>> tag=cpuhotplug02 stime=1446628761 cmdline="cpuhotplug02.sh -c 1 -l 1" contacts="" analysis=exit <<<test_output>>> Name: cpuhotplug02 Date: Wed Nov 4 09:19:21 UTC 2015 Desc: What happens to a process when its CPU is offlined? CPU is 1 /opt/ltp/testcases/bin/cpuhotplug_hotplug.sh: line 76: echo: write error: Function not implemented cpuhotplug02 1 TBROK : CPU1 cannot be onlined <<<execution_status>>> initiation_status="ok" duration=1 termination_type=exited termination_id=2 corefile=no cutime=4 cstime=6 <<<test_end>>>
diff --git a/arch/arm/kernel/devtree.c b/arch/arm/kernel/devtree.c index 65addcbf5b30..bd72ce91d7a2 100644 --- a/arch/arm/kernel/devtree.c +++ b/arch/arm/kernel/devtree.c @@ -170,15 +170,18 @@ void __init arm_dt_init_cpu_maps(void) return; } - /* - * Since the boot CPU node contains proper data, and all nodes have - * a reg property, the DT CPU list can be considered valid and the - * logical map created in smp_setup_processor_id() can be overridden - */ - for (i = 0; i < cpuidx; i++) { - set_cpu_possible(i, true); - cpu_logical_map(i) = tmp_map[i]; - pr_debug("cpu logical map 0x%x\n", cpu_logical_map(i)); + if (is_smp()) { + /* + * Since the boot CPU node contains proper data, and all + * nodes have a reg property, the DT CPU list can be + * considered valid and the logical map created in + * smp_setup_processor_id() can be overridden + */ + for (i = 0; i < cpuidx; i++) { + set_cpu_possible(i, true); + cpu_logical_map(i) = tmp_map[i]; + pr_debug("cpu logical map 0x%x\n", cpu_logical_map(i)); + } } }