diff options
| author | Prakash, Prashanth <pprakash@codeaurora.org> | 2017-03-24 00:21:53 +0300 |
|---|---|---|
| committer | Rafael J. Wysocki <rafael.j.wysocki@intel.com> | 2017-04-18 17:50:24 +0300 |
| commit | 35b68735988447485074cfa862710e49c082c013 (patch) | |
| tree | 744a792bd55dc12b6122ab28fb08ff8db52482d9 /kernel/locking/rtmutex-debug.c | |
| parent | 4f7d029b9bf009fbee76bb10c0c4351a1870d2f3 (diff) | |
| download | linux-35b68735988447485074cfa862710e49c082c013.tar.xz | |
ACPI / Processor: Drop setup_max_cpus check from acpi_processor_add()
When maxcpus=X kernel argument is used, we parse the ACPI tables only
for the first X cpus. The per-cpu ACPI data include the _PSD tables
which gives information about related cpus.
cppc_cpufreq and acpi-cpufreq parses the table once during init to
deduce the related cpus. If a user brings a new cpu online after boot
the related cpu data becomes incorrect.
acpi_get_psd_map() in acpi_cppc.c returns error if it fails to find
the parsed ACPI data for possible CPU resulting in cppc_cpufreq
initialization failure.
With this change we will probe all possible CPUs prior to cpufreq
initialization, but will bring only setup_max_cpus online. nr_cpus
kernel parameter can be used to restict even parsing per-cpu ACPI
tables.
Signed-off-by: Prashanth Prakash <pprakash@codeaurora.org>
[ rjw: Subject ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Diffstat (limited to 'kernel/locking/rtmutex-debug.c')
0 files changed, 0 insertions, 0 deletions