On 01/31/2012 04:17 PM, Venkatesh Pallipadi wrote:
> Kernel's notion of possible cpus (from include/linux/cpumask.h)
> * cpu_possible_mask- has bit 'cpu' set iff cpu is populatable
>
> * The cpu_possible_mask is fixed at boot time, as the set of CPU id's
> * that it is possible might ever be plugged in at anytime during the
> * life of that system boot.
>
> #define num_possible_cpus() cpumask_weight(cpu_possible_mask)
>
> and on x86 cpumask_weight() calls hweight64 and hweight64 (on older kernels
> and systems with !X86_FEATURE_POPCNT) or a popcnt based alternative.
>
> i.e, We needlessly go through this mask based calculation everytime
> num_possible_cpus() is called.
>
> The problem is there with cpu_online_mask() as well, which is fixed value at
> boot time in !CONFIG_HOTPLUG_CPU case and should not change that often even
> in HOTPLUG case.
>
> Though most of the callers of these two routines are init time (with few
> exceptions of runtime calls), it is cleaner to use variables
> and not go through this repeated mask based calculation.
>
> Signed-off-by: Venkatesh Pallipadi<[email protected]>
> Acked-by: Srivatsa S. Bhat<[email protected]>
> Acked-by: KOSAKI Motohiro<[email protected]>
How is it that this patch got merged to linux-next before all the
cleanup patches for nr_online_cpus?
From the looks of your follow-on patches it would seem that all MIPS,
hexagon, and um are now broken.
I know for a fact that MIPS doesn't boot because of this.
David Daney
> ---
> include/linux/cpumask.h | 9 +++++++--
> kernel/cpu.c | 13 +++++++++++++
> 2 files changed, 20 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h
> index 4f7a632..930e255 100644
> --- a/include/linux/cpumask.h
> +++ b/include/linux/cpumask.h
> @@ -27,6 +27,9 @@ typedef struct cpumask { DECLARE_BITMAP(bits, NR_CPUS); } cpumask_t;
> extern int nr_cpu_ids;
> #endif
>
> +extern unsigned int nr_online_cpus;
> +extern unsigned int nr_possible_cpus;
> +
> #ifdef CONFIG_CPUMASK_OFFSTACK
> /* Assuming NR_CPUS is huge, a runtime limit is more efficient. Also,
> * not all bits may be allocated. */
> @@ -81,8 +84,10 @@ extern const struct cpumask *const cpu_present_mask;
> extern const struct cpumask *const cpu_active_mask;
>
> #if NR_CPUS> 1
> -#define num_online_cpus() cpumask_weight(cpu_online_mask)
> -#define num_possible_cpus() cpumask_weight(cpu_possible_mask)
> +
> +#define num_online_cpus() (nr_online_cpus)
> +#define num_possible_cpus() (nr_possible_cpus)
> +
> #define num_present_cpus() cpumask_weight(cpu_present_mask)
> #define num_active_cpus() cpumask_weight(cpu_active_mask)
> #define cpu_online(cpu) cpumask_test_cpu((cpu), cpu_online_mask)
> diff --git a/kernel/cpu.c b/kernel/cpu.c
> index 2060c6e..d520d34 100644
> --- a/kernel/cpu.c
> +++ b/kernel/cpu.c
> @@ -604,9 +604,13 @@ EXPORT_SYMBOL(cpu_all_bits);
> #ifdef CONFIG_INIT_ALL_POSSIBLE
> static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly
> = CPU_BITS_ALL;
> +unsigned int nr_possible_cpus __read_mostly = CONFIG_NR_CPUS;
> #else
> static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly;
> +unsigned int nr_possible_cpus __read_mostly;
> #endif
> +EXPORT_SYMBOL(nr_possible_cpus);
> +
> const struct cpumask *const cpu_possible_mask = to_cpumask(cpu_possible_bits);
> EXPORT_SYMBOL(cpu_possible_mask);
>
> @@ -614,6 +618,9 @@ static DECLARE_BITMAP(cpu_online_bits, CONFIG_NR_CPUS) __read_mostly;
> const struct cpumask *const cpu_online_mask = to_cpumask(cpu_online_bits);
> EXPORT_SYMBOL(cpu_online_mask);
>
> +unsigned int nr_online_cpus __read_mostly;
> +EXPORT_SYMBOL(nr_online_cpus);
> +
> static DECLARE_BITMAP(cpu_present_bits, CONFIG_NR_CPUS) __read_mostly;
> const struct cpumask *const cpu_present_mask = to_cpumask(cpu_present_bits);
> EXPORT_SYMBOL(cpu_present_mask);
> @@ -628,6 +635,8 @@ void set_cpu_possible(unsigned int cpu, bool possible)
> cpumask_set_cpu(cpu, to_cpumask(cpu_possible_bits));
> else
> cpumask_clear_cpu(cpu, to_cpumask(cpu_possible_bits));
> +
> + nr_possible_cpus = cpumask_weight(cpu_possible_mask);
> }
>
> void set_cpu_present(unsigned int cpu, bool present)
> @@ -644,6 +653,8 @@ void set_cpu_online(unsigned int cpu, bool online)
> cpumask_set_cpu(cpu, to_cpumask(cpu_online_bits));
> else
> cpumask_clear_cpu(cpu, to_cpumask(cpu_online_bits));
> +
> + nr_online_cpus = cpumask_weight(cpu_online_mask);
> }
>
> void set_cpu_active(unsigned int cpu, bool active)
> @@ -662,9 +673,11 @@ void init_cpu_present(const struct cpumask *src)
> void init_cpu_possible(const struct cpumask *src)
> {
> cpumask_copy(to_cpumask(cpu_possible_bits), src);
> + nr_possible_cpus = cpumask_weight(cpu_possible_mask);
> }
>
> void init_cpu_online(const struct cpumask *src)
> {
> cpumask_copy(to_cpumask(cpu_online_bits), src);
> + nr_online_cpus = cpumask_weight(cpu_online_mask);
> }
On Mon, 27 Feb 2012 13:55:55 -0800
David Daney <[email protected]> wrote:
> On 01/31/2012 04:17 PM, Venkatesh Pallipadi wrote:
> > Kernel's notion of possible cpus (from include/linux/cpumask.h)
> > * cpu_possible_mask- has bit 'cpu' set iff cpu is populatable
> >
> > * The cpu_possible_mask is fixed at boot time, as the set of CPU id's
> > * that it is possible might ever be plugged in at anytime during the
> > * life of that system boot.
> >
> > #define num_possible_cpus() cpumask_weight(cpu_possible_mask)
> >
> > and on x86 cpumask_weight() calls hweight64 and hweight64 (on older kernels
> > and systems with !X86_FEATURE_POPCNT) or a popcnt based alternative.
> >
> > i.e, We needlessly go through this mask based calculation everytime
> > num_possible_cpus() is called.
> >
> > The problem is there with cpu_online_mask() as well, which is fixed value at
> > boot time in !CONFIG_HOTPLUG_CPU case and should not change that often even
> > in HOTPLUG case.
> >
> > Though most of the callers of these two routines are init time (with few
> > exceptions of runtime calls), it is cleaner to use variables
> > and not go through this repeated mask based calculation.
> >
> > Signed-off-by: Venkatesh Pallipadi<[email protected]>
> > Acked-by: Srivatsa S. Bhat<[email protected]>
> > Acked-by: KOSAKI Motohiro<[email protected]>
>
> How is it that this patch got merged to linux-next before all the
> cleanup patches for nr_online_cpus?
<spends five minutes searching mailing list archives>
I for one do not have a clue what patches the term "cleanup patches for
nr_online_cpus" refers to. Patches have names - please use them!
> From the looks of your follow-on patches it would seem that all MIPS,
> hexagon, and um are now broken.
>
> I know for a fact that MIPS doesn't boot because of this.
I shall drop
cpumask-avoid-mask-based-num_possible_cpus-and-num_online_cpus.patch.
That patch was sent as a single standalone patch and the changelog had
no mention of any needed preparatory patches. If resending, please
send *all* patches in a single sequence-numbered series. We know how
to do this.
On 02/27/2012 02:07 PM, Andrew Morton wrote:
> On Mon, 27 Feb 2012 13:55:55 -0800
> David Daney<[email protected]> wrote:
>
>> On 01/31/2012 04:17 PM, Venkatesh Pallipadi wrote:
>>> Kernel's notion of possible cpus (from include/linux/cpumask.h)
>>> * cpu_possible_mask- has bit 'cpu' set iff cpu is populatable
>>>
>>> * The cpu_possible_mask is fixed at boot time, as the set of CPU id's
>>> * that it is possible might ever be plugged in at anytime during the
>>> * life of that system boot.
>>>
>>> #define num_possible_cpus() cpumask_weight(cpu_possible_mask)
>>>
>>> and on x86 cpumask_weight() calls hweight64 and hweight64 (on older kernels
>>> and systems with !X86_FEATURE_POPCNT) or a popcnt based alternative.
>>>
>>> i.e, We needlessly go through this mask based calculation everytime
>>> num_possible_cpus() is called.
>>>
>>> The problem is there with cpu_online_mask() as well, which is fixed value at
>>> boot time in !CONFIG_HOTPLUG_CPU case and should not change that often even
>>> in HOTPLUG case.
>>>
>>> Though most of the callers of these two routines are init time (with few
>>> exceptions of runtime calls), it is cleaner to use variables
>>> and not go through this repeated mask based calculation.
>>>
>>> Signed-off-by: Venkatesh Pallipadi<[email protected]>
>>> Acked-by: Srivatsa S. Bhat<[email protected]>
>>> Acked-by: KOSAKI Motohiro<[email protected]>
>>
>> How is it that this patch got merged to linux-next before all the
>> cleanup patches for nr_online_cpus?
>
> <spends five minutes searching mailing list archives>
>
> I for one do not have a clue what patches the term "cleanup patches for
> nr_online_cpus" refers to. Patches have names - please use them!
>
Sorry about that. I was a little hasty:
From: Venkatesh Pallipadi <[email protected]>
Subject: [PATCH 0/3] Cleanup raw handling of online/possible map
Date: Tue, 14 Feb 2012 14:49:41 -0800
Message-Id: <[email protected]>
I don't really know a better way to refer to them.
David Daney
>> From the looks of your follow-on patches it would seem that all MIPS,
>> hexagon, and um are now broken.
>>
>> I know for a fact that MIPS doesn't boot because of this.
>
> I shall drop
> cpumask-avoid-mask-based-num_possible_cpus-and-num_online_cpus.patch.
>
> That patch was sent as a single standalone patch and the changelog had
> no mention of any needed preparatory patches. If resending, please
> send *all* patches in a single sequence-numbered series. We know how
> to do this.
It is possible that Venkatesh did not know about the breakage when the
original patch was sent
Hi Andrew,
On Mon, 27 Feb 2012 14:07:46 -0800 Andrew Morton <[email protected]> wrote:
>
> I shall drop
> cpumask-avoid-mask-based-num_possible_cpus-and-num_online_cpus.patch.
I have dropped that from the akpm tree today.
--
Cheers,
Stephen Rothwell [email protected]