| Message ID | 1476798435-8357-1-git-send-email-noamca@mellanox.com |
|---|---|
| State | New |
| Headers | show |
On 10/18/2016 06:47 AM, Noam Camus wrote: > From: Noam Camus <noamc@ezchip.com> > > num_possible_cpus() returns how many CPUs may be present on system. > However we want the highest possible CPU number. Highest possible number "Detected" at boot ? Can you explain a bit more ! > This may be differ in a sparsed possible CPUs map. > > Signed-off-by: Noam Camus <noamc@ezchip.com> > Acked-by: Vineet Gupta <vgupta@synopsys.com> > --- > arch/arc/kernel/setup.c | 2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/arch/arc/kernel/setup.c b/arch/arc/kernel/setup.c > index f52a0d0..d13ce84 100644 > --- a/arch/arc/kernel/setup.c > +++ b/arch/arc/kernel/setup.c > @@ -506,7 +506,7 @@ static void *c_start(struct seq_file *m, loff_t *pos) > * way to pass it w/o having to kmalloc/free a 2 byte string. > * Encode cpu-id as 0xFFcccc, which is decoded by show routine. > */ > - return *pos < num_possible_cpus() ? cpu_to_ptr(*pos) : NULL; > + return *pos < nr_cpu_ids ? cpu_to_ptr(*pos) : NULL; My only concern is that for smaller systems - say if I build with NR_CPUS = 4, and there are only 2 cpus present, then /proc/cpuinfo needs to print the fact. And a quick test with ur patch says that is the case - so how does this help you. Do u use nr_cpus cmdline toggle ? [ARCLinux]# cat /proc/cpuinfo IDENTITY : ARCVER [0x51] ARCNUM [0x0] CHIPID [ 0x0] processor [0] : ARC HS38 R2.0 (ARCv2 ISA) ... ... processor [2] : Offline processor [3] : Offline -Vineet > } > > static void *c_next(struct seq_file *m, void *v, loff_t *pos)
> From: Vineet Gupta [mailto:Vineet.Gupta1@synopsys.com] > Sent: Tuesday, October 18, 2016 8:49 PM >> >> num_possible_cpus() returns how many CPUs may be present on system. >> However we want the highest possible CPU number. > >Highest possible number "Detected" at boot ? Can you explain a bit more ! While trying to provide explanation I decided to "rethink" on this patch. Its effect is not visible to you due to another "local" patch I use which justify this patch. I need to reconsider both patches (maybe I will even remove both of them). Thanks for the feedback. Noam
diff --git a/arch/arc/kernel/setup.c b/arch/arc/kernel/setup.c index f52a0d0..d13ce84 100644 --- a/arch/arc/kernel/setup.c +++ b/arch/arc/kernel/setup.c @@ -506,7 +506,7 @@ static void *c_start(struct seq_file *m, loff_t *pos) * way to pass it w/o having to kmalloc/free a 2 byte string. * Encode cpu-id as 0xFFcccc, which is decoded by show routine. */ - return *pos < num_possible_cpus() ? cpu_to_ptr(*pos) : NULL; + return *pos < nr_cpu_ids ? cpu_to_ptr(*pos) : NULL; } static void *c_next(struct seq_file *m, void *v, loff_t *pos)