| Message ID | 20171129082145.23327-1-abrodkin@synopsys.com |
|---|---|
| State | New |
| Headers | show |
| Series | ARC: Enable machine_desc->init_per_cpu for non-SMP configs | expand |
On 11/29/2017 12:21 AM, Alexey Brodkin wrote: > As of today we assumed that "machine_desc->init_per_cpu" calls > are only usable on SMP systems when we want to run some piece of > code on early boot for each and every core, I guess assumption was > we have "machine_desc->init_early" for single-core cases where > the one and only master core can do all the things. > > But it turned out for platforms which might be both UP and SMP it > might be benificial to use "init_per_cpu" for both UP and SMP cases > with which we achieve 2 things simultaneously: > 1) Exactly the same one code will be used for UP&SMP for > things required to be done on each an every core regardless if it's > a master and the only core in UP system or any other slave core in SMP > setup. > 1) There will be no "ifdef CONFIG_SMP" around "init_per_cpu". > Seems fine to me. However this needs to go with the actual platform change which needs it. -Vineet
Hi Vineet, On Thu, 2017-11-30 at 09:46 -0800, Vineet Gupta wrote: > On 11/29/2017 12:21 AM, Alexey Brodkin wrote: > > As of today we assumed that "machine_desc->init_per_cpu" calls > > are only usable on SMP systems when we want to run some piece of > > code on early boot for each and every core, I guess assumption was > > we have "machine_desc->init_early" for single-core cases where > > the one and only master core can do all the things. > > > > But it turned out for platforms which might be both UP and SMP it > > might be benificial to use "init_per_cpu" for both UP and SMP cases > > with which we achieve 2 things simultaneously: > > 1) Exactly the same one code will be used for UP&SMP for > > things required to be done on each an every core regardless if it's > > a master and the only core in UP system or any other slave core in SMP > > setup. > > 1) There will be no "ifdef CONFIG_SMP" around "init_per_cpu". > > > > Seems fine to me. However this needs to go with the actual platform change which > needs it. Well for example this might get in the way of building kernel for HSDK with CONFIG_SMP disabled which is IMHO quite valid case in terms of testing code-base compiled with no CONFIG_SMP (something Waldemar was up to). Is it a strong enough reason for that patch to be applied? -Alexey
On 06/14/2018 03:26 PM, Alexey Brodkin wrote: > Hi Vineet, > > On Thu, 2017-11-30 at 09:46 -0800, Vineet Gupta wrote: >> On 11/29/2017 12:21 AM, Alexey Brodkin wrote: >>> As of today we assumed that "machine_desc->init_per_cpu" calls >>> are only usable on SMP systems when we want to run some piece of >>> code on early boot for each and every core, I guess assumption was >>> we have "machine_desc->init_early" for single-core cases where >>> the one and only master core can do all the things. >>> >>> But it turned out for platforms which might be both UP and SMP it >>> might be benificial to use "init_per_cpu" for both UP and SMP cases >>> with which we achieve 2 things simultaneously: >>> 1) Exactly the same one code will be used for UP&SMP for >>> things required to be done on each an every core regardless if it's >>> a master and the only core in UP system or any other slave core in SMP >>> setup. >>> 1) There will be no "ifdef CONFIG_SMP" around "init_per_cpu". >>> >> Seems fine to me. However this needs to go with the actual platform change which >> needs it. > Well for example this might get in the way of building kernel for HSDK with > CONFIG_SMP disabled which is IMHO quite valid case in terms of testing > code-base compiled with no CONFIG_SMP Fair enough - although world is moving away from UP. > (something Waldemar was up to). Really, why does he care if we are running SMP kernel - oh his simulation time goes up. But honestly he should try and use the SMP kernel for testing as that might uncover any system wide / SMP specific bugs in testing uclibc as well ! > Is it a strong enough reason for that patch to be applied? Well per my original reply I didn't have any objections for this per-se and wanted to see the platform patch. Seems we don't need platform patch as it is not under #ifdef in hsdk/platform.c I'll queue this up ! -Vineet
diff --git a/arch/arc/include/asm/mach_desc.h b/arch/arc/include/asm/mach_desc.h index c28e6c347b49..871f3cb16af9 100644 --- a/arch/arc/include/asm/mach_desc.h +++ b/arch/arc/include/asm/mach_desc.h @@ -34,9 +34,7 @@ struct machine_desc { const char *name; const char **dt_compat; void (*init_early)(void); -#ifdef CONFIG_SMP void (*init_per_cpu)(unsigned int); -#endif void (*init_machine)(void); void (*init_late)(void); diff --git a/arch/arc/kernel/irq.c b/arch/arc/kernel/irq.c index 538b36afe89e..62b185057c04 100644 --- a/arch/arc/kernel/irq.c +++ b/arch/arc/kernel/irq.c @@ -31,10 +31,10 @@ void __init init_IRQ(void) /* a SMP H/w block could do IPI IRQ request here */ if (plat_smp_ops.init_per_cpu) plat_smp_ops.init_per_cpu(smp_processor_id()); +#endif if (machine_desc->init_per_cpu) machine_desc->init_per_cpu(smp_processor_id()); -#endif } /*
As of today we assumed that "machine_desc->init_per_cpu" calls are only usable on SMP systems when we want to run some piece of code on early boot for each and every core, I guess assumption was we have "machine_desc->init_early" for single-core cases where the one and only master core can do all the things. But it turned out for platforms which might be both UP and SMP it might be benificial to use "init_per_cpu" for both UP and SMP cases with which we achieve 2 things simultaneously: 1) Exactly the same one code will be used for UP&SMP for things required to be done on each an every core regardless if it's a master and the only core in UP system or any other slave core in SMP setup. 1) There will be no "ifdef CONFIG_SMP" around "init_per_cpu". Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com> --- arch/arc/include/asm/mach_desc.h | 2 -- arch/arc/kernel/irq.c | 2 +- 2 files changed, 1 insertion(+), 3 deletions(-)