Date: Tue, 6 Jul 2021 10:00:20 +0100 From: Vincent Donnefort <vincent.donnefort@....com> To: jyknight@...gle.com, musl@...ts.openwall.com Subject: Re: Re: Re: [PATCH v2] sysconf: add _SC_NPROCESSORS_CONF support On Sat, Jul 03, 2021 at 11:20:18PM +0200, Szabolcs Nagy wrote: > * Vincent Donnefort <vincent.donnefort@....com> [2021-07-02 14:29:37 +0100]: > > Here's a new thread as I never received your previous email and I just > > noticed the answer today. > > > > I'm not sure I understand why I can't use "present". As per the kernel > > documentation: > > > > possible: CPUs that have been allocated resources and can be > > brought online if they are present. [cpu_possible_mask] > > > > present: CPUs that have been identified as being present in the > > system. [cpu_present_mask] > > > > > > In this example, the NR_CPUS config option is 128, but the kernel was > > started with possible_cpus=144. There are 4 CPUs in the system and cpu2 > > was manually taken offline (and is the only CPU that can be brought > > online.):: > > > > kernel_max: 127 > > offline: 2,4-127,128-143 > > online: 0-1,3 > > possible: 0-127 > > present: 0-3 > > > > So indeed I could use "possible"... but there's a chance the two masks won't > > be equal, and the sysfs entries are matching "present", not "possible". > > > > "possible" is the CPUs that have allocated resources and can be physically > > added to the system. "present" is the CPUs that are known as physically > > present but might be offline. > > i guess we just need a guarantee that these interfaces are stable > and the set of present cpus don't change during the lifetime of a > process. (can that even work with checkpoint/restore? probably not > our problem) In that case "possible" is what we want. It can't change during the kernel lifetime, while "present" can. -- Vincent
Powered by blists - more mailing lists
Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.