On 02.02.2019 19:33, Mindaugas Rasiukevicius wrote: > Jason Thorpe <thorpej%me.com@localhost> wrote: >> Patch is here: >> https://patch-diff.githubusercontent.com/raw/thorpej/netbsd-src/pull/5.diff > > Thanks for working on this. > > - Why not just add a new syscall, instead of the KSEM_PSHARED stuff? > > - If you add the 'fd' parameter to ksem_release(), then you can move > fd_putfile() there too (with fd != -1 case) and simplify a little bit. > > - ksem_alloc_pshared_id: can ~KSEM_MARKER_MASK range be exhausted as an > attach vector? > > - Can you eliminate ksem_insert_pshared_locked(), ksem_remove_pshared_locked() > and ksem_remove_pshared() wrappers? They serve no purpose. > > - ksem_lookup_pshared_locked: it makes sense to use a hash table here. > Using hashinit() is clumsy nowadays.. personally, I would replace most > of the subr_hash.c use cases with something like rhashmap [1] and much > more convenient get/put/del semantics. Anyway, that is off-topic, so > up to you if you want to bother. > > > [1] https://github.com/rmind/rhashmap > Are we far away from supporting pshared in pthread.h too? There are already function stubs for this in the header hidden with _PTHREAD_PSHARED.
Attachment:
signature.asc
Description: OpenPGP digital signature