On Sat, Jul 4, 2020 at 1:13 AM <dufa...@hda.com> wrote: > This thought is intended to be evaluated, not adopted. What I meant was I > think you have a "bitmask" of threads that need access to a stack that is > similar to a "bitmask" of threads/processes that may be blocked in select. > The select-mask is a major focus across operating systems. That's why I > thought that if you consider it to be equivalent you should use that > interface. > > Oh, that makes more sense. Thank you for the clarification!
> On Jul 3, 2020, at 08:42 , dufa...@hda.com wrote: > > > > My thought is that it matches what is needed and is expected to be > optimized. > > > >> On Jul 3, 2020, at 24:37 , Utkarsh Rai <utkarsh.ra...@gmail.com> wrote: > >> > >> > >> > >> On Fri, Jul 3, 2020 at 1:32 AM Peter Dufault <dufa...@hda.com> wrote: > >> I finally have gotten to reviewing Utkarsh's work in GSOC. One item > that I don't like is that there is a linked list management of the threads > that have access to the stack. > >> I think this access is similar to the file descriptors that are blocked > in a select call. On a given architecture I don't know exactly how many > file descriptors I can block, I don't know exactly what the FD_SET macro > does, but I have access to multiple file descriptors through the FD_* > macros. > >> > >> > >> Hello, thank you for the feedback. If I understand your suggestion > correctly, we can specify a file descriptor set 'fd_set' and then register > the stack attributes to this set, and then check for the 'current stack' > through select() and FD_ISSET()? > >> > >> Using FD_SET and friends for specifying thread access could be a good > model to specify which threads need access to which thread. > >> > >> My question is, what benefit do we get by using FD_SET and friends? > Ultimately we will have to iterate through the set to check for the > 'current stack'. > >> > >> However, it won't scale infinitely. Linked lists won't scale > infinitely in real-time either. > >> > >> An optimization that I had in mind was to use the LRU model. This way > we can mark the stack attributes of the last thread as 'not current' > without iterating through the list or the set. > >> > >> Peter > >> ----------------- > >> Peter Dufault > >> HD Associates, Inc. Software and System Engineering > >> > >> This email is delivered through the public internet using protocols > subject to interception and tampering. > >> > > > > Peter > > ----------------- > > Peter Dufault > > HD Associates, Inc. Software and System Engineering > > > > This email is delivered through the public internet using protocols > subject to interception and tampering. > > > > _______________________________________________ > > devel mailing list > > devel@rtems.org > > http://lists.rtems.org/mailman/listinfo/devel > > Peter > ----------------- > Peter Dufault > HD Associates, Inc. Software and System Engineering > > This email is delivered through the public internet using protocols > subject to interception and tampering. > >
_______________________________________________ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
