Greetings! Raymond Toy <[email protected]> writes:
>>>>>> "Camm" == Camm Maguire <[email protected]> writes: > > Camm> I haven't updated cvs head recently as I've been trying to > Camm> solidify the experimental branch, which I hope to make the > > Oh. Maybe I should try the experimental branch. > I'm working on it now. Bootstrap is very slow. I'll post a note when its ready. > >> Why do you need a heap ceiling? How would I find out a value > >> for Solaris/sparc? > > Camm> gcl uses the sbrk system call to grow memory contiguously > Camm> from the end of the .data section of the executable. > Camm> Various systems load shared libraries at some arbitrary (but > Camm> 'basically' fixed) address above this point. Alas, I have > > I'm not 100% sure about this, but there's a comment in the cmucl > sources for sparc that says: > > * It may be possible to increase the size of the dynamic spaces even > * further, but Casper H.S. Dik says shared libraries are loaded > * directly under the stack, so we need to leave some space for the C > * stack and shared libraries. He also says the top of stack is > * 0xffbf0000 for Ultrasparcs in Solaris 7+, but it's 0xf0000000 for > * sun4m (and 4u in S2.6-) 0xe0000000 for sun4d) > * > * Shared libraries can be mapped anywhere there's room. > > I guess that means they aren't fixed, but you can grow to a fairly > large value if you don't load up tons of shared libraries. > So we might be able to default to 0 for heap_ceiling, then. Take care, > Ray > > > _______________________________________________ > Gcl-devel mailing list > [email protected] > https://lists.gnu.org/mailman/listinfo/gcl-devel > > > > -- Camm Maguire [email protected] ========================================================================== "The earth is but one country, and mankind its citizens." -- Baha'u'llah _______________________________________________ Gcl-devel mailing list [email protected] https://lists.gnu.org/mailman/listinfo/gcl-devel
