Hi Dave, sorry to bother you again, I have a few questions about vppinfra/linux/mem.c that 1) in clib_mem_vm_ext_free, it just called munmap with size 1ull<<a->log2_page_size, does it assume all external allocations to be smaller than one page? 2) does clib_mem_create_hugetlb_fd need a vec_free(filename) in the end?
Peng Xia via lists.fd.io <xiapenglilai=gmail....@lists.fd.io> 于2020年4月4日周六 下午12:02写道： > Cool, thanks! > > Dave Barach (dbarach) <dbar...@cisco.com> 于2020年4月3日周五 下午9:11写道： > >> Data segment, .bss segment, stack segment addresses to name a few. >> Although it doesn’t happen every day, it’s almost too easy to free an >> object which wasn’t allocated. >> >> >> >> D >> >> >> >> *From:* vpp-dev@lists.fd.io <vpp-dev@lists.fd.io> *On Behalf Of * >> xiapengli...@gmail.com >> *Sent:* Friday, April 3, 2020 2:57 AM >> *To:* vpp-dev@lists.fd.io >> *Subject:* [vpp-dev] Naive question about mspace_is_heap_object >> >> >> >> Hi experts, >> Curious about what the following lines did in mspace_is_heap_object: >> >> >> >> if (pp > ms->least_addr && pp <= ms->least_addr + ms->footprint) >> >> >> >> return 1 >> What kind of memory allocations would fall out of the existing segments >> of mspace as we have disabled mmap_alloc? And what exactly >> ms->least_addr + ms->footprint means here? >> Btw, just saw least_addr updated during allocation, would >> release_unused_segments make the least_addr out of date and dangeous? >> >> >> > >
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#16023): https://lists.fd.io/g/vpp-dev/message/16023 Mute This Topic: https://lists.fd.io/mt/72743648/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
