[PATCH] posix: Remove dynamic memory allocation from execl{e,p}

Florian Weimer fweimer@redhat.com
Mon Feb 8 13:10:00 GMT 2016


On 01/29/2016 04:52 PM, Adhemerval Zanella wrote:

> Also mmap usage does not solve the second case (vfork/clone) in case of exec success since the call won't unmap the region.

I've written a test case to confirm this.  How unfortunate.

Are there different clone flags we could use?  Probably not, because the
goal is to share the page tables.

For posix_spawn, this doesn't really matter because posix_spawn could
allocate and deallocate the mapping.

Florian



More information about the Libc-alpha mailing list