Messages in this thread | | | Date | Fri, 21 Dec 2012 16:59:38 -0800 | Subject | Re: [PATCH 0/9] Avoid populating unbounded num of ptes with mmap_sem held | From | Michel Lespinasse <> |
| |
On Fri, Dec 21, 2012 at 4:36 PM, Andy Lutomirski <luto@amacapital.net> wrote: > On Thu, Dec 20, 2012 at 4:49 PM, Michel Lespinasse <walken@google.com> wrote: >> We have many vma manipulation functions that are fast in the typical case, >> but can optionally be instructed to populate an unbounded number of ptes >> within the region they work on: >> - mmap with MAP_POPULATE or MAP_LOCKED flags; >> - remap_file_pages() with MAP_NONBLOCK not set or when working on a >> VM_LOCKED vma; >> - mmap_region() and all its wrappers when mlock(MCL_FUTURE) is in effect; >> - brk() when mlock(MCL_FUTURE) is in effect. >> > > Something's buggy here. My evil test case is stuck with lots of > threads spinning at 100% system time. Stack traces look like: > > [<0000000000000000>] __mlock_vma_pages_range+0x66/0x70 > [<0000000000000000>] __mm_populate+0xf9/0x150 > [<0000000000000000>] vm_mmap_pgoff+0x9f/0xc0 > [<0000000000000000>] sys_mmap_pgoff+0x7e/0x150 > [<0000000000000000>] sys_mmap+0x22/0x30 > [<0000000000000000>] system_call_fastpath+0x16/0x1b > [<0000000000000000>] 0xffffffffffffffff > > perf top says: > > 38.45% [kernel] [k] __mlock_vma_pages_range > 33.04% [kernel] [k] __get_user_pages > 28.18% [kernel] [k] __mm_populate > > The tasks in question use MCL_FUTURE but not MAP_POPULATE. These > tasks are immune to SIGKILL.
Looking into it.
There seems to be a problem with mlockall - the following program fails in an unkillable way even before my changes:
#include <sys/mman.h> #include <stdio.h> #include <stdint.h>
int main(void) { void *p = mmap(NULL, 0x100000000000, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON | MAP_NORESERVE, -1, 0); printf("p: %p\n", p); mlockall(MCL_CURRENT); return 0; }
I think my changes propagate this existing problem so it now shows up in more places :/
-- Michel "Walken" Lespinasse A program is never fully debugged until the last user dies.
| |