1. 03 Apr, 2009 1 commit
  2. 10 Feb, 2009 1 commit
    • Mel Gorman's avatar
      Do not account for the address space used by hugetlbfs using VM_ACCOUNT · 5a6fe125
      Mel Gorman authored
      When overcommit is disabled, the core VM accounts for pages used by anonymous
      shared, private mappings and special mappings. It keeps track of VMAs that
      should be accounted for with VM_ACCOUNT and VMAs that never had a reserve
      with VM_NORESERVE.
      
      Overcommit for hugetlbfs is much riskier than overcommit for base pages
      due to contiguity requirements. It avoids overcommiting on both shared and
      private mappings using reservation counters that are checked and updated
      during mmap(). This ensures (within limits) that hugepages exist in the
      future when faults occurs or it is too easy to applications to be SIGKILLed.
      
      As hugetlbfs makes its own reservations of a different unit to the base page
      size, VM_ACCOUNT should never be set. Even if the units were correct, we would
      double account for the usage in the core VM and hugetlbfs. VM_NORESERVE may
      be set because an application can request no reserves be made for hugetlbfs
      at the risk of getting killed later.
      
      With commit fc8744ad, VM_NORESERVE and
      VM_ACCOUNT are getting unconditionally set for hugetlbfs-backed mappings. This
      breaks the accounting for both the core VM and hugetlbfs, can trigger an
      OOM storm when hugepage pools are too small lockups and corrupted counters
      otherwise are used. This patch brings hugetlbfs more in line with how the
      core VM treats VM_NORESERVE but prevents VM_ACCOUNT being set.
      Signed-off-by: default avatarMel Gorman <mel@csn.ul.ie>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      5a6fe125
  3. 05 Feb, 2009 2 commits
  4. 31 Jan, 2009 1 commit
    • Linus Torvalds's avatar
      Stop playing silly games with the VM_ACCOUNT flag · fc8744ad
      Linus Torvalds authored
      The mmap_region() code would temporarily set the VM_ACCOUNT flag for
      anonymous shared mappings just to inform shmem_zero_setup() that it
      should enable accounting for the resulting shm object.  It would then
      clear the flag after calling ->mmap (for the /dev/zero case) or doing
      shmem_zero_setup() (for the MAP_ANON case).
      
      This just resulted in vma merge issues, but also made for just
      unnecessary confusion.  Use the already-existing VM_NORESERVE flag for
      this instead, and let shmem_{zero|file}_setup() just figure it out from
      that.
      
      This also happens to make it obvious that the new DRI2 GEM layer uses a
      non-reserving backing store for its object allocation - which is quite
      possibly not intentional.  But since I didn't want to change semantics
      in this patch, I left it alone, and just updated the caller to use the
      new flag semantics.
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      fc8744ad
  5. 14 Jan, 2009 1 commit
  6. 08 Jan, 2009 1 commit
    • David Howells's avatar
      NOMMU: Make VMAs per MM as for MMU-mode linux · 8feae131
      David Howells authored
      Make VMAs per mm_struct as for MMU-mode linux.  This solves two problems:
      
       (1) In SYSV SHM where nattch for a segment does not reflect the number of
           shmat's (and forks) done.
      
       (2) In mmap() where the VMA's vm_mm is set to point to the parent mm by an
           exec'ing process when VM_EXECUTABLE is specified, regardless of the fact
           that a VMA might be shared and already have its vm_mm assigned to another
           process or a dead process.
      
      A new struct (vm_region) is introduced to track a mapped region and to remember
      the circumstances under which it may be shared and the vm_list_struct structure
      is discarded as it's no longer required.
      
      This patch makes the following additional changes:
      
       (1) Regions are now allocated with alloc_pages() rather than kmalloc() and
           with no recourse to __GFP_COMP, so the pages are not composite.  Instead,
           each page has a reference on it held by the region.  Anything else that is
           interested in such a page will have to get a reference on it to retain it.
           When the pages are released due to unmapping, each page is passed to
           put_page() and will be freed when the page usage count reaches zero.
      
       (2) Excess pages are trimmed after an allocation as the allocation must be
           made as a power-of-2 quantity of pages.
      
       (3) VMAs are added to the parent MM's R/B tree and mmap lists.  As an MM may
           end up with overlapping VMAs within the tree, the VMA struct address is
           appended to the sort key.
      
       (4) Non-anonymous VMAs are now added to the backing inode's prio list.
      
       (5) Holes may be punched in anonymous VMAs with munmap(), releasing parts of
           the backing region.  The VMA and region structs will be split if
           necessary.
      
       (6) sys_shmdt() only releases one attachment to a SYSV IPC shared memory
           segment instead of all the attachments at that addresss.  Multiple
           shmat()'s return the same address under NOMMU-mode instead of different
           virtual addresses as under MMU-mode.
      
       (7) Core dumping for ELF-FDPIC requires fewer exceptions for NOMMU-mode.
      
       (8) /proc/maps is now the global list of mapped regions, and may list bits
           that aren't actually mapped anywhere.
      
       (9) /proc/meminfo gains a line (tagged "MmapCopy") that indicates the amount
           of RAM currently allocated by mmap to hold mappable regions that can't be
           mapped directly.  These are copies of the backing device or file if not
           anonymous.
      
      These changes make NOMMU mode more similar to MMU mode.  The downside is that
      NOMMU mode requires some extra memory to track things over NOMMU without this
      patch (VMAs are no longer shared, and there are now region structs).
      Signed-off-by: default avatarDavid Howells <dhowells@redhat.com>
      Tested-by: default avatarMike Frysinger <vapier.adi@gmail.com>
      Acked-by: default avatarPaul Mundt <lethal@linux-sh.org>
      8feae131
  7. 06 Jan, 2009 1 commit
  8. 04 Jan, 2009 1 commit
  9. 13 Nov, 2008 3 commits
  10. 21 Oct, 2008 1 commit
  11. 20 Oct, 2008 1 commit
    • Lee Schermerhorn's avatar
      SHM_LOCKED pages are unevictable · 89e004ea
      Lee Schermerhorn authored
      Shmem segments locked into memory via shmctl(SHM_LOCKED) should not be
      kept on the normal LRU, since scanning them is a waste of time and might
      throw off kswapd's balancing algorithms.  Place them on the unevictable
      LRU list instead.
      
      Use the AS_UNEVICTABLE flag to mark address_space of SHM_LOCKed shared
      memory regions as unevictable.  Then these pages will be culled off the
      normal LRU lists during vmscan.
      
      Add new wrapper function to clear the mapping's unevictable state when/if
      shared memory segment is munlocked.
      
      Add 'scan_mapping_unevictable_page()' to mm/vmscan.c to scan all pages in
      the shmem segment's mapping [struct address_space] for evictability now
      that they're no longer locked.  If so, move them to the appropriate zone
      lru list.
      
      Changes depend on [CONFIG_]UNEVICTABLE_LRU.
      
      [kosaki.motohiro@jp.fujitsu.com: revert shm change]
      Signed-off-by: default avatarLee Schermerhorn <lee.schermerhorn@hp.com>
      Signed-off-by: default avatarRik van Riel <riel@redhat.com>
      Signed-off-by: default avatarKosaki Motohiro <kosaki.motohiro@jp.fujitsu.com>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      89e004ea
  12. 25 Jul, 2008 1 commit
  13. 24 Jul, 2008 1 commit
  14. 13 Jun, 2008 1 commit
    • Paul Menage's avatar
      /proc/sysvipc/shm: fix 32-bit truncation of segment sizes · 6c826818
      Paul Menage authored
      sysvipc_shm_proc_show() picks between format strings (based on the
      expected maximum length of a SHM segment) in a way that prevents gcc from
      performing format checks on the seq_printf() parameters.  This hid two
      format errors - shp->shm_segsz and shp->shm_nattach are both unsigned
      long, but were being printed as unsigned int and signed int respectively.
      This leads to 32-bit truncation of SHM segment sizes reported in
      /proc/sysvipc/shm.  (And for nattach, but that's less of a problem for
      most users).
      
      This patch makes the format string directly visible to gcc's format
      specifier checker, and fixes the two broken format specifiers.
      Signed-off-by: default avatarPaul Menage <menage@google.com>
      Cc: Nadia Derbey <Nadia.Derbey@bull.net>
      Cc: Manfred Spraul <manfred@colorfullife.com>
      Cc: Pierre Peiffer <peifferp@gmail.com>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      6c826818
  15. 10 Jun, 2008 1 commit
  16. 29 Apr, 2008 5 commits
  17. 28 Apr, 2008 2 commits
    • Lee Schermerhorn's avatar
      mempolicy: rework mempolicy Reference Counting [yet again] · 52cd3b07
      Lee Schermerhorn authored
      After further discussion with Christoph Lameter, it has become clear that my
      earlier attempts to clean up the mempolicy reference counting were a bit of
      overkill in some areas, resulting in superflous ref/unref in what are usually
      fast paths.  In other areas, further inspection reveals that I botched the
      unref for interleave policies.
      
      A separate patch, suitable for upstream/stable trees, fixes up the known
      errors in the previous attempt to fix reference counting.
      
      This patch reworks the memory policy referencing counting and, one hopes,
      simplifies the code.  Maybe I'll get it right this time.
      
      See the update to the numa_memory_policy.txt document for a discussion of
      memory policy reference counting that motivates this patch.
      
      Summary:
      
      Lookup of mempolicy, based on (vma, address) need only add a reference for
      shared policy, and we need only unref the policy when finished for shared
      policies.  So, this patch backs out all of the unneeded extra reference
      counting added by my previous attempt.  It then unrefs only shared policies
      when we're finished with them, using the mpol_cond_put() [conditional put]
      helper function introduced by this patch.
      
      Note that shmem_swapin() calls read_swap_cache_async() with a dummy vma
      containing just the policy.  read_swap_cache_async() can call alloc_page_vma()
      multiple times, so we can't let alloc_page_vma() unref the shared policy in
      this case.  To avoid this, we make a copy of any non-null shared policy and
      remove the MPOL_F_SHARED flag from the copy.  This copy occurs before reading
      a page [or multiple pages] from swap, so the overhead should not be an issue
      here.
      
      I introduced a new static inline function "mpol_cond_copy()" to copy the
      shared policy to an on-stack policy and remove the flags that would require a
      conditional free.  The current implementation of mpol_cond_copy() assumes that
      the struct mempolicy contains no pointers to dynamically allocated structures
      that must be duplicated or reference counted during copy.
      Signed-off-by: default avatarLee Schermerhorn <lee.schermerhorn@hp.com>
      Cc: Christoph Lameter <clameter@sgi.com>
      Cc: David Rientjes <rientjes@google.com>
      Cc: Mel Gorman <mel@csn.ul.ie>
      Cc: Andi Kleen <ak@suse.de>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      52cd3b07
    • Lee Schermerhorn's avatar
      mempolicy: fixup Fallback for Default Shmem Policy · ae4d8c16
      Lee Schermerhorn authored
      get_vma_policy() is not handling fallback to task policy correctly when the
      get_policy() vm_op returns NULL.  The NULL overwrites the 'pol' variable that
      was holding the fallback task mempolicy.  So, it was falling back directly to
      system default policy.
      
      Fix get_vma_policy() to use only non-NULL policy returned from the vma
      get_policy op.
      
      shm_get_policy() was falling back to current task's mempolicy if the "backing
      file system" [tmpfs vs hugetlbfs] does not support the get_policy vm_op and
      the vma policy is null.  This is incorrect for show_numa_maps() which is
      likely querying the numa_maps of some task other than current.  Remove this
      fallback.
      Signed-off-by: default avatarLee Schermerhorn <lee.schermerhorn@hp.com>
      Cc: Christoph Lameter <clameter@sgi.com>
      Cc: David Rientjes <rientjes@google.com>
      Cc: Mel Gorman <mel@csn.ul.ie>
      Cc: Andi Kleen <ak@suse.de>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      ae4d8c16
  18. 11 Mar, 2008 1 commit
    • Lee Schermerhorn's avatar
      mempolicy: fix reference counting bugs · 69682d85
      Lee Schermerhorn authored
      Address 3 known bugs in the current memory policy reference counting method.
      I have a series of patches to rework the reference counting to reduce overhead
      in the allocation path.  However, that series will require testing in -mm once
      I repost it.
      
      1) alloc_page_vma() does not release the extra reference taken for
         vma/shared mempolicy when the mode == MPOL_INTERLEAVE.  This can result in
         leaking mempolicy structures.  This is probably occurring, but not being
         noticed.
      
         Fix:  add the conditional release of the reference.
      
      2) hugezonelist unconditionally releases a reference on the mempolicy when
         mode == MPOL_INTERLEAVE.  This can result in decrementing the reference
         count for system default policy [should have no ill effect] or premature
         freeing of task policy.  If this occurred, the next allocation using task
         mempolicy would use the freed structure and probably BUG out.
      
         Fix:  add the necessary check to the release.
      
      3) The current reference counting method assumes that vma 'get_policy()'
         methods automatically add an extra reference a non-NULL returned mempolicy.
          This is true for shmem_get_policy() used by tmpfs mappings, including
         regular page shm segments.  However, SHM_HUGETLB shm's, backed by
         hugetlbfs, just use the vma policy without the extra reference.  This
         results in freeing of the vma policy on the first allocation, with reuse of
         the freed mempolicy structure on subsequent allocations.
      
         Fix: Rather than add another condition to the conditional reference
         release, which occur in the allocation path, just add a reference when
         returning the vma policy in shm_get_policy() to match the assumptions.
      Signed-off-by: default avatarLee Schermerhorn <lee.schermerhorn@hp.com>
      Cc: Greg KH <greg@kroah.com>
      Cc: Andi Kleen <ak@suse.de>
      Cc: Christoph Lameter <clameter@sgi.com>
      Cc: Mel Gorman <mel@csn.ul.ie>
      Cc: David Rientjes <rientjes@google.com>
      Cc: <eric.whitney@hp.com>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      69682d85
  19. 08 Feb, 2008 3 commits
    • Pierre Peiffer's avatar
      IPC: consolidate sem_exit_ns(), msg_exit_ns() and shm_exit_ns() · 01b8b07a
      Pierre Peiffer authored
      sem_exit_ns(), msg_exit_ns() and shm_exit_ns() are all called when an
      ipc_namespace is released to free all ipcs of each type.  But in fact, they
      do the same thing: they loop around all ipcs to free them individually by
      calling a specific routine.
      
      This patch proposes to consolidate this by introducing a common function,
      free_ipcs(), that do the job.  The specific routine to call on each
      individual ipcs is passed as parameter.  For this, these ipc-specific
      'free' routines are reworked to take a generic 'struct ipc_perm' as
      parameter.
      Signed-off-by: default avatarPierre Peiffer <pierre.peiffer@bull.net>
      Cc: Cedric Le Goater <clg@fr.ibm.com>
      Cc: Pavel Emelyanov <xemul@openvz.org>
      Cc: Nadia Derbey <Nadia.Derbey@bull.net>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      01b8b07a
    • Pierre Peiffer's avatar
      IPC: make struct ipc_ids static in ipc_namespace · ed2ddbf8
      Pierre Peiffer authored
      Each ipc_namespace contains a table of 3 pointers to struct ipc_ids (3 for
      msg, sem and shm, structure used to store all ipcs) These 'struct ipc_ids'
      are dynamically allocated for each icp_namespace as the ipc_namespace
      itself (for the init namespace, they are initialized with pointers to
      static variables instead)
      
      It is so for historical reason: in fact, before the use of idr to store the
      ipcs, the ipcs were stored in tables of variable length, depending of the
      maximum number of ipc allowed.  Now, these 'struct ipc_ids' have a fixed
      size.  As they are allocated in any cases for each new ipc_namespace, there
      is no gain of memory in having them allocated separately of the struct
      ipc_namespace.
      
      This patch proposes to make this table static in the struct ipc_namespace.
      Thus, we can allocate all in once and get rid of all the code needed to
      allocate and free these ipc_ids separately.
      Signed-off-by: default avatarPierre Peiffer <pierre.peiffer@bull.net>
      Acked-by: default avatarCedric Le Goater <clg@fr.ibm.com>
      Cc: Pavel Emelyanov <xemul@openvz.org>
      Cc: Nadia Derbey <Nadia.Derbey@bull.net>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      ed2ddbf8
    • Pavel Emelyanov's avatar
      namespaces: move the IPC namespace under IPC_NS option · ae5e1b22
      Pavel Emelyanov authored
      Currently the IPC namespace management code is spread over the ipc/*.c files.
      I moved this code into ipc/namespace.c file which is compiled out when needed.
      
      The linux/ipc_namespace.h file is used to store the prototypes of the
      functions in namespace.c and the stubs for NAMESPACES=n case.  This is done
      so, because the stub for copy_ipc_namespace requires the knowledge of the
      CLONE_NEWIPC flag, which is in sched.h.  But the linux/ipc.h file itself in
      included into many many .c files via the sys.h->sem.h sequence so adding the
      sched.h into it will make all these .c depend on sched.h which is not that
      good.  On the other hand the knowledge about the namespaces stuff is required
      in 4 .c files only.
      
      Besides, this patch compiles out some auxiliary functions from ipc/sem.c,
      msg.c and shm.c files.  It turned out that moving these functions into
      namespaces.c is not that easy because they use many other calls and macros
      from the original file.  Moving them would make this patch complicated.  On
      the other hand all these functions can be consolidated, so I will send a
      separate patch doing this a bit later.
      Signed-off-by: default avatarPavel Emelyanov <xemul@openvz.org>
      Acked-by: default avatarSerge Hallyn <serue@us.ibm.com>
      Cc: Cedric Le Goater <clg@fr.ibm.com>
      Cc: "Eric W. Biederman" <ebiederm@xmission.com>
      Cc: Herbert Poetzl <herbert@13thfloor.at>
      Cc: Kirill Korotaev <dev@sw.ru>
      Cc: Sukadev Bhattiprolu <sukadev@us.ibm.com>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      ae5e1b22
  20. 06 Feb, 2008 1 commit
    • Pierre Peiffer's avatar
      IPC: fix error check in all new xxx_lock() and xxx_exit_ns() functions · b1ed88b4
      Pierre Peiffer authored
      In the new implementation of the [sem|shm|msg]_lock[_check]() routines, we
      use the return value of ipc_lock() in container_of() without any check.
      But ipc_lock may return a errcode.  The use of this errcode in
      container_of() may alter this errcode, and we don't want this.
      
      And in xxx_exit_ns, the pointer return by idr_find is of type 'struct
      kern_ipc_per'...
      
      Today, the code will work as is because the member used in these
      container_of() is the first member of its container (offset == 0), the
      errcode isn't changed then.  But in the general case, we can't count on
      this assumption and this may lead later to a real bug if we don't correct
      this.
      
      Again, the proposed solution is simple and correct.  But, as pointed by
      Nadia, with this solution, the same check will be done several times (in
      all sub-callers...), what is not very funny/optimal...
      Signed-off-by: default avatarPierre Peiffer <pierre.peiffer@bull.net>
      Cc: Nadia Derbey <Nadia.Derbey@bull.net>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      b1ed88b4
  21. 19 Oct, 2007 10 commits