Skip to content
  1. Oct 19, 2015
  3. Sep 12, 2015
  5. Sep 11, 2015
    • Mathieu Desnoyers's avatar
      sys_membarrier(): system-wide memory barrier (generic, x86) · 5b25b13a
      Mathieu Desnoyers authored 
      Here is an implementation of a new system call, sys_membarrier(), which
executes a memory barrier on all threads running on the system.  It is
implemented by calling synchronize_sched().  It can be used to
distribute the cost of user-space memory barriers asymmetrically by
transforming pairs of memory barriers into pairs consisting of
sys_membarrier() and a compiler barrier.  For synchronization primitives
that distinguish between read-side and write-side (e.g.  userspace RCU
[1], rwlocks), the read-side can be accelerated significantly by moving
the bulk of the memory barrier overhead to the write-side.

The existing applications of which I am aware that would be improved by
this system call are as follows:

* Through Userspace RCU library (http://urcu.so)
  - DNS server (Knot DNS) https://www.knot-dns.cz/
  - Network sniffer (http://netsniff-ng.org/)
  - Distributed object storage (https://sheepdog.github.io/sheepdog/)
  - User-space tracing (http://lttng.org)
  - Network storage system (https://www.gluster.org/)
  - Virtual routers (https://events.linuxfoundation.org/sites/events/files/slides/DPDK_RCU_0MQ.pdf)
  - Financial software (https://lkml.org/lkml/2015/3/23/189)

Those projects use RCU in userspace to increase read-side speed and
scalability compared to locking.  Especially in the case of RCU used by
libraries, sys_membarrier can speed up the read-side by moving the bulk of
the memory barrier cost to synchronize_rcu().

* Direct users of sys_membarrier
  - core dotnet garbage collector (https://github.com/dotnet/coreclr/issues/198)

Microsoft core dotnet GC developers are planning to use the mprotect()
side-effect of issuing memory barriers through IPIs as a way to implement
Windows FlushProcessWriteBuffers() on Linux.  They are referring to
sys_membarrier in their github thread, specifically stating that
sys_membarrier() is what they are looking for.

To explain the benefit of this scheme, let's introduce two example threads:

Thread A (non-frequent, e.g. executing liburcu synchronize_rcu())
Thread B (frequent, e.g. executing liburcu
rcu_read_lock()/rcu_read_unlock())

In a scheme where all smp_mb() in thread A are ordering memory accesses
with respect to smp_mb() present in Thread B, we can change each
smp_mb() within Thread A into calls to sys_membarrier() and each
smp_mb() within Thread B into compiler barriers "barrier()".

Before the change, we had, for each smp_mb() pairs:

Thread A                    Thread B
previous mem accesses       previous mem accesses
smp_mb()                    smp_mb()
following mem accesses      following mem accesses

After the change, these pairs become:

Thread A                    Thread B
prev mem accesses           prev mem accesses
sys_membarrier()            barrier()
follow mem accesses         follow mem accesses

As we can see, there are two possible scenarios: either Thread B memory
accesses do not happen concurrently with Thread A accesses (1), or they
do (2).

1) Non-concurrent Thread A vs Thread B accesses:

Thread A                    Thread B
prev mem accesses
sys_membarrier()
follow mem accesses
                            prev mem accesses
                            barrier()
                            follow mem accesses

In this case, thread B accesses will be weakly ordered. This is OK,
because at that point, thread A is not particularly interested in
ordering them with respect to its own accesses.

2) Concurrent Thread A vs Thread B accesses

Thread A                    Thread B
prev mem accesses           prev mem accesses
sys_membarrier()            barrier()
follow mem accesses         follow mem accesses

In this case, thread B accesses, which are ensured to be in program
order thanks to the compiler barrier, will be "upgraded" to full
smp_mb() by synchronize_sched().

* Benchmarks

On Intel Xeon E5405 (8 cores)
(one thread is calling sys_membarrier, the other 7 threads are busy
looping)

1000 non-expedited sys_membarrier calls in 33s =3D 33 milliseconds/call.

* User-space user of this system call: Userspace RCU library

Both the signal-based and the sys_membarrier userspace RCU schemes
permit us to remove the memory barrier from the userspace RCU
rcu_read_lock() and rcu_read_unlock() primitives, thus significantly
accelerating them. These memory barriers are replaced by compiler
barriers on the read-side, and all matching memory barriers on the
write-side are turned into an invocation of a memory barrier on all
active threads in the process. By letting the kernel perform this
synchronization rather than dumbly sending a signal to every process
threads (as we currently do), we diminish the number of unnecessary wake
ups and only issue the memory barriers on active threads. Non-running
threads do not need to execute such barrier anyway, because these are
implied by the scheduler context switches.

Results in liburcu:

Operations in 10s, 6 readers, 2 writers:

memory barriers in reader:    1701557485 reads, 2202847 writes
signal-based scheme:          9830061167 reads,    6700 writes
sys_membarrier:               9952759104 reads,     425 writes
sys_membarrier (dyn. check):  7970328887 reads,     425 writes

The dynamic sys_membarrier availability check adds some overhead to
the read-side compared to the signal-based scheme, but besides that,
sys_membarrier slightly outperforms the signal-based scheme. However,
this non-expedited sys_membarrier implementation has a much slower grace
period than signal and memory barrier schemes.

Besides diminishing the number of wake-ups, one major advantage of the
membarrier system call over the signal-based scheme is that it does not
need to reserve a signal. This plays much more nicely with libraries,
and with processes injected into for tracing purposes, for which we
cannot expect that signals will be unused by the application.

An expedited version of this system call can be added later on to speed
up the grace period. Its implementation will likely depend on reading
the cpu_curr()->mm without holding each CPU's rq lock.

This patch adds the system call to x86 and to asm-generic.

[1] http://urcu.so



membarrier(2) man page:

MEMBARRIER(2)              Linux Programmer's Manual             MEMBARRIER(2)

NAME
       membarrier - issue memory barriers on a set of threads

SYNOPSIS
       #include <linux/membarrier.h>

       int membarrier(int cmd, int flags);

DESCRIPTION
       The cmd argument is one of the following:

       MEMBARRIER_CMD_QUERY
              Query  the  set  of  supported commands. It returns a bitmask of
              supported commands.

       MEMBARRIER_CMD_SHARED
              Execute a memory barrier on all threads running on  the  system.
              Upon  return from system call, the caller thread is ensured that
              all running threads have passed through a state where all memory
              accesses  to  user-space  addresses  match program order between
              entry to and return from the system  call  (non-running  threads
              are de facto in such a state). This covers threads from all pro=E2=80=90
              cesses running on the system.  This command returns 0.

       The flags argument needs to be 0. For future extensions.

       All memory accesses performed  in  program  order  from  each  targeted
       thread is guaranteed to be ordered with respect to sys_membarrier(). If
       we use the semantic "barrier()" to represent a compiler barrier forcing
       memory  accesses  to  be performed in program order across the barrier,
       and smp_mb() to represent explicit memory barriers forcing full  memory
       ordering  across  the barrier, we have the following ordering table for
       each pair of barrier(), sys_membarrier() and smp_mb():

       The pair ordering is detailed as (O: ordered, X: not ordered):

                              barrier()   smp_mb() sys_membarrier()
              barrier()          X           X            O
              smp_mb()           X           O            O
              sys_membarrier()   O           O            O

RETURN VALUE
       On success, these system calls return zero.  On error, -1 is  returned,
       and errno is set appropriately. For a given command, with flags
       argument set to 0, this system call is guaranteed to always return the
       same value until reboot.

ERRORS
       ENOSYS System call is not implemented.

       EINVAL Invalid arguments.

Linux                             2015-04-15                     MEMBARRIER(2)

Signed-off-by: default avatarMathieu Desnoyers <mathieu.desnoyers@efficios.com>
Reviewed-by: default avatarPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: default avatarJosh Triplett <josh@joshtriplett.org>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Nicholas Miell <nmiell@comcast.net>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Alan Cox <gnomes@lxorguk.ukuu.org.uk>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Pranith Kumar <bobby.prani@gmail.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Shuah Khan <shuahkh@osg.samsung.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      5b25b13a
  7. Sep 10, 2015
    • Christoph Hellwig's avatar
      dma-mapping: consolidate dma_set_mask · 452e06af
      Christoph Hellwig authored 
      

Almost everyone implements dma_set_mask the same way, although some time
that's hidden in ->set_dma_mask methods.

This patch consolidates those into a common implementation that either
calls ->set_dma_mask if present or otherwise uses the default
implementation.  Some architectures used to only call ->set_dma_mask
after the initial checks, and those instance have been fixed to do the
full work.  h8300 implemented dma_set_mask bogusly as a no-ops and has
been fixed.

Unfortunately some architectures overload unrelated semantics like changing
the dma_ops into it so we still need to allow for an architecture override
for now.

[jcmvbkbc@gmail.com: fix xtensa]
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Chris Metcalf <cmetcalf@ezchip.com>
Cc: Guan Xuetao <gxt@mprc.pku.edu.cn>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: default avatarMax Filippov <jcmvbkbc@gmail.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      452e06af
    • Christoph Hellwig's avatar
      dma-mapping: consolidate dma_supported · ee196371
      Christoph Hellwig authored 
      

Most architectures just call into ->dma_supported, but some also return 1
if the method is not present, or 0 if no dma ops are present (although
that should never happeb). Consolidate this more broad version into
common code.

Also fix h8300 which inorrectly always returned 0, which would have been
a problem if it's dma_set_mask implementation wasn't a similarly buggy
noop.

As a few architectures have much more elaborate implementations, we
still allow for arch overrides.

[jcmvbkbc@gmail.com: fix xtensa]
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Chris Metcalf <cmetcalf@ezchip.com>
Cc: Guan Xuetao <gxt@mprc.pku.edu.cn>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: default avatarMax Filippov <jcmvbkbc@gmail.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      ee196371
    • Christoph Hellwig's avatar
      dma-mapping: cosolidate dma_mapping_error · efa21e43
      Christoph Hellwig authored 
      

Currently there are three valid implementations of dma_mapping_error:

 (1) call ->mapping_error
 (2) check for a hardcoded error code
 (3) always return 0

This patch provides a common implementation that calls ->mapping_error
if present, then checks for DMA_ERROR_CODE if defined or otherwise
returns 0.

[jcmvbkbc@gmail.com: fix xtensa]
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Chris Metcalf <cmetcalf@ezchip.com>
Cc: Guan Xuetao <gxt@mprc.pku.edu.cn>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: default avatarMax Filippov <jcmvbkbc@gmail.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      efa21e43
    • Christoph Hellwig's avatar
      dma-mapping: consolidate dma_{alloc,free}_noncoherent · 1e893752
      Christoph Hellwig authored 
      

Most architectures do not support non-coherent allocations and either
define dma_{alloc,free}_noncoherent to their coherent versions or stub
them out.

Openrisc uses dma_{alloc,free}_attrs to implement them, and only Mips
implements them directly.

This patch moves the Openrisc version to common code, and handles the
DMA_ATTR_NON_CONSISTENT case in the mips dma_map_ops instance.

Note that actual non-coherent allocations require a dma_cache_sync
implementation, so if non-coherent allocations didn't work on
an architecture before this patch they still won't work after it.

[jcmvbkbc@gmail.com: fix xtensa]
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Chris Metcalf <cmetcalf@ezchip.com>
Cc: Guan Xuetao <gxt@mprc.pku.edu.cn>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: default avatarMax Filippov <jcmvbkbc@gmail.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      1e893752
    • Christoph Hellwig's avatar
      dma-mapping: consolidate dma_{alloc,free}_{attrs,coherent} · 6894258e
      Christoph Hellwig authored 
      

Since 2009 we have a nice asm-generic header implementing lots of DMA API
functions for architectures using struct dma_map_ops, but unfortunately
it's still missing a lot of APIs that all architectures still have to
duplicate.

This series consolidates the remaining functions, although we still need
arch opt outs for two of them as a few architectures have very
non-standard implementations.

This patch (of 5):

The coherent DMA allocator works the same over all architectures supporting
dma_map operations.

This patch consolidates them and converges the minor differences:

 - the debug_dma helpers are now called from all architectures, including
   those that were previously missing them
 - dma_alloc_from_coherent and dma_release_from_coherent are now always
   called from the generic alloc/free routines instead of the ops
   dma-mapping-common.h always includes dma-coherent.h to get the defintions
   for them, or the stubs if the architecture doesn't support this feature
 - checks for ->alloc / ->free presence are removed.  There is only one
   magic instead of dma_map_ops without them (mic_dma_ops) and that one
   is x86 only anyway.

Besides that only x86 needs special treatment to replace a default devices
if none is passed and tweak the gfp_flags.  An optional arch hook is provided
for that.

[linux@roeck-us.net: fix build]
[jcmvbkbc@gmail.com: fix xtensa]
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Chris Metcalf <cmetcalf@ezchip.com>
Cc: Guan Xuetao <gxt@mprc.pku.edu.cn>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: default avatarGuenter Roeck <linux@roeck-us.net>
Signed-off-by: default avatarMax Filippov <jcmvbkbc@gmail.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      6894258e
    • Oleg Nesterov's avatar
      mm, mpx: add "vm_flags_t vm_flags" arg to do_mmap_pgoff() · 1fcfd8db
      Oleg Nesterov authored 
      

Add the additional "vm_flags_t vm_flags" argument to do_mmap_pgoff(),
rename it to do_mmap(), and re-introduce do_mmap_pgoff() as a simple
wrapper on top of do_mmap().  Perhaps we should update the callers of
do_mmap_pgoff() and kill it later.

This way mpx_mmap() can simply call do_mmap(vm_flags => VM_MPX) and do not
play with vm internals.

After this change mmap_region() has a single user outside of mmap.c,
arch/tile/mm/elf.c:arch_setup_additional_pages().  It would be nice to
change arch/tile/ and unexport mmap_region().

[kirill@shutemov.name: fix build]
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: default avatarOleg Nesterov <oleg@redhat.com>
Acked-by: default avatarDave Hansen <dave.hansen@linux.intel.com>
Tested-by: default avatarDave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: default avatarKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      1fcfd8db
    • Kirill A. Shutemov's avatar
      mm: mark most vm_operations_struct const · 7cbea8dc
      Kirill A. Shutemov authored 
      

With two exceptions (drm/qxl and drm/radeon) all vm_operations_struct
structs should be constant.

Signed-off-by: default avatarKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: default avatarOleg Nesterov <oleg@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      7cbea8dc
    • Yinghai Lu's avatar
      lib/decompressors: use real out buf size for gunzip with kernel · 2d3862d2
      Yinghai Lu authored 
      

When loading x86 64bit kernel above 4GiB with patched grub2, got kernel
gunzip error.

| early console in decompress_kernel
| decompress_kernel:
|       input: [0x807f2143b4-0x807ff61aee]
|      output: [0x807cc00000-0x807f3ea29b] 0x027ea29c: output_len
| boot via startup_64
| KASLR using RDTSC...
|  new output: [0x46fe000000-0x470138cfff] 0x0338d000: output_run_size
|  decompress: [0x46fe000000-0x47007ea29b] <=== [0x807f2143b4-0x807ff61aee]
|
| Decompressing Linux... gz...
|
| uncompression error
|
| -- System halted

the new buffer is at 0x46fe000000ULL, decompressor_gzip is using
0xffffffb901ffffff as out_len.  gunzip in lib/zlib_inflate/inflate.c cap
that len to 0x01ffffff and decompress fails later.

We could hit this problem with crashkernel booting that uses kexec loading
kernel above 4GiB.

We have decompress_* support:
    1. inbuf[]/outbuf[] for kernel preboot.
    2. inbuf[]/flush() for initramfs
    3. fill()/flush() for initrd.
This bug only affect kernel preboot path that use outbuf[].

Add __decompress and take real out_buf_len for gunzip instead of guessing
wrong buf size.

Fixes: 1431574a (lib/decompressors: fix "no limit" output buffer length)
Signed-off-by: default avatarYinghai Lu <yinghai@kernel.org>
Cc: Alexandre Courbot <acourbot@nvidia.com>
Cc: Jon Medhurst <tixy@linaro.org>
Cc: Stephen Warren <swarren@wwwdotorg.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      2d3862d2