Simplify memory allocation by sticking with a single API.  GSlice
is not that fast anyway (tcmalloc/jemalloc are better).
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

Let dataplane allocate different region for the desc/avail/used
ring regions.
Take VIRTIO_RING_F_EVENT_IDX into account to increase the used/avail
rings accordingly.

[Fix 32-bit builds by changing 16lx format specifier to HWADDR_PRIx.
--Stefan]
Signed-off-by: NPierre Morel <pmorel@linux.vnet.ibm.com>
Tested-by: NGreg Kurz <gkurz@linux.vnet.ibm.com>
Signed-off-by: NGreg Kurz <gkurz@linux.vnet.ibm.com>
Message-id: 1441625636-23773-1-git-send-email-pmorel@linux.vnet.ibm.com
(changed __virtio16 into uint16_t,
 map descriptor table and available ring read-only)
Signed-off-by: NGreg Kurz <gkurz@linux.vnet.ibm.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

The raw-posix block driver implements Linux AIO batching so multiple
requests can be submitted with a single io_submit(2) system call.
Batching is currently only used by virtio-scsi and
virtio-blk-data-plane.

Enable batching for regular virtio-blk so the number of io_submit(2)
system calls is reduced for workloads with queue depth > 1.

In 4KB random read performance tests with queue depth 32, the CPU
utilization on the host is reduced by 9.4%.  The fio job is as follows:

  [global]
  bs=4k
  ioengine=libaio
  iodepth=32
  direct=1
  sync=0
  time_based=1
  runtime=30
  clocksource=gettimeofday
  ramp_time=5

  [job1]
  rw=randread
  filename=/dev/vdb
  size=4096M
  write_bw_log=fio
  write_iops_log=fio
  write_lat_log=fio
  log_avg_msec=1000

This benchmark was run on an raw image on LVM.  The disk was an SSD
drive and -drive cache=none,aio=native was used.
Tested-by: NPradeep Surisetty <psuriset@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NFam Zheng <famz@redhat.com>

Commit 19109131 disabled the sdhci-pci support because it used
drive_get_next().  This patch reenables sdhci-pci and changes it to
pass the drive via a qdev property - for example:
 -device sdhci-pci,drive=drive0 -drive id=drive0,if=sd,file=myimage
Reviewed-by: NStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NKevin O'Connor <kevin@koconnor.net>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

# gpg: Signature made Fri 09 Oct 2015 10:15:13 BST using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"

* remotes/stefanha/tags/tracing-pull-request:
  trace: remove malloc tracing
  docs: update the usage example of "dtrace" backend in tracing.txt
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

trivial patches for 2015-10-08

# gpg: Signature made Thu 08 Oct 2015 17:51:05 BST using RSA key ID A4C3D7DB
# gpg: Good signature from "Michael Tokarev <mjt@tls.msk.ru>"
# gpg:                 aka "Michael Tokarev <mjt@corpit.ru>"
# gpg:                 aka "Michael Tokarev <mjt@debian.org>"

* remotes/mjt/tags/pull-trivial-patches-2015-10-08:
  tests: Unique test path for /string-visitor/output
  linux-user: Remove type casts to union type
  linux-user: Use g_new() & friends where that makes obvious sense
  rocker: Use g_new() & friends where that makes obvious sense
  .travis.yml: Run make check for all targets, not just some
  hw: char: Remove unnecessary variable
  hw: timer: Remove unnecessary variable
  qapi: add missing @
  MAINTAINERS: Add NSIS file for W32, W64 hosts
  target-ppc: Remove unnecessary variable
  target-microblaze: Remove unnecessary variable
  s/cpu_get_real_ticks/cpu_get_host_ticks/
  pc: check for underflow in load_linux
  pci-assign: do not include sys/io.h
  block/ssh: remove dead code
  imx_serial: Generate interrupt on tx empty if enabled
  sdhci: Change debug prints to compile unconditionally
  sdhci: use PRIx64 for uint64_t type
  Add .dir-locals.el file to configure emacs coding style
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

The malloc vtable is not supported anymore in glib, because it broke
when constructors called g_malloc.  Remove tracing of g_malloc,
g_realloc and g_free calls.

Note that, for systemtap users, glib also provides tracepoints
glib.mem_alloc, glib.mem_free, glib.mem_realloc, glib.slice_alloc
and glib.slice_free.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
Reviewed-by: NAlberto Garcia <berto@igalia.com>
Message-id: 1442417924-25831-1-git-send-email-pbonzini@redhat.com
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

The usage example of dtrace is quite ancient, We have tracetool.py with
different parameters instead of the original tracetool shell script for
a long time, So update the old information.
Signed-off-by: NLin Ma <lma@suse.com>
Message-id: 1441954730-17341-1-git-send-email-lma@suse.com
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

Newer GLib's want unique test paths, and thus moan at dupes.
(Seen on Fedora 23 which has glib 2.46)

Uniquify the paths.
Signed-off-by: NDr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Casting to a union type is a gcc (and clang) extension. Other compilers
might not support it. This is not a problem today, but the type casts
can be removed easily. Smatch now no longer complains like before:

linux-user/syscall.c:3190:18: warning: cast to non-scalar
linux-user/syscall.c:7348:44: warning: cast to non-scalar

Cc: Riku Voipio <riku.voipio@iki.fi>
Signed-off-by: NStefan Weil <sw@weilnetz.de>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

g_new(T, n) is neater than g_malloc(sizeof(T) * n).  It's also safer,
for two reasons.  One, it catches multiplication overflowing size_t.
Two, it returns T * rather than void *, which lets the compiler catch
more type errors.

This commit only touches allocations with size arguments of the form
sizeof(T).  Same Coccinelle semantic patch as in commit b45c03f5.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NStefan Weil <sw@weilnetz.de>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

g_new(T, n) is neater than g_malloc(sizeof(T) * n).  It's also safer,
for two reasons.  One, it catches multiplication overflowing size_t.
Two, it returns T * rather than void *, which lets the compiler catch
more type errors.

This commit only touches allocations with size arguments of the form
sizeof(T).  Same Coccinelle semantic patchas in commit b45c03f5.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Acked-by: NJiri Pirko <jiri@resnulli.us>
Reviewed-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

ed173cb7 ".travis.yml: remove "make check" from main matrix" stopped running
make check for all the Travis build targets for various reasons.  It
continued to run make check on one Travis build, which builds for a big
list of all (? nearly all) our supported softmmu targets.

Unfortunately, due to a spacing / quoting error it only actually builds for
the alpha, arm, aarch64 and cris targets.  Specifically, the list of
targets is split over several lines.  Even with YAML folding, this will
leave spaces in the list, meaning $TARGETS won't have the value we need.

I had a look at the YAML spec and I couldn't quickly see a way of splitting
the list so that it doesn't end up with spaces, so this patch fixes the
problem by putting the whole list on one huge line.
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Compress lines and remove the variable.
Signed-off-by: NShraddha Barke <shraddha.6596@gmail.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Compress lines and remove the variable.
Signed-off-by: NShraddha Barke <shraddha.6596@gmail.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Signed-off-by: NMarc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

The NSIS installer configuration is maintained by me.
Signed-off-by: NStefan Weil <sw@weilnetz.de>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Compress lines and remove the variable.
Signed-off-by: NShraddha Barke <shraddha.6596@gmail.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Compress lines and remove the variable.
Signed-off-by: NShraddha Barke <shraddha.6596@gmail.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

This should help clarify the purpose of the function that returns
the host system's CPU cycle count.
Signed-off-by: NChristopher Covington <cov@codeaurora.org>
Acked-by: NPaolo Bonzini <pbonzini@redhat.com>
ppc portion
Acked-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

If (setup_size+1)*512 is small enough, kernel_size -= setup_size can allocate
a huge amount of memory.  Avoid that.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

This file does not exist on bionic libc and the functions it defines
are in fact not used by pci-assign.c.  Remove it.
Reported-by: NHoucheng Lin <houcheng@gmail.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

The "err" label cannot be reached with qp != NULL.  Remove the free-ing
of qp and avoid future regressions by removing the initializer.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
ACKed-by: NRichard W.M. Jones <rjones@redhat.com>
Reviewed-by: NFam Zheng <famz@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Generate an interrupt if the tx buffer is empty and the tx empty interrupt
is enabled. This fixes a problem seen when running a Linux image since
Linux commit 55c3cb1358e ("serial: imx: remove unneeded imx_transmit_buffer()
from imx_start_tx()"). Linux now waits for the tx empty interrupt before
starting to send data, causing transmit stalls until there is an interrupt
for another reason.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Reviewed-by: NPeter Crosthwaite <crosthwaite.peter@gmail.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Conditional compilation hides few type mismatch warnings, fix it to
compile unconditionally.
Signed-off-by: NSai Pavan Boddu <saipava@xilinx.com>
Suggested-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NPeter Crosthwaite <crosthwaite.peter@gmail.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Fix compile time warnings, because of type mismatch for unsigned long
long type.
Signed-off-by: NSai Pavan Boddu <saipava@xilinx.com>
Reviewed-by: NPeter Crosthwaite <crosthwaite.peter@gmail.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Some default emacs setups indent by 2 spaces and uses tabs
which is counter to the QEMU coding style rules. Adding a
.dir-locals.el file in the top level of the GIT repo will
inform emacs about the QEMU coding style, and so assist
contributors in avoiding common style mistakes before
they submit patches.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NMarkus Armbruster <armbru@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

VFIO updates 2015-10-07

 - Change platform device IRQ setup sequence for compatibility
   with upcoming IRQ forwarding (Eric Auger)
 - Extensions to support vfio-pci devices on spapr-pci-host-bridge
   (David Gibson) [clang problem patch dropped]

# gpg: Signature made Wed 07 Oct 2015 16:30:52 BST using RSA key ID 3BB08B22
# gpg: Good signature from "Alex Williamson <alex.williamson@redhat.com>"
# gpg:                 aka "Alex Williamson <alex@shazbot.org>"
# gpg:                 aka "Alex Williamson <alwillia@redhat.com>"
# gpg:                 aka "Alex Williamson <alex.l.williamson@gmail.com>"

* remotes/awilliam/tags/vfio-update-20151007.0:
  vfio: Allow hotplug of containers onto existing guest IOMMU mappings
  memory: Allow replay of IOMMU mapping notifications
  vfio: Record host IOMMU's available IO page sizes
  vfio: Check guest IOVA ranges against host IOMMU capabilities
  vfio: Generalize vfio_listener_region_add failure path
  vfio: Remove unneeded union from VFIOContainer
  hw/vfio/platform: do not set resamplefd for edge-sensitive IRQS
  hw/vfio/platform: change interrupt/unmask fields into pointer
  hw/vfio/platform: irqfd setup sequence update
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Do away with TB retranslation

# gpg: Signature made Wed 07 Oct 2015 10:42:08 BST using RSA key ID 4DD0279B
# gpg: Good signature from "Richard Henderson <rth7680@gmail.com>"
# gpg:                 aka "Richard Henderson <rth@redhat.com>"
# gpg:                 aka "Richard Henderson <rth@twiddle.net>"

* remotes/rth/tags/pull-tcg-20151007: (26 commits)
  tcg: Adjust CODE_GEN_AVG_BLOCK_SIZE
  tcg: Check for overflow via highwater mark
  tcg: Allocate a guard page after code_gen_buffer
  tcg: Emit prologue to the beginning of code_gen_buffer
  tcg: Remove tcg_gen_code_search_pc
  tcg: Remove gen_intermediate_code_pc
  tcg: Save insn data and use it in cpu_restore_state_from_tb
  tcg: Pass data argument to restore_state_to_opc
  tcg: Add TCG_MAX_INSNS
  target-*: Drop cpu_gen_code define
  tcg: Merge cpu_gen_code into tb_gen_code
  target-sparc: Add npc state to insn_start
  target-sparc: Remove gen_opc_jump_pc
  target-sparc: Split out gen_branch_n
  target-sparc: Tidy gen_branch_a interface
  target-cris: Mirror gen_opc_pc into insn_start
  target-sh4: Add flags state to insn_start
  target-s390x: Add cc_op state to insn_start
  target-mips: Add delayed branch state to insn_start
  target-i386: Add cc_op state to insn_start
  ...
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Collected patches

# gpg: Signature made Wed 07 Oct 2015 10:30:17 BST using RSA key ID 4DD0279B
# gpg: Good signature from "Richard Henderson <rth7680@gmail.com>"
# gpg:                 aka "Richard Henderson <rth@redhat.com>"
# gpg:                 aka "Richard Henderson <rth@twiddle.net>"

* remotes/rth/tags/pull-tile-20151007:
  target-tilegx: Support iret instruction and related special registers
  target-tilegx: Use TILEGX_EXCP_OPCODE_UNKNOWN and TILEGX_EXCP_OPCODE_UNIMPLEMENTED correctly
  target-tilegx: Implement v2mults instruction
  target-tilegx: Implement v?int_* instructions.
  target-tilegx: Implement v2sh* instructions
  target-tilegx: Handle nofault prefetch instructions
  target-tilegx: Fix a typo for mnemonic about "ld_add"
  target-tilegx: Use TILEGX_EXCP_SIGNAL instead of TILEGX_EXCP_SEGV
  target-tilegx: Decode ill pseudo-instructions
  linux-user/tilegx: Implement tilegx signal features
  linux-user/syscall_defs.h: Sync the latest si_code from Linux kernel
  target-tilegx: Let x1 pipe process bpt instruction only
  target-tilegx: Implement complex multiply instructions
  target-tilegx: Implement table index instructions
  target-tilegx: Implement crc instructions
  target-tilegx: Implement v1multu instruction
  target-tilegx: Implement v*add and v*sub instructions
  target-tilegx: Implement v*shl, v*shru, and v*shrs instructions
  target-tilegx: Tidy simd_helper.c
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

NUMA queue, 2015-10-06

# gpg: Signature made Tue 06 Oct 2015 20:53:42 BST using RSA key ID 984DC5A6
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>"

* remotes/ehabkost/tags/numa-pull-request:
  pc-dimm: Fail realization for invalid nodes in non-NUMA config
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

At present, the "average" guestimate of TB size is way too small, leading
to many unused entries in the pre-allocated TB array.  For a guest with 1GB
ram, we're currently allocating 256MB for the array.

Survey arm, alpha, aarch64, ppc, sparc, i686, x86_64 guests running on
x86_64 and ppc64 hosts and select a new average.  The size of the array
drops to 81MB with no more flushing than before.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

We currently pre-compute an worst case code size for any TB, which
works out to be 122kB.  Since the average TB size is near 1kB, this
wastes quite a lot of storage.

Instead, check for overflow in between generating code for each opcode.
The overhead of the check isn't measurable and wastage is minimized.
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

This will catch any overflow of the buffer.

Add a native win32 alternative for alloc_code_gen_buffer;
remove the malloc alternative.
Signed-off-by: NRichard Henderson <rth@twiddle.net>

By putting the prologue at the end, we risk overwriting the
prologue should our estimate of maximum TB size.  Given the
two different placements of the call to tcg_prologue_init,
move the high water mark computation into tcg_prologue_init.
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

It's no longer used, so tidy up everything reached by it.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

It is no longer used, so tidy up everything reached by it.
This includes the gen_opc_* arrays, the search_pc parameter
and the inline gen_intermediate_code_internal functions.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

We can now restore state without retranslation.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

The gen_opc_* arrays are already redundant with the data stored in
the insn_start arguments.  Transition restore_state_to_opc to use
data from the latter.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Adjust all translators to respect it.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>