qemu.git
3 years agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Tue, 3 Jul 2018 08:49:20 +0000 (09:49 +0100)] 
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

pc, virtio: fixes

A couple of fixes to amd iommu, and a fix to virtio iommu.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Thu 28 Jun 2018 02:46:45 BST
# gpg:                using RSA key 281F0DB8D28D5469
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>"
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>"
# Primary key fingerprint: 0270 606B 6F3C DF3D 0B17  0970 C350 3912 AFBE 8E67
#      Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA  8A0D 281F 0DB8 D28D 5469

* remotes/mst/tags/for_upstream:
  virtio-rng: process pending requests on DRIVER_OK
  hw/i386: Fix AMDVI GATS and HATS encodings
  hw/i386: Fix IVHD entry length for AMD IOMMU

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agoMerge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging
Peter Maydell [Mon, 2 Jul 2018 18:07:19 +0000 (19:07 +0100)] 
Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

* IEC units series (Philippe)
* Hyper-V PV TLB flush (Vitaly)
* git archive detection (Daniel)
* host serial passthrough fix (David)
* NPT support for SVM emulation (Jan)
* x86 "info mem" and "info tlb" fix (Doug)

# gpg: Signature made Mon 02 Jul 2018 16:18:21 BST
# gpg:                using RSA key BFFBD25F78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>"
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>"
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* remotes/bonzini/tags/for-upstream: (50 commits)
  tcg: simplify !CONFIG_TCG handling of tb_invalidate_*
  i386/monitor.c: make addresses canonical for "info mem" and "info tlb"
  target-i386: Add NPT support
  serial: Open non-block
  bsd-user: Use the IEC binary prefix definitions
  linux-user: Use the IEC binary prefix definitions
  tests/crypto: Use the IEC binary prefix definitions
  vl: Use the IEC binary prefix definitions
  monitor: Use the IEC binary prefix definitions
  cutils: Do not include "qemu/units.h" directly
  hw/rdma: Use the IEC binary prefix definitions
  hw/virtio: Use the IEC binary prefix definitions
  hw/vfio: Use the IEC binary prefix definitions
  hw/sd: Use the IEC binary prefix definitions
  hw/usb: Use the IEC binary prefix definitions
  hw/net: Use the IEC binary prefix definitions
  hw/i386: Use the IEC binary prefix definitions
  hw/ppc: Use the IEC binary prefix definitions
  hw/mips: Use the IEC binary prefix definitions
  hw/mips/r4k: Constify params_size
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agoMerge remote-tracking branch 'remotes/rth/tags/pull-tcg-20180702' into staging
Peter Maydell [Mon, 2 Jul 2018 16:57:46 +0000 (17:57 +0100)] 
Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20180702' into staging

Assorted tlb and tb caching fixes

# gpg: Signature made Mon 02 Jul 2018 17:03:07 BST
# gpg:                using RSA key 64DF38E8AF7E215F
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>"
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* remotes/rth/tags/pull-tcg-20180702:
  cpu: Assert asidx_from_attrs return value in range
  accel/tcg: Avoid caching overwritten tlb entries
  accel/tcg: Don't treat invalid TLB entries as needing recheck
  accel/tcg: Correct "is this a TLB miss" check in get_page_addr_code()
  tcg: Define and use new tlb_hit() and tlb_hit_page() functions
  translate-all: fix locking of TBs whose two pages share the same physical page

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agoMerge remote-tracking branch 'remotes/kraxel/tags/seabios-1.11.2-20180702-pull-reques...
Peter Maydell [Mon, 2 Jul 2018 16:04:20 +0000 (17:04 +0100)] 
Merge remote-tracking branch 'remotes/kraxel/tags/seabios-1.11.2-20180702-pull-request' into staging

seabios: update to release 1.11.2, add/update configuration.
vgabios: remove (old unused lgpl'ed vgabios).
ramfb,bochs-display: use new vgabios roms.

# gpg: Signature made Mon 02 Jul 2018 16:42:06 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/seabios-1.11.2-20180702-pull-request:
  ramfb: enable vgabios
  bochs-display: enable vgabios
  seabios: update bios and vgabios binaries
  vgabios: remove submodule and build rules.
  seabios: enable ide dma
  seabios: add vga configs for bochs-display and ramfb
  seabios: update submodule to release 1.11.2

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agoramfb: enable vgabios
Gerd Hoffmann [Fri, 8 Jun 2018 10:38:47 +0000 (12:38 +0200)] 
ramfb: enable vgabios

Add vgabios binary to fw_cfg vgaroms.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
3 years agobochs-display: enable vgabios
Gerd Hoffmann [Wed, 30 May 2018 13:02:16 +0000 (15:02 +0200)] 
bochs-display: enable vgabios

Add vgabios binary to pci rom bar.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
3 years agoseabios: update bios and vgabios binaries
Gerd Hoffmann [Wed, 20 Jun 2018 10:17:34 +0000 (12:17 +0200)] 
seabios: update bios and vgabios binaries

Adds two new vgabios binaries, for ramfb and bochs-display.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
3 years agovgabios: remove submodule and build rules.
Gerd Hoffmann [Wed, 20 Jun 2018 10:11:42 +0000 (12:11 +0200)] 
vgabios: remove submodule and build rules.

It's the old, lgpl vgabios implementation.

Was left in as fallback when we switched to seavgabios, so we could
easily switch back in case we see regressions.  It's unused since years
now, reportedly doesn't even build, and lacks support for recently (and
not so recently) added display devices.

Zap it.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
3 years agoseabios: enable ide dma
Gerd Hoffmann [Wed, 20 Jun 2018 10:07:40 +0000 (12:07 +0200)] 
seabios: enable ide dma

QNX reportedly requires this to boot.
Should also speed up booting other guests.

Note: Upstream seabios defaults this to 'n' to due to known problems
on physical hardware (qemu not affected), and wouldn't flip the default
to 'y'.  So we adjust our local build config accordingly.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
3 years agoseabios: add vga configs for bochs-display and ramfb
Gerd Hoffmann [Wed, 20 Jun 2018 10:05:55 +0000 (12:05 +0200)] 
seabios: add vga configs for bochs-display and ramfb

Both bochs-display and ramfb are devices with a simple framebuffer and
no vga emulation or text mode.  seavgabios has support for text mode
emulation (at vgabios call level), we are using that to provide some
vga compatibility support for these devices.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
3 years agoseabios: update submodule to release 1.11.2
Gerd Hoffmann [Wed, 20 Jun 2018 10:02:28 +0000 (12:02 +0200)] 
seabios: update submodule to release 1.11.2

git shortlog rel-1.11.1..rel-1.11.2
-----------------------------------

Gerd Hoffmann (11):
      optionrom: enable non-vga display devices
      cbvga: factor out cbvga_setup_modes()
      qemu: add bochs-display support
      cbvga_setup_modes: use real mode number instead of 0x140
      cbvga_list_modes: don't list current mode twice
      cbvga_set_mode: disable clearmem in windows x86 emulator.
      bochs_display_setup: return error on failure
      pmm: use tmp zone on oom
      vgasrc: add allocate_pmm()
      qemu: add qemu ramfb support
      cbvga_set_mode: refine clear display logic

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
3 years agocpu: Assert asidx_from_attrs return value in range
Richard Henderson [Fri, 29 Jun 2018 23:30:28 +0000 (16:30 -0700)] 
cpu: Assert asidx_from_attrs return value in range

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
3 years agoaccel/tcg: Avoid caching overwritten tlb entries
Richard Henderson [Fri, 29 Jun 2018 20:07:08 +0000 (13:07 -0700)] 
accel/tcg: Avoid caching overwritten tlb entries

When installing a TLB entry, remove any cached version of the
same page in the VTLB.  If the existing TLB entry matches, do
not copy into the VTLB, but overwrite it.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
3 years agoaccel/tcg: Don't treat invalid TLB entries as needing recheck
Peter Maydell [Fri, 29 Jun 2018 16:17:31 +0000 (17:17 +0100)] 
accel/tcg: Don't treat invalid TLB entries as needing recheck

In get_page_addr_code() when we check whether the TLB entry
is marked as TLB_RECHECK, we should not go down that code
path if the TLB entry is not valid at all (ie the TLB_INVALID
bit is set).

Tested-by: Laurent Vivier <laurent@vivier.eu>
Reported-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20180629161731.16239-1-peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
3 years agoaccel/tcg: Correct "is this a TLB miss" check in get_page_addr_code()
Peter Maydell [Fri, 29 Jun 2018 16:21:22 +0000 (17:21 +0100)] 
accel/tcg: Correct "is this a TLB miss" check in get_page_addr_code()

In commit 71b9a45330fe220d1 we changed the condition we use
to determine whether we need to refill the TLB in
get_page_addr_code() to
    if (unlikely(env->tlb_table[mmu_idx][index].addr_code !=
                 (addr & (TARGET_PAGE_MASK | TLB_INVALID_MASK)))) {

This isn't the right check (it will falsely fail if the
input addr happens to have the low bit corresponding to
TLB_INVALID_MASK set, for instance). Replace it with a
use of the new tlb_hit() function, which is the correct test.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20180629162122.19376-3-peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
3 years agotcg: Define and use new tlb_hit() and tlb_hit_page() functions
Peter Maydell [Fri, 29 Jun 2018 16:21:21 +0000 (17:21 +0100)] 
tcg: Define and use new tlb_hit() and tlb_hit_page() functions

The condition to check whether an address has hit against a particular
TLB entry is not completely trivial. We do this in various places, and
in fact in one place (get_page_addr_code()) we have got the condition
wrong. Abstract it out into new tlb_hit() and tlb_hit_page() inline
functions (one for a known-page-aligned address and one for an
arbitrary address), and use them in all the places where we had the
condition correct.

This is a no-behaviour-change patch; we leave fixing the buggy
code in get_page_addr_code() to a subsequent patch.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20180629162122.19376-2-peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
3 years agotranslate-all: fix locking of TBs whose two pages share the same physical page
Emilio G. Cota [Mon, 25 Jun 2018 16:31:42 +0000 (12:31 -0400)] 
translate-all: fix locking of TBs whose two pages share the same physical page

Commit 0b5c91f ("translate-all: use per-page locking in !user-mode",
2018-06-15) introduced per-page locking. It assumed that the physical
pages corresponding to a TB (at most two pages) are always distinct,
which is wrong. For instance, an xtensa test provided by Max Filippov
is broken by the commit, since the test maps two virtual pages
to the same physical page:

virt1: 7fff, virt2: 8000
phys1 6000fff, phys2 6000000

Fix it by removing the assumption from page_lock_pair.
If the two physical page addresses are equal, we only lock
the PageDesc once. Note that the two callers of page_lock_pair,
namely page_unlock_tb and tb_link_page, are also updated so that
we do not try to unlock the same PageDesc twice.

Fixes: 0b5c91f74f3c83a36f37740969df8c775c997e69
Reported-by: Max Filippov <jcmvbkbc@gmail.com>
Tested-by: Max Filippov <jcmvbkbc@gmail.com>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1529944302-14186-1-git-send-email-cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
3 years agoMerge remote-tracking branch 'remotes/cohuck/tags/s390x-20180702' into staging
Peter Maydell [Mon, 2 Jul 2018 13:57:43 +0000 (14:57 +0100)] 
Merge remote-tracking branch 'remotes/cohuck/tags/s390x-20180702' into staging

s390x updates:
- add bpb/ppa15 features to default cpu model for z196 and later
- rework TOD handling and fix cpu hotplug under tcg
- various fixes

# gpg: Signature made Mon 02 Jul 2018 12:09:40 BST
# gpg:                using RSA key DECF6B93C6F02FAF
# gpg: Good signature from "Cornelia Huck <conny@cornelia-huck.de>"
# gpg:                 aka "Cornelia Huck <huckc@linux.vnet.ibm.com>"
# gpg:                 aka "Cornelia Huck <cornelia.huck@de.ibm.com>"
# gpg:                 aka "Cornelia Huck <cohuck@kernel.org>"
# gpg:                 aka "Cornelia Huck <cohuck@redhat.com>"
# Primary key fingerprint: C3D0 D66D C362 4FF6 A8C0  18CE DECF 6B93 C6F0 2FAF

* remotes/cohuck/tags/s390x-20180702:
  s390x/tcg: fix locking problem with tcg_s390_tod_updated
  s390x/kvm: indicate alignment in legacy_s390_alloc()
  s390x/kvm: legacy_s390_alloc() only supports one allocation
  s390x/tcg: fix CPU hotplug with single-threaded TCG
  s390x/tcg: rearm the CKC timer during migration
  s390x/tcg: implement SET CLOCK
  s390x/tcg: SET CLOCK COMPARATOR can clear CKC interrupts
  s390x/tcg: properly implement the TOD
  s390x/tcg: drop tod_basetime
  s390x/tod: factor out TOD into separate device
  s390x/kvm: pass values instead of pointers to kvm_s390_set_clock_*()
  s390x/tcg: avoid overflows in time2tod/tod2time
  s390x/cpumodel: default enable bpb and ppa15 for z196 and later
  loader: Check access size when calling rom_ptr() to avoid crashes
  s390/ipl: fix ipl with -no-reboot

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agotcg: simplify !CONFIG_TCG handling of tb_invalidate_*
Paolo Bonzini [Mon, 2 Jul 2018 12:45:25 +0000 (14:45 +0200)] 
tcg: simplify !CONFIG_TCG handling of tb_invalidate_*

There is no need for a stub, since tb_invalidate_phys_addr can be excised
altogether when TCG is disabled.  This is a bit cleaner since it avoids
using code that is clearly specific to user-mode emulation (it calls
mmap_lock/unlock) for the !CONFIG_TCG case.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoi386/monitor.c: make addresses canonical for "info mem" and "info tlb"
Doug Gale [Sun, 17 Jun 2018 08:40:25 +0000 (04:40 -0400)] 
i386/monitor.c: make addresses canonical for "info mem" and "info tlb"

Correct the output of the "info mem" and "info tlb" monitor commands to
correctly show canonical addresses.

In 48-bit addressing mode, the upper 16 bits of linear addresses are
equal to bit 47. In 57-bit addressing mode (LA57), the upper 7 bits of
linear addresses are equal to bit 56.

Signed-off-by: Doug Gale <doug16k@gmail.com>
Message-Id: <20180617084025.29198-1-doug16k@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agotarget-i386: Add NPT support
Jan Kiszka [Sat, 30 Jun 2018 06:08:23 +0000 (08:08 +0200)] 
target-i386: Add NPT support

This implements NPT suport for SVM by hooking into
x86_cpu_handle_mmu_fault where it reads the stage-1 page table. Whether
we need to perform this 2nd stage translation, and how, is decided
during vmrun and stored in hflags2, along with nested_cr3 and
nested_pg_mode.

As get_hphys performs a direct cpu_vmexit in case of NPT faults, we need
retaddr in that function. To avoid changing the signature of
cpu_handle_mmu_fault, this passes the value from tlb_fill to get_hphys
via the CPU state.

This was tested successfully via the Jailhouse hypervisor.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Message-Id: <567473a0-6005-5843-4c73-951f476085ca@web.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoserial: Open non-block
Dr. David Alan Gilbert [Mon, 26 Feb 2018 13:04:52 +0000 (13:04 +0000)] 
serial: Open non-block

On a real serial device, the open can block if the handshake
lines are in a particular state.  If a QEMU is passing the serial
device to the guest, the QEMU startup is blocked opening the device
(with a symptom seen as a timeout from libvirt).

Open the serial port with O_NONBLOCK.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agobsd-user: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:38 +0000 (09:42 -0300)] 
bsd-user: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -n '[<>][<>]= ?[1-5]0'

and modified manually.

Suggested-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-47-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agolinux-user: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:37 +0000 (09:42 -0300)] 
linux-user: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -n '[<>][<>]= ?[1-5]0'

and modified manually.

Suggested-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-46-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agotests/crypto: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:36 +0000 (09:42 -0300)] 
tests/crypto: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -n '[<>][<>]= ?[1-5]0'

and modified manually.

Suggested-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-45-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agovl: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:35 +0000 (09:42 -0300)] 
vl: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -n '[<>][<>]= ?[1-5]0'

and modified manually.

Suggested-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-44-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agomonitor: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:34 +0000 (09:42 -0300)] 
monitor: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -n '[<>][<>]= ?[1-5]0'

and modified manually.

Suggested-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-43-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agocutils: Do not include "qemu/units.h" directly
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:32 +0000 (09:42 -0300)] 
cutils: Do not include "qemu/units.h" directly

All files using "qemu/units.h" definitions already include it directly,
we can now remove it from "qemu/cutils.h".

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Message-Id: <20180625124238.25339-41-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/rdma: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:31 +0000 (09:42 -0300)] 
hw/rdma: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Yuval Shaia <yuval.shaia@oracle.com>
Message-Id: <20180625124238.25339-40-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/virtio: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:30 +0000 (09:42 -0300)] 
hw/virtio: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-39-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/vfio: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:29 +0000 (09:42 -0300)] 
hw/vfio: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-38-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/sd: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:28 +0000 (09:42 -0300)] 
hw/sd: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-37-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/usb: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:27 +0000 (09:42 -0300)] 
hw/usb: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-36-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/net: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:26 +0000 (09:42 -0300)] 
hw/net: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Message-Id: <20180625124238.25339-35-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/i386: Use the IEC binary prefix definitions
Paolo Bonzini [Fri, 29 Jun 2018 14:22:13 +0000 (16:22 +0200)] 
hw/i386: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '[<>][<>]=? ?[1-5]0' hw/ include/hw/

and modified manually.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/ppc: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:24 +0000 (09:42 -0300)] 
hw/ppc: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20180625124238.25339-33-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/mips: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:22 +0000 (09:42 -0300)] 
hw/mips: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/
  $ git grep -n '[<>][<>]= ?[1-5]0'

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20180625124238.25339-31-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/mips/r4k: Constify params_size
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:21 +0000 (09:42 -0300)] 
hw/mips/r4k: Constify params_size

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-30-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/sh4: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:20 +0000 (09:42 -0300)] 
hw/sh4: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-29-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/lm32: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:19 +0000 (09:42 -0300)] 
hw/lm32: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Acked-by: Michael Walle <michael@walle.cc>
Message-Id: <20180625124238.25339-28-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/cris: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:18 +0000 (09:42 -0300)] 
hw/cris: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-27-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/nios2: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:17 +0000 (09:42 -0300)] 
hw/nios2: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-26-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/microblaze: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:16 +0000 (09:42 -0300)] 
hw/microblaze: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-25-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/tricore: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:15 +0000 (09:42 -0300)] 
hw/tricore: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Acked-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Message-Id: <20180625124238.25339-24-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/alpha: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:14 +0000 (09:42 -0300)] 
hw/alpha: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-23-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/xtensa: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:13 +0000 (09:42 -0300)] 
hw/xtensa: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/
  $ git grep -n '[<>][<>]= ?[1-5]0'

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Acked-by: Max Filippov <jcmvbkbc@gmail.com>
Message-Id: <20180625124238.25339-22-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/hppa: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:11 +0000 (09:42 -0300)] 
hw/hppa: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20180625124238.25339-20-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/s390x: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:11 +0000 (09:42 -0300)] 
hw/s390x: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20180625124238.25339-20-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/sparc: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:10 +0000 (09:42 -0300)] 
hw/sparc: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-19-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/m68k: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:09 +0000 (09:42 -0300)] 
hw/m68k: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Thomas Huth <thuth@tuxfamily.org>
Message-Id: <20180625124238.25339-18-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/riscv: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:08 +0000 (09:42 -0300)] 
hw/riscv: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Michael Clark <mjc@sifive.com>
Message-Id: <20180625124238.25339-17-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/misc: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:07 +0000 (09:42 -0300)] 
hw/misc: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-16-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/display: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:06 +0000 (09:42 -0300)] 
hw/display: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20180625124238.25339-15-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/block: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:05 +0000 (09:42 -0300)] 
hw/block: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-14-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/tpm: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:04 +0000 (09:42 -0300)] 
hw/tpm: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180625124238.25339-13-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/xen: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:03 +0000 (09:42 -0300)] 
hw/xen: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Alan Robinson <Alan.Robinson@ts.fujitsu.com>
Message-Id: <20180625124238.25339-12-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/smbios: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:02 +0000 (09:42 -0300)] 
hw/smbios: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-11-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/scsi: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:01 +0000 (09:42 -0300)] 
hw/scsi: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-10-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/ipack: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:42:00 +0000 (09:42 -0300)] 
hw/ipack: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Message-Id: <20180625124238.25339-9-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw/ivshmem: Use the IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:41:59 +0000 (09:41 -0300)] 
hw/ivshmem: Use the IEC binary prefix definitions

It eases code review, unit is explicit.

Patch generated using:

  $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ include/hw/

and modified manually.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180625124238.25339-8-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw: Directly use "qemu/units.h" instead of "qemu/cutils.h"
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:41:58 +0000 (09:41 -0300)] 
hw: Directly use "qemu/units.h" instead of "qemu/cutils.h"

These files don't use anything exposed by "qemu/cutils.h",
simplify preprocessing including directly "qemu/units.h".

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au> (ppc parts)
Message-Id: <20180625124238.25339-7-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agohw: Use IEC binary prefix definitions from "qemu/units.h"
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:41:57 +0000 (09:41 -0300)] 
hw: Use IEC binary prefix definitions from "qemu/units.h"

Code change produced with:

  $ git ls-files | egrep '\.[ch]$' | \
    xargs sed -i -e 's/\(\W[KMGTPE]\)_BYTE/\1iB/g'

Suggested-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au> (ppc parts)
Message-Id: <20180625124238.25339-6-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agocheckpatch: Recognize IEC binary prefix definitions
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:41:56 +0000 (09:41 -0300)] 
checkpatch: Recognize IEC binary prefix definitions

Do not match the IEC binary prefix as camelcase typedefs.

This fixes:

    ERROR: "foo * bar" should be "foo *bar"
    #310: FILE: hw/ppc/ppc440_uc.c:564:
    +        size = 8 * MiB * sh;
    total: 1 errors, 0 warnings, 433 lines checked

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-5-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agox86/cpu: Use definitions from "qemu/units.h"
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:41:55 +0000 (09:41 -0300)] 
x86/cpu: Use definitions from "qemu/units.h"

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Acked-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20180625124238.25339-4-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agovdi: Use definitions from "qemu/units.h"
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:41:54 +0000 (09:41 -0300)] 
vdi: Use definitions from "qemu/units.h"

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Message-Id: <20180625124238.25339-3-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoinclude: Add IEC binary prefixes in "qemu/units.h"
Philippe Mathieu-Daudé [Mon, 25 Jun 2018 12:41:53 +0000 (09:41 -0300)] 
include: Add IEC binary prefixes in "qemu/units.h"

Loosely based on 076b35b5a56.

Suggested-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180625124238.25339-2-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoconfigure: add sanity check to catch builds from "git archive"
Daniel P. Berrangé [Wed, 18 Apr 2018 17:11:51 +0000 (18:11 +0100)] 
configure: add sanity check to catch builds from "git archive"

The "git archive" feature creates tarballs which are missing all
submodule content. GitHub unhelpfully provides users with "Download"
links that claim to give them valid source release tarballs. These
GitHub archives will not be buildable as they are created by the
"git archive" feature and so are missing content. The user gets
unhelpful messages from make such as:

  fatal error: ui/input-keymap-atset1-to-qcode.c: No such file or directory

By adding a sanity check we can give users an informative message about
what they've done wrong.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20180418171151.5263-1-berrange@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoi386/kvm: add support for Hyper-V TLB flush
Vitaly Kuznetsov [Sun, 10 Jun 2018 18:49:27 +0000 (20:49 +0200)] 
i386/kvm: add support for Hyper-V TLB flush

Add support for Hyper-V TLB flush which recently got added to KVM.

Just like regular Hyper-V we announce HV_EX_PROCESSOR_MASKS_RECOMMENDED
regardless of how many vCPUs we have. Windows is 'smart' and uses less
expensive non-EX Hypercall whenever possible (when it wants to flush TLB
for all vCPUs or the maximum vCPU index in the vCPU set requires flushing
is less than 64).

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20180610184927.19309-1-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
3 years agoMerge remote-tracking branch 'remotes/vivier/tags/m68k-for-3.0-pull-request' into...
Peter Maydell [Mon, 2 Jul 2018 12:43:10 +0000 (13:43 +0100)] 
Merge remote-tracking branch 'remotes/vivier/tags/m68k-for-3.0-pull-request' into staging

Fix move16 instruction disassembly

# gpg: Signature made Mon 02 Jul 2018 11:25:51 BST
# gpg:                using RSA key F30C38BD3F2FBE3C
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>"
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>"
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>"
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* remotes/vivier/tags/m68k-for-3.0-pull-request:
  target/m68k: correctly disassemble move16

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agotcg: Fix --disable-tcg build breakage
Philippe Mathieu-Daudé [Fri, 29 Jun 2018 20:07:10 +0000 (17:07 -0300)] 
tcg: Fix --disable-tcg build breakage

Fix the --disable-tcg breakage introduced by 8bca9a03ec60d:

    $ configure --disable-tcg
    [...]
    $ make -C i386-softmmu exec.o
    make: Entering directory 'i386-softmmu'
      CC      exec.o
    In file included from source/qemu/exec.c:62:0:
    source/qemu/include/exec/ram_addr.h:96:6: error: conflicting types for ‘tb_invalidate_phys_range’
     void tb_invalidate_phys_range(ram_addr_t start, ram_addr_t end);
          ^~~~~~~~~~~~~~~~~~~~~~~~
    In file included from source/qemu/exec.c:24:0:
    source/qemu/include/exec/exec-all.h:309:6: note: previous declaration of ‘tb_invalidate_phys_range’ was here
     void tb_invalidate_phys_range(target_ulong start, target_ulong end);
          ^~~~~~~~~~~~~~~~~~~~~~~~
    source/qemu/exec.c:1043:6: error: conflicting types for ‘tb_invalidate_phys_addr’
     void tb_invalidate_phys_addr(AddressSpace *as, hwaddr addr, MemTxAttrs attrs)
          ^~~~~~~~~~~~~~~~~~~~~~~
    In file included from source/qemu/exec.c:24:0:
    source/qemu/include/exec/exec-all.h:308:6: note: previous declaration of ‘tb_invalidate_phys_addr’ was here
     void tb_invalidate_phys_addr(target_ulong addr);
          ^~~~~~~~~~~~~~~~~~~~~~~
    make: *** [source/qemu/rules.mak:69: exec.o] Error 1
    make: Leaving directory 'i386-softmmu'

Tested to build x86_64-softmmu and i386-softmmu targets.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20180629200710.27626-1-f4bug@amsat.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agotarget/m68k: correctly disassemble move16
Laurent Vivier [Mon, 25 Jun 2018 20:35:59 +0000 (22:35 +0200)] 
target/m68k: correctly disassemble move16

"move16 %a0@+,%a1@" and "fmovel (cpid=3) %a0@-,%fpcr"
share the same opcode.

To fix that, backport the fix from binutils:

  2005-11-10  Andreas Schwab  <schwab@suse.de>

     * m68k-dis.c (print_insn_m68k): Only match FPU insns with
     coprocessor ID 1.

Reported-by: Thomas Huth <huth@tuxfamily.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Tested-by: Thomas Huth <huth@tuxfamily.org>
Message-Id: <20180625203559.21370-2-laurent@vivier.eu>

3 years agoaudio/hda: drop atomics
Gerd Hoffmann [Wed, 27 Jun 2018 11:19:36 +0000 (13:19 +0200)] 
audio/hda: drop atomics

Doesn't build on 32bit clang.  And because we run under qemu mutex
anyway they are not needed.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20180627111936.31019-1-kraxel@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agos390x/tcg: fix locking problem with tcg_s390_tod_updated
David Hildenbrand [Fri, 29 Jun 2018 17:05:20 +0000 (19:05 +0200)] 
s390x/tcg: fix locking problem with tcg_s390_tod_updated

tcg_s390_tod_updated() is always called with the iothread being locked
(e.g. from S390TODClass->set() e.g. via HELPER(sck) or on incoming
migration). The helper we call takes the lock itself - bad.

Let's change that by factoring out updating the ckc timer. This now looks
much nicer than having to call a helper from another function.

While touching it we also make sure that env->ckc is updated even if the
new value is -1ULL, for now it would not have been modified in that case.

Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180629170520.13671-1-david@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/kvm: indicate alignment in legacy_s390_alloc()
David Hildenbrand [Thu, 28 Jun 2018 11:38:17 +0000 (13:38 +0200)] 
s390x/kvm: indicate alignment in legacy_s390_alloc()

Let's do this for completeness reason, although we don't support e.g.
PCDIMM/NVDIMM, which would use the alignment for placing the memory
region in guest physical memory. But maybe someday we would want to
support something like this - then we don't forget about this if
allowing multiple allocations in legacy_s390_alloc().

Use the same alignment as we would set in qemu_anon_ram_alloc(). Our
fixed address satisfies this alignment (1MB). This implicitly sets the
alignment of the underlying memory region.

Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180628113817.30814-3-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/kvm: legacy_s390_alloc() only supports one allocation
David Hildenbrand [Thu, 28 Jun 2018 11:38:16 +0000 (13:38 +0200)] 
s390x/kvm: legacy_s390_alloc() only supports one allocation

We always allocate at a fixed address, a second allocation can therefore
of course never work. We would simply overwrite mappings.

This can e.g. happen in s390_memory_init(), if trying to allocate more
than > 8TB. Let's just bail out, as there is no need for supporting it
(legacy handling for z/VM).

Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180628113817.30814-2-david@redhat.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/tcg: fix CPU hotplug with single-threaded TCG
David Hildenbrand [Wed, 27 Jun 2018 13:44:10 +0000 (15:44 +0200)] 
s390x/tcg: fix CPU hotplug with single-threaded TCG

run_on_cpu() doesn't seem to work reliably until the CPU has been fully
created if the single-threaded TCG main loop is already running.

Therefore, hotplugging a CPU under single-threaded TCG does currently
not work. We should use the direct call instead of going via
run_on_cpu().

So let's use run_on_cpu() for KVM only - KVM requires it due to the initial
CPU reset ioctl. As a nice side effect, we get rid of the ifdef.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180627134410.4901-10-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/tcg: rearm the CKC timer during migration
David Hildenbrand [Wed, 27 Jun 2018 13:44:09 +0000 (15:44 +0200)] 
s390x/tcg: rearm the CKC timer during migration

If the CPU data is migrated after the TOD clock, the CKC timer of a CPU
is not rearmed. Let's rearm it when loading the CPU state.

Introduce tcg-stub.c just like kvm-stub.c for tcg specific stubs.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180627134410.4901-9-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/tcg: implement SET CLOCK
David Hildenbrand [Wed, 27 Jun 2018 13:44:08 +0000 (15:44 +0200)] 
s390x/tcg: implement SET CLOCK

This allows a guest to change its TOD. We already take care of updating
all CKC timers from within S390TODClass.

Use MO_ALIGN to load the operand manually - this will properly trigger a
SPECIFICATION exception.

Acked-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180627134410.4901-8-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/tcg: SET CLOCK COMPARATOR can clear CKC interrupts
David Hildenbrand [Wed, 27 Jun 2018 13:44:07 +0000 (15:44 +0200)] 
s390x/tcg: SET CLOCK COMPARATOR can clear CKC interrupts

Let's stop the timer and delete any pending CKC IRQ before doing
anything else.

While at it, add a comment why the check for ckc == -1ULL is needed.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180627134410.4901-7-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/tcg: properly implement the TOD
David Hildenbrand [Wed, 27 Jun 2018 13:44:06 +0000 (15:44 +0200)] 
s390x/tcg: properly implement the TOD

Right now, each CPU has its own TOD. Especially, the TOD will differ
based on creation time of a CPU - e.g. when hotplugging a CPU the times
will differ quite a lot, resulting in stall warnings in the guest.

Let's use a single TOD by implementing our new TOD device. Prepare it
for TOD-clock epoch extension.

Most importantly, whenever we set the TOD, we have to update the CKC
timer.

Introduce "tcg_s390x.h" just like "kvm_s390x.h" for tcg specific
function declarations that should not go into cpu.h.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180627134410.4901-6-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/tcg: drop tod_basetime
David Hildenbrand [Wed, 27 Jun 2018 13:44:05 +0000 (15:44 +0200)] 
s390x/tcg: drop tod_basetime

Never set to anything but 0.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180627134410.4901-5-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/tod: factor out TOD into separate device
David Hildenbrand [Wed, 27 Jun 2018 13:44:04 +0000 (15:44 +0200)] 
s390x/tod: factor out TOD into separate device

Let's treat this like a separate device. TCG will have to store the
actual state/time later on.

Include cpu-qom.h in kvm_s390x.h (due to S390CPU) to compile tod-kvm.c.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180627134410.4901-4-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/kvm: pass values instead of pointers to kvm_s390_set_clock_*()
David Hildenbrand [Wed, 27 Jun 2018 13:44:03 +0000 (15:44 +0200)] 
s390x/kvm: pass values instead of pointers to kvm_s390_set_clock_*()

We are going to factor out the TOD into a separate device and use const
pointers for device class functions where possible. We are passing right
now ordinary pointers that should never be touched when setting the TOD.
Let's just pass the values directly.

Note that s390_set_clock() will be removed in a follow-on patch and
therefore its calling convention is not changed.

Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180627134410.4901-3-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/tcg: avoid overflows in time2tod/tod2time
David Hildenbrand [Wed, 27 Jun 2018 13:44:02 +0000 (15:44 +0200)] 
s390x/tcg: avoid overflows in time2tod/tod2time

Big values for the TOD/ns clock can result in some overflows that can be
avoided. Not all overflows can be handled however, as the conversion either
multiplies by 4.096 or divided by 4.096.

Apply the trick used in the Linux kernel in arch/s390/include/asm/timex.h
for tod_to_ns() and use the same trick also for the conversion in the
other direction.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180627134410.4901-2-david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390x/cpumodel: default enable bpb and ppa15 for z196 and later
Christian Borntraeger [Tue, 26 Jun 2018 12:38:30 +0000 (14:38 +0200)] 
s390x/cpumodel: default enable bpb and ppa15 for z196 and later

Most systems and host kernels provide the necessary building blocks for
bpb and ppa15. We can reverse the logic and default enable those
features, while still allowing to disable it via cpu model.

So let us add bpb and ppa15 to z196 and later default CPU model for the
qemu 3.0 machine. (like -cpu z13).  Older machine types (e.g.
s390-ccw-virtio-2.12) will retain the old value and not provide those
bits in the default model.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Message-Id: <20180626123830.18282-1-borntraeger@de.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agoloader: Check access size when calling rom_ptr() to avoid crashes
Thomas Huth [Tue, 26 Jun 2018 09:35:40 +0000 (11:35 +0200)] 
loader: Check access size when calling rom_ptr() to avoid crashes

The rom_ptr() function allows direct access to the ROM blobs that we
load during startup. However, there are currently no checks for the
size of the accesses, so it's currently possible to crash QEMU for
example with:

$ echo "Insane in the mainframe" > /tmp/test.txt
$ s390x-softmmu/qemu-system-s390x -kernel /tmp/test.txt -append xyz
Segmentation fault (core dumped)
$ s390x-softmmu/qemu-system-s390x -kernel /tmp/test.txt -initrd /tmp/test.txt
Segmentation fault (core dumped)
$ echo -n HdrS > /tmp/hdr.txt
$ sparc64-softmmu/qemu-system-sparc64 -kernel /tmp/hdr.txt -initrd /tmp/hdr.txt
Segmentation fault (core dumped)

We need a possibility to check the size of the ROM area that we want
to access, thus let's add a size parameter to the rom_ptr() function
to avoid these problems.

Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1530005740-25254-1-git-send-email-thuth@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agos390/ipl: fix ipl with -no-reboot
Christian Borntraeger [Fri, 22 Jun 2018 10:29:28 +0000 (12:29 +0200)] 
s390/ipl: fix ipl with -no-reboot

kexec/kdump as well as the bootloader use a subcode of diagnose 308
that is supposed to reset the I/O subsystem but not comprise a full
"reboot". With the latest refactoring this is now broken when
-no-reboot is used or when libvirt acts on a reboot QMP event, for
example a virt-install from iso images.

We need to mark these "subsystem resets" as special.

Fixes: a30fb811cbe9 (s390x: refactor reset/reipl handling)
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180622102928.173420-1-borntraeger@de.ibm.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
3 years agoMerge remote-tracking branch 'remotes/xtensa/tags/20180630-xtensa' into staging
Peter Maydell [Sat, 30 Jun 2018 21:23:51 +0000 (22:23 +0100)] 
Merge remote-tracking branch 'remotes/xtensa/tags/20180630-xtensa' into staging

target/xtensa updates:

- add diagnostic for zero-overhead loop alignment;
- convert to TranslatorOps;
- don't call get_page_addr_code() from helper functions.

# gpg: Signature made Sat 30 Jun 2018 22:16:30 BST
# gpg:                using RSA key 51F9CC91F83FA044
# gpg: Good signature from "Max Filippov <filippov@cadence.com>"
# gpg:                 aka "Max Filippov <max.filippov@cogentembedded.com>"
# gpg:                 aka "Max Filippov <jcmvbkbc@gmail.com>"
# Primary key fingerprint: 2B67 854B 98E5 327D CDEB  17D8 51F9 CC91 F83F A044

* remotes/xtensa/tags/20180630-xtensa:
  xtensa: Avoid calling get_page_addr_code() from helper function
  target/xtensa: Convert to TranslatorOps
  target/xtensa: Change gen_intermediate_code dc to pointer
  target/xtensa: Convert to DisasContextBase
  target/xtensa: Replace DISAS_UPDATE with DISAS_NORETURN
  target/xtensa: check zero overhead loop alignment

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agoxtensa: Avoid calling get_page_addr_code() from helper function
Peter Maydell [Fri, 22 Jun 2018 13:58:23 +0000 (14:58 +0100)] 
xtensa: Avoid calling get_page_addr_code() from helper function

The xtensa frontend calls get_page_addr_code() from its
itlb_hit_test helper function. This function is really part
of the TCG core's internals, and calling it from a target
helper makes it awkward to make changes to that core code.
It also means that we don't pass the correct retaddr to
tlb_fill(), so we won't correctly handle the case where
an exception is generated.

The helper is used for the instructions IHI, IHU and IPFL.

Change it to call cpu_ldb_code_ra() instead.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agotarget/xtensa: Convert to TranslatorOps
Richard Henderson [Sat, 12 May 2018 17:57:24 +0000 (10:57 -0700)] 
target/xtensa: Convert to TranslatorOps

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
3 years agotarget/xtensa: Change gen_intermediate_code dc to pointer
Richard Henderson [Sat, 12 May 2018 17:57:23 +0000 (10:57 -0700)] 
target/xtensa: Change gen_intermediate_code dc to pointer

This will reduce the size of the patch in the next patch,
where the context will have to be a pointer.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
3 years agotarget/xtensa: Convert to DisasContextBase
Richard Henderson [Sat, 12 May 2018 17:57:22 +0000 (10:57 -0700)] 
target/xtensa: Convert to DisasContextBase

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
3 years agotarget/xtensa: Replace DISAS_UPDATE with DISAS_NORETURN
Richard Henderson [Sat, 12 May 2018 17:57:21 +0000 (10:57 -0700)] 
target/xtensa: Replace DISAS_UPDATE with DISAS_NORETURN

The usage of DISAS_UPDATE is after noreturn helpers.
It is thus indistinguishable from DISAS_NORETURN.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
3 years agotarget/xtensa: check zero overhead loop alignment
Max Filippov [Fri, 27 Apr 2018 20:07:53 +0000 (13:07 -0700)] 
target/xtensa: check zero overhead loop alignment

ISA book documents that the first instruction of zero overhead loop
must fit completely into naturally aligned region of an instruction
fetch unit size. Check that condition and log a message if it's
violated.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
3 years agoMerge remote-tracking branch 'remotes/armbru/tags/pull-monitor-2018-06-30' into staging
Peter Maydell [Sat, 30 Jun 2018 16:30:09 +0000 (17:30 +0100)] 
Merge remote-tracking branch 'remotes/armbru/tags/pull-monitor-2018-06-30' into staging

Monitor patches for 2018-06-30

# gpg: Signature made Sat 30 Jun 2018 17:22:12 BST
# gpg:                using RSA key 3870B400EB918653
# gpg: Good signature from "Markus Armbruster <armbru@redhat.com>"
# gpg:                 aka "Markus Armbruster <armbru@pond.sub.org>"
# Primary key fingerprint: 354B C8B3 D7EB 2A6B 6867  4E5F 3870 B400 EB91 8653

* remotes/armbru/tags/pull-monitor-2018-06-30:
  docs: mention shared state protect for OOB
  tests: iotests: drop some stderr line
  monitor: flush qmp responses when CLOSED
  monitor: rename *_pop_one to *_pop_any
  chardev: comment details for CLOSED event

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
3 years agodocs: mention shared state protect for OOB
Peter Xu [Wed, 20 Jun 2018 07:32:21 +0000 (15:32 +0800)] 
docs: mention shared state protect for OOB

Out-Of-Band handlers need to protect shared state if there is any.
Mention it in the document.  Meanwhile, touch up some other places too,
either with better English, or reordering of bullets.

Suggested-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180620073223.31964-6-peterx@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
3 years agotests: iotests: drop some stderr line
Peter Xu [Wed, 20 Jun 2018 07:32:20 +0000 (15:32 +0800)] 
tests: iotests: drop some stderr line

In my Out-Of-Band test, "check -qcow2 060" fail with this:

  --- /home/peterx/git/qemu/tests/qemu-iotests/060.out
  +++ /home/peterx/git/qemu/bin/tests/qemu-iotests/060.out.bad
  @@ -427,8 +427,8 @@
  QMP_VERSION
  {"return": {}}
  qcow2: Image is corrupt: L2 table offset 0x2a2a2a00 unaligned (L1
  index: 0); further non-fatal corruption events will be suppressed
  -{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_IMAGE_CORRUPTED", "data": {"device": "", "msg": "L2 table offset 0x2a2a2a0
  0 unaligned (L1 index: 0)", "node-name": "drive", "fatal": false}}
  read failed: Input/output error
  +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_IMAGE_CORRUPTED", "data": {"device": "", "msg": "L2 table offset 0x2a2a2a0
  0 unaligned (L1 index: 0)", "node-name": "drive", "fatal": false}}
  {"return": ""}
  {"return": {}}
  {"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP},
  "event": "SHUTDOWN", "data": {"guest": false}}

The order of the event and the in/out error line is swapped.  I didn't
dig up the reason, but AFAIU what we want to verify is the event rather
than stderr.  Let's drop the stderr line directly for this test.

Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180620073223.31964-5-peterx@redhat.com>
[Commit message touched up]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
3 years agomonitor: flush qmp responses when CLOSED
Peter Xu [Wed, 20 Jun 2018 07:32:19 +0000 (15:32 +0800)] 
monitor: flush qmp responses when CLOSED

Previously we clean up the queues when we got CLOSED event.  It was used
to make sure we won't send leftover replies/events of a old client to a
new client which makes perfect sense. However this will also drop the
replies/events even if the output port of the previous chardev backend
is still open, which can lead to missing of the last replies/events.
Now this patch does an extra operation to flush the response queue
before cleaning up.

In most cases, a QMP session will be based on a bidirectional channel (a
TCP port, for example, we read/write to the same socket handle), so in
port and out port of the backend chardev are fundamentally the same
port. In these cases, it does not really matter much on whether we'll
flush the response queue since flushing will fail anyway.  However there
can be cases where in & out ports of the QMP monitor's backend chardev
are separated.  Here is an example:

  cat $QMP_COMMANDS | qemu -qmp stdio ... | filter_commands

In this case, the backend is fd-typed, and it is connected to stdio
where in port is stdin and out port is stdout.  Now if we drop all the
events on the response queue then filter_command process might miss some
events that it might expect.  The thing is that, when stdin closes,
stdout might still be there alive!

In practice, I encountered SHUTDOWN event missing when running test with
iotest 087 with Out-Of-Band enabled.  Here is one of the ways that this
can happen (after "quit" command is executed and QEMU quits the main
loop):

1. [main thread] QEMU queues a SHUTDOWN event into response queue.

2. "cat" terminates (to distinguish it from the animal, I quote it).

3. [monitor iothread] QEMU's monitor iothread reads EOF from stdin.

4. [monitor iothread] QEMU's monitor iothread calls the CLOSED event
   hook for the monitor, which will destroy the response queue of the
   monitor, then the SHUTDOWN event is dropped.

5. [main thread] QEMU's main thread cleans up the monitors in
   monitor_cleanup().  When trying to flush pending responses, it sees
   nothing.  SHUTDOWN is lost forever.

Note that before the monitor iothread was introduced, step [4]/[5] could
never happen since the main loop was the only place to detect the EOF
event of stdin and run the CLOSED event hooks.  Now things can happen in
parallel in the iothread.

Without this patch, iotest 087 will have ~10% chance to miss the
SHUTDOWN event and fail when with Out-Of-Band enabled:

  --- /home/peterx/git/qemu/tests/qemu-iotests/087.out
  +++ /home/peterx/git/qemu/bin/tests/qemu-iotests/087.out.bad
  @@ -8,7 +8,6 @@
  {"return": {}}
  {"error": {"class": "GenericError", "desc": "'node-name' must be
  specified for the root node"}}
  {"return": {}}
  -{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}

  === Duplicate ID ===
  @@ -53,7 +52,6 @@
  {"return": {}}
  {"return": {}}
  {"return": {}}

  -{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}

This patch fixes the problem.

Fixes: 6d2d563f8c ("qmp: cleanup qmp queues properly", 2018-03-27)
Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180620073223.31964-4-peterx@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Commit message and a comment touched up]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
3 years agomonitor: rename *_pop_one to *_pop_any
Peter Xu [Wed, 20 Jun 2018 07:32:18 +0000 (15:32 +0800)] 
monitor: rename *_pop_one to *_pop_any

The old names are confusing since both of the old functions are popping
an item from multiple queues rather than a single queue.  In that
sense, *_pop_any() suites better than *_pop_one().

Since at it, touch up the function monitor_qmp_response_pop_any() a bit
to let the callers pass in a QMPResponse struct instead of returning a
struct.  Change the return value to boolean to mark whether we have
popped a valid response instead.

Suggested-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180620073223.31964-3-peterx@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
3 years agochardev: comment details for CLOSED event
Peter Xu [Wed, 20 Jun 2018 07:32:17 +0000 (15:32 +0800)] 
chardev: comment details for CLOSED event

It was unclear before on what does the CLOSED event mean.  Meanwhile we
add a TODO to fix up the CLOSED event in the future when the in/out
ports are different for a chardev.

CC: Paolo Bonzini <pbonzini@redhat.com>
CC: "Marc-André Lureau" <marcandre.lureau@redhat.com>
CC: Stefan Hajnoczi <stefanha@redhat.com>
CC: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180620073223.31964-2-peterx@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>