tcg: Split out target/arch/cpu-param.h

For all targets, into this new file move TARGET_LONG_BITS,
TARGET_PAGE_BITS, TARGET_PHYS_ADDR_SPACE_BITS,
TARGET_VIRT_ADDR_SPACE_BITS, and NB_MMU_MODES.

Include this new file from exec/cpu-defs.h.

This now removes the somewhat odd requirement that target/arch/cpu.h
defines TARGET_LONG_BITS before including exec/cpu-defs.h, so push the
bulk of the includes within target/arch/cpu.h to the top.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Fold CPUTLBWindow into CPUTLBDesc

Both structures are allocated once per mmu_idx.
There is no reason for them to be separate.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge remote-tracking branch 'remotes/kraxel/tags/usb-20190607-pull-request' into staging

usb-mtp: refactor the flow of usb_mtp_write_data

# gpg: Signature made Fri 07 Jun 2019 17:56:31 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/usb-20190607-pull-request:
  usb-mtp: refactor the flow of usb_mtp_write_data

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/ehabkost/tags/machine-next-pull-request' into staging

Machine Core queue, 2019-06-07

* Improve CPU hotplug error message
  (Laurent Vivier)
* Remove unused AccelClass::opt_name field
  (Wainer dos Santos Moschetta)

# gpg: Signature made Fri 07 Jun 2019 19:31:28 BST
# gpg:                using RSA key 2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>" [full]
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* remotes/ehabkost/tags/machine-next-pull-request:
  accel: Remove unused AccelClass::opt_name attribute
  numa: improve cpu hotplug error message with a wrong node-id

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

accel: Remove unused AccelClass::opt_name attribute

The AccelType type was converted to AccelClass QOM
object on b14a0b7469f, and the original data type had
a field to store the option name which in turn was
used to search an accelerator. The lookup method
(accel_find) changed too, making the option field
unnecessary but it became AccelClass::opt_name despite
that. Therefore, and given that none accelerator
implementation sets AccelClass::opt_name, let's
remove this attribute.

Signed-off-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Message-Id: <20190531165334.20403-2-wainersm@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

numa: improve cpu hotplug error message with a wrong node-id

On pseries, core-ids are strongly binded to a node-id by the command
line option. If an user tries to add a CPU to the wrong node, he has
an error but it is not really helpful:

  qemu-system-ppc64 ... -smp 1,maxcpus=64,cores=1,threads=1,sockets=1 \
                        -numa node,nodeid=0 -numa node,nodeid=1 ...

  (qemu) device_add power9_v2.0-spapr-cpu-core,core-id=30,node-id=1
  Error: node-id=1 must match numa node specified with -numa option

This patch improves this error message by giving to the user the good
node-id to use with the core-id he's providing

  Error: invalid node-id, must be 0

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20190529160747.778-1-lvivier@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

usb-mtp: refactor the flow of usb_mtp_write_data

There's no functional change but the flow is (hopefully)
more consistent for both file and folder object types.

Signed-off-by: Bandan Das <bsd@redhat.com>
Message-Id: <20190401211712.19012-4-bsd@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Merge remote-tracking branch 'remotes/cohuck/tags/s390x-20190607-2' into staging

s390x updates:
- tcg: finalize implementation for the vector facility and
  add it to the 'qemu' cpu model
- linux-user: properly generate ELF_HWCAP

# gpg: Signature made Fri 07 Jun 2019 15:14:42 BST
# gpg:                using RSA key C3D0D66DC3624FF6A8C018CEDECF6B93C6F02FAF
# gpg:                issuer "cohuck@redhat.com"
# gpg: Good signature from "Cornelia Huck <conny@cornelia-huck.de>" [unknown]
# gpg:                 aka "Cornelia Huck <huckc@linux.vnet.ibm.com>" [full]
# gpg:                 aka "Cornelia Huck <cornelia.huck@de.ibm.com>" [full]
# gpg:                 aka "Cornelia Huck <cohuck@kernel.org>" [unknown]
# gpg:                 aka "Cornelia Huck <cohuck@redhat.com>" [unknown]
# Primary key fingerprint: C3D0 D66D C362 4FF6 A8C0  18CE DECF 6B93 C6F0 2FAF

* remotes/cohuck/tags/s390x-20190607-2: (34 commits)
  linux-user: elf: ELF_HWCAP for s390x
  s390x/tcg: Use tcg_gen_gvec_bitsel for VECTOR SELECT
  s390x: Bump the "qemu" CPU model up to a stripped-down z13
  s390x/tcg: We support the Vector Facility
  s390x/tcg: Allow linux-user to use vector instructions
  s390x/tcg: Implement VECTOR FP TEST DATA CLASS IMMEDIATE
  s390x/tcg: Implement VECTOR FP SUBTRACT
  s390x/tcg: Implement VECTOR FP SQUARE ROOT
  s390x/tcg: Implement VECTOR FP PERFORM SIGN OPERATION
  s390x/tcg: Implement VECTOR FP MULTIPLY AND (ADD|SUBTRACT)
  s390x/tcg: Implement VECTOR FP MULTIPLY
  s390x/tcg: Implement VECTOR LOAD ROUNDED
  s390x/tcg: Implement VECTOR LOAD LENGTHENED
  s390x/tcg: Implement VECTOR LOAD FP INTEGER
  s390x/tcg: Implement VECTOR FP DIVIDE
  s390x/tcg: Implement VECTOR FP CONVERT TO LOGICAL 64-BIT
  s390x/tcg: Implement VECTOR FP CONVERT TO FIXED 64-BIT
  s390x/tcg: Implement VECTOR FP CONVERT FROM LOGICAL 64-BIT
  s390x/tcg: Implement VECTOR FP CONVERT FROM FIXED 64-BIT
  s390x/tcg: Implement VECTOR FP COMPARE (EQUAL|HIGH|HIGH OR EQUAL)
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge tag 's390x-tcg-2019-06-07' into s390-next-staging

Finalize implementation of the "Vector Facility" for s390x TCG. Add it
to the QEMU CPU model, so it is enabled as default.
Also:
- One fix (and one workaround) for the STFLE instruction
- Fix the alignment of vector registers (and change the data type)
- Properly generate ELF_HWCAP for s390x for linux-user
- Use a gvec helper for VECTOR SELECT

# gpg: Signature made Fri 07 Jun 2019 02:58:01 PM CEST
# gpg:                using RSA key 1BD9CAAD735C4C3A460DFCCA4DDE10F700FF835A
# gpg:                issuer "david@redhat.com"
# gpg: Good signature from "David Hildenbrand <david@redhat.com>" [full]
# gpg:                 aka "David Hildenbrand <davidhildenbrand@gmail.com>" [full]

* tag 's390x-tcg-2019-06-07': (33 commits)
  linux-user: elf: ELF_HWCAP for s390x
  s390x/tcg: Use tcg_gen_gvec_bitsel for VECTOR SELECT
  s390x: Bump the "qemu" CPU model up to a stripped-down z13
  s390x/tcg: We support the Vector Facility
  s390x/tcg: Allow linux-user to use vector instructions
  s390x/tcg: Implement VECTOR FP TEST DATA CLASS IMMEDIATE
  s390x/tcg: Implement VECTOR FP SUBTRACT
  s390x/tcg: Implement VECTOR FP SQUARE ROOT
  s390x/tcg: Implement VECTOR FP PERFORM SIGN OPERATION
  s390x/tcg: Implement VECTOR FP MULTIPLY AND (ADD|SUBTRACT)
  s390x/tcg: Implement VECTOR FP MULTIPLY
  s390x/tcg: Implement VECTOR LOAD ROUNDED
  s390x/tcg: Implement VECTOR LOAD LENGTHENED
  s390x/tcg: Implement VECTOR LOAD FP INTEGER
  s390x/tcg: Implement VECTOR FP DIVIDE
  s390x/tcg: Implement VECTOR FP CONVERT TO LOGICAL 64-BIT
  s390x/tcg: Implement VECTOR FP CONVERT TO FIXED 64-BIT
  s390x/tcg: Implement VECTOR FP CONVERT FROM LOGICAL 64-BIT
  s390x/tcg: Implement VECTOR FP CONVERT FROM FIXED 64-BIT
  s390x/tcg: Implement VECTOR FP COMPARE (EQUAL|HIGH|HIGH OR EQUAL)
  ...

Signed-off-by: Cornelia Huck <cohuck@redhat.com>

Merge remote-tracking branch 'remotes/kraxel/tags/ui-20190607-pull-request' into staging

curses: 32bit build fix.
egl: dmabuf modifier support.

# gpg: Signature made Fri 07 Jun 2019 14:18:41 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/ui-20190607-pull-request:
  egl-helpers: add modifier support to egl_dmabuf_import_texture()
  egl-helpers: add modifier support to egl_get_fd_for_texture().
  vfio/display: set dmabuf modifier field
  console: add dmabuf modifier field.
  ui/curses: Fix build with -m32

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

linux-user: elf: ELF_HWCAP for s390x

Let's add all HWCAPs that we can support under TCG right now, when the
respective CPU facilities are enabled.

Cc: Riku Voipio <riku.voipio@iki.fi>
Cc: Laurent Vivier <laurent@vivier.eu>
Cc: Cornelia Huck <cohuck@redhat.com>
Cc: Laurent Vivier <laurent@vivier.eu>
Cc: Richard Henderson <richard.henderson@linaro.org>
Acked-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Use tcg_gen_gvec_bitsel for VECTOR SELECT

This replaces the target-specific implementations for VSEL.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x: Bump the "qemu" CPU model up to a stripped-down z13

We don't care about the other two missing base features:
- S390_FEAT_DFP_PACKED_CONVERSION
- S390_FEAT_GROUP_GEN13_PTFF

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: We support the Vector Facility

Let's add it to the max model, so we can enable it.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Allow linux-user to use vector instructions

Once we unlock S390_FEAT_VECTOR for TCG, we want linux-user to be
able to make use of it.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP TEST DATA CLASS IMMEDIATE

We can reuse float64_dcmask().

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP SUBTRACT

Similar to VECTOR FP ADD.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP SQUARE ROOT

Simulate XxC=0 and ERM=0 (current mode), so we can use the existing
helper function.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP PERFORM SIGN OPERATION

The only FP instruction we can implement without an helper.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP MULTIPLY AND (ADD|SUBTRACT)

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP MULTIPLY

Very similar to VECTOR FP DIVIDE.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR LOAD ROUNDED

We can reuse some of the infrastructure introduced for
VECTOR FP CONVERT FROM FIXED 64-BIT and friends.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR LOAD LENGTHENED

Take care of reading/indicating the 32-bit elements.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR LOAD FP INTEGER

We can reuse most of the infrastructure introduced for
VECTOR FP CONVERT FROM FIXED 64-BIT and friends.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP DIVIDE

We can reuse most of the infrastructure added for VECTOR FP ADD.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP CONVERT TO LOGICAL 64-BIT

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP CONVERT TO FIXED 64-BIT

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP CONVERT FROM LOGICAL 64-BIT

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP CONVERT FROM FIXED 64-BIT

1. We'll reuse op_vcdg() for similar instructions later, prepare for
   that.
2. We'll reuse vop64_2() later for other instructions.

We have to mangle the erm (effective rounding mode) and the m4 into
the simd_data(), and properly unmangle them again.

Make sure to restore the erm before triggering an exception.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP COMPARE (EQUAL|HIGH|HIGH OR EQUAL)

Provide for all three instructions all four combinations of cc bit and
s bit.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP COMPARE (AND SIGNAL) SCALAR

As far as I can see, there is only a tiny difference.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FP ADD

1. We'll reuse op_vfa() for similar instructions later, prepare for
   that.
2. We'll reuse vop64_3() for other instructions later.
3. Take care of modifying the vector register only if no trap happened.
 - on traps, flags are not updated and no elements are modified
 - traps don't modify the fpc flags
 - without traps, all exceptions of all elements are merged
4. We'll reuse check_ieee_exc() later when we need the XxC flag.

We have to check for exceptions after processing each element.
Provide separate handlers for single/all element processing. We'll do
the same for all applicable FP instructions.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Export float_comp_to_cc() and float(32|64|128)_dcmask()

Vector floating-point instructions will require these functions, so
allow to use them from other files.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Introduce tcg_s390_vector_exception()

Handling is similar to data exceptions, however we can always store the
VXC into the lowore and the FPC:

z14 PoP, 6-20, "Vector-Exception Code"
    When a vector-processing exception causes a pro-
    gram interruption, a vector-exception code (VXC) is
    stored at location 147, and zeros are stored at loca-
    tions 144-146. The VXC is also placed in the DXC
    field of the floating-point-control (FPC) register if bit
    45 of control register 0 is one. When bit 45 of control
    register 0 is zero and bit 46 of control register 0 is
    one, the DXC field of the FPC register and the con-
    tents of storage at location 147 are unpredictable.

Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Store only the necessary amount of doublewords for STFLE

The PoP (z14, 7-382) says:
    Doublewords to the right of the doubleword in which the
    highest-numbered facility bit is assigned for a model
    may or may not be stored.

However, stack protection in certain binaries can't deal with that.
"gzip" example code:

f1b4:       a7 08 00 03             lhi     %r0,3
f1b8:       b2 b0 f0 a0             stfle   160(%r15)
f1bc:       e3 20 f0 b2 00 90       llgc    %r2,178(%r15)
f1c2:       c0 2b 00 00 00 01       nilf    %r2,1
f1c8:       b2 4f 00 10             ear     %r1,%a0
f1cc:       b9 14 00 22             lgfr    %r2,%r2
f1d0:       eb 11 00 20 00 0d       sllg    %r1,%r1,32
f1d6:       b2 4f 00 11             ear     %r1,%a1
f1da:       d5 07 f0 b8 10 28       clc     184(8,%r15),40(%r1)
f1e0:       a7 74 00 06             jne     f1ec <file_read@@Base+0x1bc>
f1e4:       eb ef f1 30 00 04       lmg     %r14,%r15,304(%r15)
f1ea:       07 fe                   br      %r14
f1ec:       c0 e5 ff ff 9d 6e       brasl   %r14,2cc8 <__stack_chk_fail@plt>

In QEMU, we currently have:
    max_bytes = 24
the code asks for (3 + 1) doublewords == 32 bytes.

If we write 32 bytes instead of only 24, and return "2 + 1" doublewords
("one less than the number of doulewords needed to contain all of the
 facility bits"), the example code detects a stack corruption.

In my opinion, the code is wrong. However, it seems to work fine on
real machines. So let's limit storing to the minimum of the requested
and the maximum doublewords.

Cc: Stefan Liebler <stli@linux.ibm.com>
Cc: Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Fix max_byte detection for stfle

used_stfl_bytes is 0, before initialized via prepare_stfl() on the
first invocation. We have to move the calculation of max_bytes after
prepare_stfl().

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x: Use uint64_t for vector registers

CPU_DoubleU is primarily used to reinterpret between integer and floats.
We don't really need this functionality. So let's just keep it simple
and use an uint64_t.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x: Align vector registers to 16 bytes

11e2bfef7990 ("tcg/i386: Use MOVDQA for TCG_TYPE_V128 load/store")
revealed that the vregs are not aligned to 16 bytes. Align them to
16 bytes, to avoid segfault'ing on x86.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR STRING RANGE COMPARE

Unfortunately, there is no easy way to avoid looping over all elements
in v2. Provide specialized variants for !cc,!rt/!cc,rt/cc,!rt/cc,rt and
all element types. Especially for different values of rt, the compiler
might be able to optimize the code a lot.

Add s390_vec_write_element().

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR ISOLATE STRING

Logic mostly courtesy of Richard H.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FIND ELEMENT NOT EQUAL

Similar to VECTOR FIND ELEMENT EQUAL. Core logic courtesy of Richard H.

Add s390_vec_read_element() that can deal with element sizes.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FIND ELEMENT EQUAL

Core logic courtesy of Richard H.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

s390x/tcg: Implement VECTOR FIND ANY ELEMENT EQUAL

Complicated stuff. Provide two different helpers for CC an !CC handling.
We might want to add more helpers later.

zero_search() and match_index() are courtesy of Richard H.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>

MAINTAINERS: cover tests/migration/s390x/

The generic s390 section looks like the best resting place.

Message-Id: <20190529091243.25562-1-cohuck@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

Merge remote-tracking branch 'remotes/amarkovic/tags/mips-queue-jun-7-2019' into staging

MIPS Queue for June 7th, 2019

# gpg: Signature made Fri 07 Jun 2019 10:59:15 BST
# gpg:                using RSA key D4972A8967F75A65
# gpg: Good signature from "Aleksandar Markovic <amarkovic@wavecomp.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 8526 FBF1 5DA3 811F 4A01  DD75 D497 2A89 67F7 5A65

* remotes/amarkovic/tags/mips-queue-jun-7-2019:
  tests/tcg: target/mips: Add README for MSA tests
  tests/tcg: target/mips: Add tests for MSA FP max/min instructions
  tests/tcg: target/mips: Add utility function reset_msa_registers()
  tests/tcg: target/mips: Move four tests to a better location
  tests/tcg: target/mips: Add tests for MSA shift instructions
  tests/tcg: target/mips: Amend and rearrange MSA wrappers
  target/mips: Unroll loops in helpers for MSA logic instructions
  target/mips: Outline places for future MSA helpers
  target/mips: Fix block-comment-related issues in msa_helper.c
  target/mips: Fix space-related format issues in msa_helper.c

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

tests/tcg: target/mips: Add README for MSA tests

Add README for MSA tests. This is just to explain how to run tests even
without Makefile. Makefile will be provided later on.

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-11-git-send-email-aleksandar.markovic@rt-rk.com>

tests/tcg: target/mips: Add tests for MSA FP max/min instructions

Add tests for MSA FP max/min instructions. This includes following
instructions:

  * FMAX.W - float maximum (words)
  * FMAX.D - float maximum (doublewords)
  * FMAX_A.W - float maximum absolute (words)
  * FMAX_A.D - float maximum absolute (doublewords)
  * FMIN.W - float minimum (words)
  * FMIN.D - float minimum (doublewords)
  * FMIN_A.W - float minimum absolute (words)
  * FMIN_A.D - float minimum absolute (doublewords)

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-10-git-send-email-aleksandar.markovic@rt-rk.com>

tests/tcg: target/mips: Add utility function reset_msa_registers()

Add function reset_msa_registers() and utilize it in each MSA test.
This is needed to ensure independency of test results on the state of
MSA registers before test execution. This also allows for correction
of tests for VSHF* instructions, that are now independent on the
previous state of MSA registers.

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-9-git-send-email-aleksandar.markovic@rt-rk.com>

tests/tcg: target/mips: Move four tests to a better location

Move tests for <MUL|MULR>_Q.<H|B> from "integer multiply" directory
to "fixed-point multiply" directory, since they do not operate on
integers, but on fixed point numbers.

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-8-git-send-email-aleksandar.markovic@rt-rk.com>

tests/tcg: target/mips: Add tests for MSA shift instructions

Add tests for MSA shift instructions. This includes following
instructions:

  * SLL.B - shift left logical (bytes)
  * SLL.H - shift left logical (halfwords)
  * SLL.W - shift left logical (words)
  * SLL.D - shift left logical (doublewords)
  * SRA.B - shift right arithmetic (bytes)
  * SRA.H - shift right arithmetic (halfwords)
  * SRA.W - shift right arithmetic (words)
  * SRA.D - shift right arithmetic (doublewords)
  * SRAR.B - shift right arithmetic rounded (bytes)
  * SRAR.H - shift right arithmetic rounded (halfwords)
  * SRAR.W - shift right arithmetic rounded (words)
  * SRAR.D - shift right arithmetic rounded (doublewords)
  * SRL.B - shift right logical (bytes)
  * SRL.H - shift right logical (halfwords)
  * SRL.W - shift right logical (words)
  * SRL.D - shift right logical (doublewords)
  * SRLR.B - shift right logical rounded (bytes)
  * SRLR.H - shift right logical rounded (halfwords)
  * SRLR.W - shift right logical rounded (words)
  * SRLR.D - shift right logical rounded (doublewords)

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-7-git-send-email-aleksandar.markovic@rt-rk.com>

tests/tcg: target/mips: Amend and rearrange MSA wrappers

Amend and rearrange MSA wrappers to follow the same organization as
in MSA tests.

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-6-git-send-email-aleksandar.markovic@rt-rk.com>

target/mips: Unroll loops in helpers for MSA logic instructions

Unroll loops in helpers for MSA logic instructions for better
performance.

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-5-git-send-email-aleksandar.markovic@rt-rk.com>

target/mips: Outline places for future MSA helpers

Outline places for future MSA helpers to follow the same organization
as in MSA tests.

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-4-git-send-email-aleksandar.markovic@rt-rk.com>

egl-helpers: add modifier support to egl_dmabuf_import_texture()

Check and use QemuDmaBuf->modifier in egl_dmabuf_import_texture()
for dmabuf imports.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20190529072144.26737-5-kraxel@redhat.com

egl-helpers: add modifier support to egl_get_fd_for_texture().

Add modifier parameter to egl_get_fd_for_texture(), to return the used
modifier on dmabuf exports.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20190529072144.26737-4-kraxel@redhat.com

vfio/display: set dmabuf modifier field

Fill the new QemuDmaBuf->modifier field properly from plane info.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
Message-id: 20190529072144.26737-3-kraxel@redhat.com

console: add dmabuf modifier field.

dmabufs can have a format modifier (DRM_FORMAT_MOD_*) which is used for
tiled layouts for example.  Add a field to QemuDmaBuf so we can carry
around that information.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-id: 20190529072144.26737-2-kraxel@redhat.com

ui/curses: Fix build with -m32

wchar_t may resolve to be an unsigned long on 32-bit architectures.
Using the %x conversion specifier will then give a compiler warning:

ui/curses.c: In function ‘get_ucs’:
ui/curses.c:492:49: error: format ‘%x’ expects argument of type ‘unsigned int’, but argument 3 has type ‘wchar_t’ {aka ‘long int’} [-Werror=format=]
  492 |         fprintf(stderr, "Could not convert 0x%04x "
      |                                              ~~~^
      |                                                 |
      |                                                 unsigned int
      |                                              %04lx
  493 |                         "from wchar_t to a multibyte character: %s\n",
  494 |                         wch, strerror(errno));
      |                         ~~~
      |                         |
      |                         wchar_t {aka long int}
ui/curses.c:504:49: error: format ‘%x’ expects argument of type ‘unsigned int’, but argument 3 has type ‘wchar_t’ {aka ‘long int’} [-Werror=format=]
  504 |         fprintf(stderr, "Could not convert 0x%04x "
      |                                              ~~~^
      |                                                 |
      |                                                 unsigned int
      |                                              %04lx
  505 |                         "from a multibyte character to UCS-2 : %s\n",
  506 |                         wch, strerror(errno));
      |                         ~~~
      |                         |
      |                         wchar_t {aka long int}

Fix this by casting the wchar_t value to an unsigned long and using %lx
as the conversion specifier.

Fixes: b7b664a4fe9a955338f2e11a0f7433b29c8cbad0
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Message-id: 20190527142540.23255-1-mreitz@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

target/mips: Fix block-comment-related issues in msa_helper.c

Fix block-comment-related issues reported by checkpatch for file
msa_helper.c.

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-3-git-send-email-aleksandar.markovic@rt-rk.com>

target/mips: Fix space-related format issues in msa_helper.c

Fix space-related format issues reported by checkpatch in file
msa_helper.c.

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Aleksandar Rikalo <arikalo@wavecomp.com>
Message-Id: <1559838440-9866-2-git-send-email-aleksandar.markovic@rt-rk.com>

Merge remote-tracking branch 'remotes/vivier2/tags/trivial-branch-pull-request' into staging

Trivial fixes 06/06/2019

# gpg: Signature made Thu 06 Jun 2019 12:05:50 BST
# gpg:                using RSA key CD2F75DDC8E3A4DC2E4F5173F30C38BD3F2FBE3C
# gpg:                issuer "laurent@vivier.eu"
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>" [full]
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>" [full]
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>" [full]
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* remotes/vivier2/tags/trivial-branch-pull-request:
  hw/watchdog/wdt_i6300esb: Use DEVICE() macro to access DeviceState.qdev
  hw/scsi: Use the QOM BUS() macro to access BusState.qbus
  hw/sd: Use the QOM BUS() macro to access BusState.qbus
  hw/audio/ac97: Use the QOM DEVICE() macro to access DeviceState.qdev
  hw/vfio/pci: Use the QOM DEVICE() macro to access DeviceState.qdev
  hw/usb-storage: Use the QOM DEVICE() macro to access DeviceState.qdev
  hw/isa: Use the QOM DEVICE() macro to access DeviceState.qdev
  hw/s390x/event-facility: Use the QOM BUS() macro to access BusState.qbus
  hw/pci-bridge: Use the QOM BUS() macro to access BusState.qbus
  hw/scsi/vmw_pvscsi: Use qbus_reset_all() directly
  docs/devel/build-system: Update an example
  test: Fix make target check-report.tap
  util: Adjust qemu_guest_getrandom_nofail for Coverity
  vhost: fix incorrect print type
  migration: fix a typo
  hw/rdma: Delete unused headers inclusion

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

virtio, pci, pc: cleanups, features

stricter rules for acpi tables: we now fail
on any difference that isn't whitelisted.

vhost-scsi migration.

some cleanups all over the place

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Wed 05 Jun 2019 20:55:04 BST
# gpg:                using RSA key 281F0DB8D28D5469
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>" [full]
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>" [full]
# 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:
  bios-tables-test: ignore identical binaries
  tests: acpi: add simple arm/virt testcase
  tests: add expected ACPI tables for arm/virt board
  bios-tables-test: list all tables that differ
  vhost-scsi: Allow user to enable migration
  vhost-scsi: Add VMState descriptor
  vhost-scsi: The vhost backend should be stopped when the VM is not running
  bios-tables-test: add diff allowed list
  vhost: fix memory leak in vhost_user_scsi_realize
  vhost: fix incorrect print type
  vhost: remove the dead code
  docs: smbios: remove family=x from type2 entry description
  pci: Fold pci_get_bus_devfn() into its sole caller
  pci: Make is_bridge a bool
  pcie: Simplify pci_adjust_config_limit()
  acpi: pci: use build_append_foo() API to construct MCFG
  hw/acpi: Consolidate build_mcfg to pci.c

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

* Fix pr-manager-helper (Markus)

# gpg: Signature made Wed 05 Jun 2019 15:15:34 BST
# gpg:                using RSA key BFFBD25F78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]
# 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:
  vl: Document why objects are delayed
  vl: Fix -drive / -blockdev persistent reservation management

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/watchdog/wdt_i6300esb: Use DEVICE() macro to access DeviceState.qdev

Rather than looking inside the definition of a DeviceState with
"s->qdev", use the QOM prefered style: "DEVICE(s)".

This patch was generated using the following Coccinelle script:

    // Use DEVICE() macros to access DeviceState.qdev
    @use_device_macro_to_access_qdev@
    expression obj;
    identifier dev;
    @@
    -&obj->dev.qdev
    +DEVICE(obj)

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20190528164020.32250-11-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/scsi: Use the QOM BUS() macro to access BusState.qbus

Rather than looking inside the definition of a BusState with "s->bus.qbus",
use the QOM prefered style: "BUS(&s->bus)".

This patch was generated using the following Coccinelle script:

    // Use BUS() macros to access BusState.qbus
    @use_bus_macro_to_access_qbus@
    expression obj;
    identifier bus;
    @@
    -&obj->bus.qbus
    +BUS(&obj->bus)

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20190528164020.32250-3-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

Merge remote-tracking branch 'remotes/juanquintela/tags/migration-pull-request' into staging

Migration Pull request

# gpg: Signature made Wed 05 Jun 2019 12:52:06 BST
# gpg:                using RSA key 1899FF8EDEBF58CCEE034B82F487EF185872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>" [full]
# gpg:                 aka "Juan Quintela <quintela@trasno.org>" [full]
# Primary key fingerprint: 1899 FF8E DEBF 58CC EE03  4B82 F487 EF18 5872 D723

* remotes/juanquintela/tags/migration-pull-request:
  migratioin/ram: leave RAMBlock->bmap blank on allocating
  migration-test: Add a test for fd protocol
  migration: Fix fd protocol for incoming defer
  migration/ram.c: multifd_send_state->count is not really used
  migration/ram.c: MultiFDSendParams.sem_sync is not really used

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/sd: Use the QOM BUS() macro to access BusState.qbus

Rather than looking inside the definition of a BusState with "s->bus.qbus",
use the QOM prefered style: "BUS(&s->bus)".

This patch was generated using the following Coccinelle script:

    // Use BUS() macros to access BusState.qbus
    @use_bus_macro_to_access_qbus@
    expression obj;
    identifier bus;
    @@
    -&obj->bus.qbus
    +BUS(&obj->bus)

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20190528164020.32250-6-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/audio/ac97: Use the QOM DEVICE() macro to access DeviceState.qdev

Rather than looking inside the definition of a DeviceState with
"s->qdev", use the QOM prefered style: "DEVICE(s)".

This patch was generated using the following Coccinelle script
(with a bit of manual fix-up, removing an extra space to please
checkpatch.pl):

    // Use DEVICE() macros to access DeviceState.qdev
    @use_device_macro_to_access_qdev@
    expression obj;
    identifier dev;
    @@
    -&obj->dev.qdev
    +DEVICE(obj)

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>.
Message-Id: <20190528164020.32250-7-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/vfio/pci: Use the QOM DEVICE() macro to access DeviceState.qdev

Rather than looking inside the definition of a DeviceState with
"s->qdev", use the QOM prefered style: "DEVICE(s)".

This patch was generated using the following Coccinelle script:

    // Use DEVICE() macros to access DeviceState.qdev
    @use_device_macro_to_access_qdev@
    expression obj;
    identifier dev;
    @@
    -&obj->dev.qdev
    +DEVICE(obj)

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
Message-Id: <20190528164020.32250-10-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/usb-storage: Use the QOM DEVICE() macro to access DeviceState.qdev

Rather than looking inside the definition of a DeviceState with
"s->qdev", use the QOM prefered style: "DEVICE(s)".

This patch was generated using the following Coccinelle script:

    // Use DEVICE() macros to access DeviceState.qdev
    @use_device_macro_to_access_qdev@
    expression obj;
    identifier dev;
    @@
    -&obj->dev.qdev
    +DEVICE(obj)

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Message-Id: <20190528164020.32250-9-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/isa: Use the QOM DEVICE() macro to access DeviceState.qdev

Rather than looking inside the definition of a DeviceState with
"s->qdev", use the QOM prefered style: "DEVICE(s)".

This patch was generated using the following Coccinelle script:

    // Use DEVICE() macros to access DeviceState.qdev
    @use_device_macro_to_access_qdev@
    expression obj;
    identifier dev;
    @@
    -&obj->dev.qdev
    +DEVICE(obj)

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Message-Id: <20190528164020.32250-8-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/s390x/event-facility: Use the QOM BUS() macro to access BusState.qbus

Rather than looking inside the definition of a BusState with "s->bus.qbus",
use the QOM prefered style: "BUS(&s->bus)".

This patch was generated using the following Coccinelle script:

    // Use BUS() macros to access BusState.qbus
    @use_bus_macro_to_access_qbus@
    expression obj;
    identifier bus;
    @@
    -&obj->bus.qbus
    +BUS(&obj->bus)

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190528164020.32250-5-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/pci-bridge: Use the QOM BUS() macro to access BusState.qbus

Rather than looking inside the definition of a BusState with "s->bus.qbus",
use the QOM prefered style: "BUS(&s->bus)".

This patch was generated using the following Coccinelle script:

    // Use BUS() macros to access BusState.qbus
    @use_bus_macro_to_access_qbus@
    expression obj;
    identifier bus;
    @@
    -&obj->bus.qbus
    +BUS(&obj->bus)

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Message-Id: <20190528164020.32250-4-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/scsi/vmw_pvscsi: Use qbus_reset_all() directly

Since the BusState is accesible from the SCSIBus object,
it is pointless to use qbus_reset_all_fn.
Use qbus_reset_all() directly.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Dmitry Fleytman <dmitry.fleytman@gmail.com>
Message-Id: <20190528164020.32250-2-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

docs/devel/build-system: Update an example

The default-configs/ example added in 717171bd2025 is no
more accurate since fa212a2b8b60 (and various further other
commits).

The Kconfig build system is now in place.
Use the aarch64-softmmu config as example.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190529140504.21580-1-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

test: Fix make target check-report.tap

Fix a fat-fingered invocation of tap-merge.pl in the recipe of target
check-report.tap.

Fixes: 9df43317b82 "test: replace gtester with a TAP driver"
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190604080010.23186-1-armbru@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

util: Adjust qemu_guest_getrandom_nofail for Coverity

Explicitly ignore the return value of qemu_guest_getrandom.
Because we use error_fatal, all errors are already caught.

Fixes: CID 1401701
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190530173824.30699-1-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

vhost: fix incorrect print type

fix incorrect print type in vhost_virtqueue_stop

Signed-off-by: Jie Wang <wangjie88@huawei.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <1556605773-42019-1-git-send-email-wangjie88@huawei.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

migration: fix a typo

'postocpy' should be 'postcopy'.

CC: qemu-trivial@nongnu.org
Signed-off-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20190525062832.18009-1-liq3ea@163.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/rdma: Delete unused headers inclusion

This is a trivial cleanup patch.

Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com>
Reviewed-by: Kamal Heib <kheib@redhat.com>
Message-Id: <20190505105112.22691-1-yuval.shaia@oracle.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

bios-tables-test: ignore identical binaries

when binary of the tables is identical, there is no need to run iasl
to check that they are functionally equivalent.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vl: Document why objects are delayed

Objects should not be "delayed" without a reason, as the previous
commit demonstrates.  The remaining ones have reasons.  State them.
and demand future ones come with such a statement.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20190604151251.9903-3-armbru@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

vl: Fix -drive / -blockdev persistent reservation management

qemu-system-FOO's main() acts on command line arguments in its own
idiosyncratic order.  There's not much method to its madness.
Whenever we find a case where one kind of command line argument needs
to refer to something created for another kind later, we rejigger the
order.

Recent commit cda4aa9a5a "vl: Create block backends before setting
machine properties" was such a rejigger.  Block backends are now
created before "delayed" objects.  This broke persistent reservation
management.  Reproducer:

    $ qemu-system-x86_64 -object pr-manager-helper,id=pr-helper0,path=/tmp/pr-helper0.sock-drive -drive file=/dev/mapper/crypt,file.pr-manager=pr-helper0,format=raw,if=none,id=drive-scsi0-0-0-2
    qemu-system-x86_64: -drive file=/dev/mapper/crypt,file.pr-manager=pr-helper0,format=raw,if=none,id=drive-scsi0-0-0-2: No persistent reservation manager with id 'pr-helper0'

The delayed pr-manager-helper object is created too late for use by
-drive or -blockdev.  Normal objects are still created in time.

pr-manager-helper has always been a delayed object (commit 7c9e527659
"scsi, file-posix: add support for persistent reservation
management").  Turns out there's no real reason for that.  Make it a
normal object.

Fixes: cda4aa9a5a08777cf13e164c0543bd4888b8adce
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20190604151251.9903-2-armbru@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

migratioin/ram: leave RAMBlock->bmap blank on allocating

During migration, we would sync bitmap from ram_list.dirty_memory to
RAMBlock.bmap in cpu_physical_memory_sync_dirty_bitmap().

Since we set RAMBlock.bmap and ram_list.dirty_memory both to all 1, this
means at the first round this sync is meaningless and is a duplicated
work.

Leaving RAMBlock->bmap blank on allocating would have a side effect on
migration_dirty_pages, since it is calculated from the result of
cpu_physical_memory_sync_dirty_bitmap(). To keep it right, we need to
set migration_dirty_pages to 0 in ram_state_init().

Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration-test: Add a test for fd protocol

Signed-off-by: Yury Kotov <yury-kotov@yandex-team.ru>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration: Fix fd protocol for incoming defer

Currently, incoming migration through fd supports only command-line case:
E.g.
    fork();
    fd = open();
    exec("qemu ... -incoming fd:%d", fd);

It's possible to use add-fd commands to pass fd for migration, but it's
invalid case. add-fd works with fdset but not with particular fds.

To work with getfd in incoming defer it's enough to use monitor_fd_param
instead of strtol. monitor_fd_param supports both cases:
* fd:123
* fd:fd_name (added by getfd).

And also the use of monitor_fd_param improves error messages.

Signed-off-by: Yury Kotov <yury-kotov@yandex-team.ru>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration/ram.c: multifd_send_state->count is not really used

Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration/ram.c: MultiFDSendParams.sem_sync is not really used

Besides init and destroy, MultiFDSendParams.sem_sync is not really used.

Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block layer patches:

- block: AioContext management, part 2
- Avoid recursive block_status call (i.e. lseek() calls) if possible
- linux-aio: Drop unused BlockAIOCB submission method
- nvme: add Get/Set Feature Timestamp support
- Fix crash on commit job start with active I/O on base node
- Fix crash in bdrv_drained_end
- Fix integer overflow in qcow2 discard

# gpg: Signature made Tue 04 Jun 2019 16:20:02 BST
# gpg:                using RSA key 7F09B272C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>" [full]
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* remotes/kevin/tags/for-upstream: (29 commits)
  iotests: Fix duplicated diff output on failure
  iotests: test big qcow2 shrink
  block/io: bdrv_pdiscard: support int64_t bytes parameter
  block/qcow2-refcount: add trace-point to qcow2_process_discards
  block: Remove bdrv_set_aio_context()
  test-bdrv-drain: Use bdrv_try_set_aio_context()
  iotests: Attach new devices to node in non-default iothread
  virtio-scsi-test: Test attaching new overlay with iothreads
  block: Remove wrong bdrv_set_aio_context() calls
  blockdev: Use bdrv_try_set_aio_context() for monitor commands
  block: Move node without parents to main AioContext
  test-block-iothread: BlockBackend AioContext across root node change
  test-block-iothread: Test adding parent to iothread node
  block: Adjust AioContexts when attaching nodes
  scsi-disk: Use qdev_prop_drive_iothread
  block: Add qdev_prop_drive_iothread property type
  block: Add BlockBackend.ctx
  block: Add Error to blk_set_aio_context()
  nbd-server: Call blk_set_allow_aio_context_change()
  test-block-iothread: Check filter node in test_propagate_mirror
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

iotests: Fix duplicated diff output on failure

Commit 70ff5b07 wanted to move the diff between actual and reference
output to the end after printing the test result line. It really only
copied it, though, so the diff is now displayed twice. Remove the old
one.

Fixes: 70ff5b07fcdd378180ad2d5cc0b0d5e67e7ef325
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

iotests: test big qcow2 shrink

This test checks bug in qcow2_process_discards, fixed by previous
commit.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block/io: bdrv_pdiscard: support int64_t bytes parameter

This fixes at least one overflow in qcow2_process_discards, which
passes 64bit region length to bdrv_pdiscard where bytes (or sectors in
the past) parameter is int since its introduction in 0b919fae.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block/qcow2-refcount: add trace-point to qcow2_process_discards

Let's at least trace ignored failure.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Remove bdrv_set_aio_context()

All callers of bdrv_set_aio_context() are eliminated now, they have
moved to bdrv_try_set_aio_context() and related safe functions. Remove
bdrv_set_aio_context().

With this, we can now know that the .set_aio_ctx callback must be
present in bdrv_set_aio_context_ignore() because
bdrv_can_set_aio_context() would have returned false previously, so
instead of checking the condition, we can assert it.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

test-bdrv-drain: Use bdrv_try_set_aio_context()

No reason to use the unchecked version in tests, even more so when these
are the last callers of bdrv_set_aio_context() outside of block.c.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

iotests: Attach new devices to node in non-default iothread

This tests that devices refuse to be attached to a node that has already
been moved to a different iothread if they can't be or aren't configured
to work in the same iothread.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

virtio-scsi-test: Test attaching new overlay with iothreads

This tests that blockdev-add can correctly add a qcow2 overlay to an
image used by a virtio-scsi disk in an iothread. The interesting point
here is whether the newly added node gets correctly moved into the
iothread AioContext.

If it isn't, we get an assertion failure in virtio-scsi while processing
the next request:

    virtio_scsi_ctx_check: Assertion `blk_get_aio_context(d->conf.blk) == s->ctx' failed.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Remove wrong bdrv_set_aio_context() calls

The mirror and commit block jobs use bdrv_set_aio_context() to move
their filter node into the right AioContext before hooking it up in the
graph. Similarly, bdrv_open_backing_file() explicitly moves the backing
file node into the right AioContext first.

This isn't necessary any more, they get automatically moved into the
right context now when attaching them.

However, in the case of bdrv_open_backing_file() with a node reference,
it's actually not only unnecessary, but even wrong: The unchecked
bdrv_set_aio_context() changes the AioContext of the child node even if
other parents require it to retain the old context. So this is not only
a simplification, but a bug fix, too.

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1684342
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

blockdev: Use bdrv_try_set_aio_context() for monitor commands

Monitor commands can handle errors, so they can easily be converted to
using the safer bdrv_try_set_aio_context() function.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Move node without parents to main AioContext

A node should only be in a non-default AioContext if a user is attached
to it that requires this. When the last parent of a node is gone, it can
move back to the main AioContext.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>