OSDN Git Service

qmiga/qemu.git
7 years agovirtio: add virtqueue_alloc_element tracepoint
Paolo Bonzini [Thu, 1 Jun 2017 13:54:47 +0000 (15:54 +0200)]
virtio: add virtqueue_alloc_element tracepoint

This tracepoint can help diagnosing failures due to memory
fragmentation in the guest.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
7 years agovirtio-serial-bus: Unset hotplug handler when unrealize
Ladi Prosek [Tue, 30 May 2017 08:59:43 +0000 (10:59 +0200)]
virtio-serial-bus: Unset hotplug handler when unrealize

Virtio serial device controls the lifetime of virtio-serial-bus and
virtio-serial-bus links back to the device via its hotplug-handler
property. This extra ref-count prevents the device from getting
finalized, leaving the VirtIODevice memory listener registered and
leading to use-after-free later on.

This patch addresses the same issue as Fam Zheng's
"virtio-scsi: Unset hotplug handler when unrealize"
only for a different virtio device.

Cc: qemu-stable@nongnu.org
Signed-off-by: Ladi Prosek <lprosek@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/juanquintela/tags/migration/20170601' into...
Peter Maydell [Fri, 2 Jun 2017 13:07:53 +0000 (14:07 +0100)]
Merge remote-tracking branch 'remotes/juanquintela/tags/migration/20170601' into staging

migration/next for 20170601

# gpg: Signature made Thu 01 Jun 2017 17:51:04 BST
# gpg:                using RSA key 0xF487EF185872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>"
# gpg:                 aka "Juan Quintela <quintela@trasno.org>"
# Primary key fingerprint: 1899 FF8E DEBF 58CC EE03  4B82 F487 EF18 5872 D723

* remotes/juanquintela/tags/migration/20170601:
  migration: Move include/migration/block.h into migration/
  migration: Export ram.c functions in its own file
  migration: Create include for migration snapshots
  migration: Export rdma.c functions in its own file
  migration: Export tls.c functions in its own file
  migration: Export socket.c functions in its own file
  migration: Export fd.c functions in its own file
  migration: Export exec.c functions in its own file
  migration: Split qemu-file.h
  migration: Remove unneeded includes of migration/vmstate.h
  migration: shut src return path unconditionally
  migration: fix leak of src file on dst
  migration: Remove section_id parameter from vmstate_load
  migration: loadvm handlers are not used
  migration: Use savevm_handlers instead of loadvm copy

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20170602' into...
Peter Maydell [Fri, 2 Jun 2017 12:05:06 +0000 (13:05 +0100)]
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20170602' into staging

target-arm queue:
 * virt: numa: provide ACPI distance info when needed
 * aspeed: fix i2c controller bugs
 * M profile: support MPU
 * gicv3: fix mishandling of BPR1, VBPR1
 * load_uboot_image: don't assume a full header read
 * libvixl: Correct build failures on NetBSD

# gpg: Signature made Fri 02 Jun 2017 12:00:42 BST
# gpg:                using RSA key 0x3C2525ED14360CDE
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>"
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>"
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>"
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83  15CF 3C25 25ED 1436 0CDE

* remotes/pmaydell/tags/pull-target-arm-20170602: (25 commits)
  hw/arm/virt: fdt: generate distance-map when needed
  hw/arm/virt-acpi-build: build SLIT when needed
  aspeed: add some I2C devices to the Aspeed machines
  aspeed/i2c: introduce a state machine
  aspeed/i2c: handle LAST command under the RX command
  aspeed/i2c: improve command handling
  arm: Implement HFNMIENA support for M profile MPU
  arm: add MPU support to M profile CPUs
  armv7m: Classify faults as MemManage or BusFault
  arm: All M profile cores are PMSA
  armv7m: Implement M profile default memory map
  armv7m: Improve "-d mmu" tracing for PMSAv7 MPU
  arm: Remove unnecessary check on cpu->pmsav7_dregion
  arm: Don't let no-MPU PMSA cores write to SCTLR.M
  arm: Don't clear ARM_FEATURE_PMSA for no-mpu configs
  arm: Clean up handling of no-MPU PMSA CPUs
  arm: Use different ARMMMUIdx values for M profile
  arm: Add support for M profile CPUs having different MMU index semantics
  arm: Use the mmu_idx we're passed in arm_cpu_do_unaligned_access()
  target/arm: clear PMUVER field of AA64DFR0 when vPMU=off
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agohw/arm/virt: fdt: generate distance-map when needed
Andrew Jones [Fri, 2 Jun 2017 10:51:49 +0000 (11:51 +0100)]
hw/arm/virt: fdt: generate distance-map when needed

This is based on patch Shannon Zhao originally posted.

Cc: Shannon Zhao <zhaoshenglong@huawei.com>
Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Shannon Zhao <shannon.zhao@linaro.org>
Message-id: 20170529173751.3443-3-drjones@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agohw/arm/virt-acpi-build: build SLIT when needed
Andrew Jones [Fri, 2 Jun 2017 10:51:49 +0000 (11:51 +0100)]
hw/arm/virt-acpi-build: build SLIT when needed

Cc: Shannon Zhao <zhaoshenglong@huawei.com>
Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Shannon Zhao <shannon.zhao@linaro.org>
Message-id: 20170529173751.3443-2-drjones@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoaspeed: add some I2C devices to the Aspeed machines
Cédric Le Goater [Fri, 2 Jun 2017 10:51:49 +0000 (11:51 +0100)]
aspeed: add some I2C devices to the Aspeed machines

Let's add an RTC to the palmetto BMC and a LM75 temperature sensor to
the AST2500 EVB to start with.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-id: 1494827476-1487-5-git-send-email-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoaspeed/i2c: introduce a state machine
Cédric Le Goater [Fri, 2 Jun 2017 10:51:49 +0000 (11:51 +0100)]
aspeed/i2c: introduce a state machine

The Aspeed I2C controller maintains a state machine in the command
register, which is mostly used for debug.

Let's start adding a few states to handle abnormal STOP
commands. Today, the model uses the busy status of the bus as a
condition to do so but it is not precise enough.

Also remove the ABNORMAL bit for failing TX commands. This is
incorrect with respect to the specs.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-id: 1494827476-1487-4-git-send-email-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoaspeed/i2c: handle LAST command under the RX command
Cédric Le Goater [Fri, 2 Jun 2017 10:51:49 +0000 (11:51 +0100)]
aspeed/i2c: handle LAST command under the RX command

Today, the LAST command is handled with the STOP command but this is
incorrect. Also nack the I2C bus when a LAST is issued.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-id: 1494827476-1487-3-git-send-email-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoaspeed/i2c: improve command handling
Cédric Le Goater [Fri, 2 Jun 2017 10:51:49 +0000 (11:51 +0100)]
aspeed/i2c: improve command handling

Multiple I2C commands can be fired simultaneously and the controller
execute the commands following these priorities:

  (1) Master Start Command
  (2) Master Transmit Command
  (3) Slave Transmit Command or Master Receive Command
  (4) Master Stop Command

The current code is incorrect with respect to the above sequence and
needs to be reworked to handle each individual command.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-id: 1494827476-1487-2-git-send-email-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoarm: Implement HFNMIENA support for M profile MPU
Peter Maydell [Fri, 2 Jun 2017 10:51:49 +0000 (11:51 +0100)]
arm: Implement HFNMIENA support for M profile MPU

Implement HFNMIENA support for the M profile MPU. This bit controls
whether the MPU is treated as enabled when executing at execution
priorities of less than zero (in NMI, HardFault or with the FAULTMASK
bit set).

Doing this requires us to use a different MMU index for "running
at execution priority < 0", because we will have different
access permissions for that case versus the normal case.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1493122030-32191-14-git-send-email-peter.maydell@linaro.org

7 years agoarm: add MPU support to M profile CPUs
Michael Davidsaver [Fri, 2 Jun 2017 10:51:48 +0000 (11:51 +0100)]
arm: add MPU support to M profile CPUs

The M series MPU is almost the same as the already implemented R
profile MPU (v7 PMSA).  So all we need to implement here is the MPU
register interface in the system register space.

This implementation has the same restriction as the R profile MPU
that it doesn't permit regions to be sized down smaller than 1K.

We also do not yet implement support for MPU_CTRL.HFNMIENA; this
bit should if zero disable use of the MPU when running HardFault,
NMI or with FAULTMASK set to 1 (ie at an execution priority of
less than zero) -- if the MPU is enabled we don't treat these
cases any differently.

Signed-off-by: Michael Davidsaver <mdavidsaver@gmail.com>
Message-id: 1493122030-32191-13-git-send-email-peter.maydell@linaro.org
[PMM: Keep all the bits in mpu_ctrl field, rather than
 using SCTLR bits for them; drop broken HFNMIENA support;
 various cleanup]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoarmv7m: Classify faults as MemManage or BusFault
Michael Davidsaver [Fri, 2 Jun 2017 10:51:48 +0000 (11:51 +0100)]
armv7m: Classify faults as MemManage or BusFault

General logic is that operations stopped by the MPU are MemManage,
and those which go through the MPU and are caught by the unassigned
handle are BusFault. Distinguish these by looking at the
exception.fsr values, and set the CFSR bits and (if appropriate)
fill in the BFAR or MMFAR with the exception address.

Signed-off-by: Michael Davidsaver <mdavidsaver@gmail.com>
Message-id: 1493122030-32191-12-git-send-email-peter.maydell@linaro.org
[PMM: i-side faults do not set BFAR/MMFAR, only d-side;
 added some CPU_LOG_INT logging]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
7 years agoarm: All M profile cores are PMSA
Peter Maydell [Fri, 2 Jun 2017 10:51:48 +0000 (11:51 +0100)]
arm: All M profile cores are PMSA

All M profile CPUs are PMSA, so set the feature bit.
(We haven't actually implemented the M profile MPU register
interface yet, but setting this feature bit gives us closer
to correct behaviour for the MPU-disabled case.)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493122030-32191-11-git-send-email-peter.maydell@linaro.org

7 years agoarmv7m: Implement M profile default memory map
Michael Davidsaver [Fri, 2 Jun 2017 10:51:48 +0000 (11:51 +0100)]
armv7m: Implement M profile default memory map

Add support for the M profile default memory map which is used
if the MPU is not present or disabled.

The main differences in behaviour from implementing this
correctly are that we set the PAGE_EXEC attribute on
the right regions of memory, such that device regions
are not executable.

Signed-off-by: Michael Davidsaver <mdavidsaver@gmail.com>
Message-id: 1493122030-32191-10-git-send-email-peter.maydell@linaro.org
[PMM: rephrased comment and commit message; don't mark
 the flash memory region as not-writable; list all
 the cases in the default map explicitly rather than
 using a 'default' case for the non-executable regions]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoarmv7m: Improve "-d mmu" tracing for PMSAv7 MPU
Michael Davidsaver [Fri, 2 Jun 2017 10:51:48 +0000 (11:51 +0100)]
armv7m: Improve "-d mmu" tracing for PMSAv7 MPU

Improve the "-d mmu" tracing for the PMSAv7 MPU translation
process as an aid in debugging guest MPU configurations:
 * fix a missing newline for a guest-error log
 * report the region number with guest-error or unimp
   logs of bad region register values
 * add a log message for the overall result of the lookup
 * print "0x" prefix for hex values

Signed-off-by: Michael Davidsaver <mdavidsaver@gmail.com>
Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493122030-32191-9-git-send-email-peter.maydell@linaro.org
[PMM: a little tidyup, report region number in all messages
 rather than just one]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoarm: Remove unnecessary check on cpu->pmsav7_dregion
Peter Maydell [Fri, 2 Jun 2017 10:51:48 +0000 (11:51 +0100)]
arm: Remove unnecessary check on cpu->pmsav7_dregion

Now that we enforce both:
 * pmsav7_dregion == 0 implies has_mpu == false
 * PMSA with has_mpu == false means SCTLR.M cannot be set
we can remove a check on pmsav7_dregion from get_phys_addr_pmsav7(),
because we can only reach this code path if the MPU is enabled
(and so region_translation_disabled() returned false).

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493122030-32191-8-git-send-email-peter.maydell@linaro.org

7 years agoarm: Don't let no-MPU PMSA cores write to SCTLR.M
Peter Maydell [Fri, 2 Jun 2017 10:51:48 +0000 (11:51 +0100)]
arm: Don't let no-MPU PMSA cores write to SCTLR.M

If the CPU is a PMSA config with no MPU implemented, then the
SCTLR.M bit should be RAZ/WI, so that the guest can never
turn on the non-existent MPU.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493122030-32191-7-git-send-email-peter.maydell@linaro.org

7 years agoarm: Don't clear ARM_FEATURE_PMSA for no-mpu configs
Peter Maydell [Fri, 2 Jun 2017 10:51:47 +0000 (11:51 +0100)]
arm: Don't clear ARM_FEATURE_PMSA for no-mpu configs

Fix the handling of QOM properties for PMSA CPUs with no MPU:

Allow no-MPU to be specified by either:
 * has-mpu = false
 * pmsav7_dregion = 0
and make setting one imply the other. Don't clear the PMSA
feature bit in this situation.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493122030-32191-6-git-send-email-peter.maydell@linaro.org

7 years agoarm: Clean up handling of no-MPU PMSA CPUs
Peter Maydell [Fri, 2 Jun 2017 10:51:47 +0000 (11:51 +0100)]
arm: Clean up handling of no-MPU PMSA CPUs

ARM CPUs come in two flavours:
 * proper MMU ("VMSA")
 * only an MPU ("PMSA")
For PMSA, the MPU may be implemented, or not (in which case there
is default "always acts the same" behaviour, but it isn't guest
programmable).

QEMU is a bit confused about how we indicate this: we have an
ARM_FEATURE_MPU, but it's not clear whether this indicates
"PMSA, not VMSA" or "PMSA and MPU present" , and sometimes we
use it for one purpose and sometimes the other.

Currently trying to implement a PMSA-without-MPU core won't
work correctly because we turn off the ARM_FEATURE_MPU bit
and then a lot of things which should still exist get
turned off too.

As the first step in cleaning this up, rename the feature
bit to ARM_FEATURE_PMSA, which indicates a PMSA CPU (with
or without MPU).

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493122030-32191-5-git-send-email-peter.maydell@linaro.org

7 years agoarm: Use different ARMMMUIdx values for M profile
Peter Maydell [Fri, 2 Jun 2017 10:51:47 +0000 (11:51 +0100)]
arm: Use different ARMMMUIdx values for M profile

Make M profile use completely separate ARMMMUIdx values from
those that A profile CPUs use. This is a prelude to adding
support for the MPU and for v8M, which together will require
6 MMU indexes which don't map cleanly onto the A profile
uses:
 non secure User
 non secure Privileged
 non secure Privileged, execution priority < 0
 secure User
 secure Privileged
 secure Privileged, execution priority < 0

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1493122030-32191-4-git-send-email-peter.maydell@linaro.org

7 years agoarm: Add support for M profile CPUs having different MMU index semantics
Peter Maydell [Fri, 2 Jun 2017 10:51:47 +0000 (11:51 +0100)]
arm: Add support for M profile CPUs having different MMU index semantics

The M profile CPU's MPU has an awkward corner case which we
would like to implement with a different MMU index.

We can avoid having to bump the number of MMU modes ARM
uses, because some of our existing MMU indexes are only
used by non-M-profile CPUs, so we can borrow one.
To avoid that getting too confusing, clean up the code
to try to keep the two meanings of the index separate.

Instead of ARMMMUIdx enum values being identical to core QEMU
MMU index values, they are now the core index values with some
high bits set. Any particular CPU always uses the same high
bits (so eventually A profile cores and M profile cores will
use different bits). New functions arm_to_core_mmu_idx()
and core_to_arm_mmu_idx() convert between the two.

In general core index values are stored in 'int' types, and
ARM values are stored in ARMMMUIdx types.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1493122030-32191-3-git-send-email-peter.maydell@linaro.org

7 years agoarm: Use the mmu_idx we're passed in arm_cpu_do_unaligned_access()
Peter Maydell [Fri, 2 Jun 2017 10:51:47 +0000 (11:51 +0100)]
arm: Use the mmu_idx we're passed in arm_cpu_do_unaligned_access()

When identifying the DFSR format for an alignment fault, use
the mmu index that we are passed, rather than calling cpu_mmu_index()
to get the mmu index for the current CPU state. This doesn't actually
make any difference since the only cases where the current MMU index
differs from the index used for the load are the "unprivileged
load/store" instructions, and in that case the mmu index may
differ but the translation regime is the same (apart from the
"use from Hyp mode" case which is UNPREDICTABLE).
However it's the more logical thing to do.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493122030-32191-2-git-send-email-peter.maydell@linaro.org

7 years agotarget/arm: clear PMUVER field of AA64DFR0 when vPMU=off
Wei Huang [Fri, 2 Jun 2017 10:51:47 +0000 (11:51 +0100)]
target/arm: clear PMUVER field of AA64DFR0 when vPMU=off

The PMUv3 driver of linux kernel (in arch/arm64/kernel/perf_event.c)
relies on the PMUVER field of id_aa64dfr0_el1 to decide if PMU support
is present or not. This patch clears the PMUVER field under TCG mode
when vPMU=off. Without it, PMUv3 will init insider guest VMs even
with vPMU=off. This patch also removes a redundant line inside the
if-statement.

Signed-off-by: Wei Huang <wei@redhat.com>
Message-id: 1495123889-32301-1-git-send-email-wei@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agohw/intc/arm_gicv3_cpuif: Fix priority masking for NS BPR1
Peter Maydell [Fri, 2 Jun 2017 10:51:47 +0000 (11:51 +0100)]
hw/intc/arm_gicv3_cpuif: Fix priority masking for NS BPR1

When we calculate the mask to use to get the group priority from
an interrupt priority, the way that NS BPR1 is handled differs
from how BPR0 and S BPR1 work -- a BPR1 value of 1 means
the group priority is in bits [7:1], whereas for BPR0 and S BPR1
this is indicated by a 0 BPR value.

Subtract 1 from the BPR value before creating the mask if
we're using the NS BPR value, for both hardware and virtual
interrupts, as the GICv3 pseudocode does, and fix the comments
accordingly.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493226792-3237-4-git-send-email-peter.maydell@linaro.org

7 years agohw/intc/arm_gicv3_cpuif: Don't let BPR be set below its minimum
Peter Maydell [Fri, 2 Jun 2017 10:51:47 +0000 (11:51 +0100)]
hw/intc/arm_gicv3_cpuif: Don't let BPR be set below its minimum

icc_bpr_write() was not enforcing that writing a value below the
minimum for the BPR should behave as if the BPR was set to the
minimum value. This doesn't make a difference for the secure
BPRs (since we define the minimum for the QEMU implementation
as zero) but did mean we were allowing the NS BPR1 to be set to
0 when 1 should be the lowest value.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493226792-3237-3-git-send-email-peter.maydell@linaro.org

7 years agohw/intc/arm_gicv3_cpuif: Fix reset value for VMCR_EL2.VBPR1
Peter Maydell [Fri, 2 Jun 2017 10:51:46 +0000 (11:51 +0100)]
hw/intc/arm_gicv3_cpuif: Fix reset value for VMCR_EL2.VBPR1

We were setting the VBPR1 field of VMCR_EL2 to icv_min_vbpr()
on reset, but this is not correct. The field should reset to
the minimum value of ICV_BPR0_EL1 plus one.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1493226792-3237-2-git-send-email-peter.maydell@linaro.org

7 years agoload_uboot_image: don't assume a full header read
Andrew Jones [Fri, 2 Jun 2017 10:51:46 +0000 (11:51 +0100)]
load_uboot_image: don't assume a full header read

Don't allow load_uboot_image() to proceed when less bytes than
header-size was read.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Message-id: 20170524091315.20284-1-drjones@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agolibvixl: Correct build failures on NetBSD
Kamil Rytarowski [Fri, 2 Jun 2017 10:51:46 +0000 (11:51 +0100)]
libvixl: Correct build failures on NetBSD

Ensure that C99 macros are defined regardless of the inclusion order of
headers in vixl. This is required at least on NetBSD.

The vixl/globals.h headers defines __STDC_CONSTANT_MACROS and must be
included before other system headers.

This file defines unconditionally the following macros, without altering
the original sources:
 - __STDC_CONSTANT_MACROS
 - __STDC_LIMIT_MACROS
 - __STDC_FORMAT_MACROS

Signed-off-by: Kamil Rytarowski <n54@gmx.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20170514051820.15985-1-n54@gmx.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agomigration: Move include/migration/block.h into migration/
Juan Quintela [Fri, 21 Apr 2017 12:31:22 +0000 (14:31 +0200)]
migration: Move include/migration/block.h into migration/

All functions were internal, except blk_mig_init() that is exported in
misc.h now.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Export ram.c functions in its own file
Juan Quintela [Mon, 17 Apr 2017 18:26:27 +0000 (20:26 +0200)]
migration: Export ram.c functions in its own file

All functions are internal except for ram_mig_init().  Create
migration/misc.h for this kind of functions.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Create include for migration snapshots
Juan Quintela [Thu, 20 Apr 2017 12:25:55 +0000 (14:25 +0200)]
migration: Create include for migration snapshots

Start removing migration code from sysemu/sysemu.h.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Export rdma.c functions in its own file
Juan Quintela [Mon, 17 Apr 2017 18:32:36 +0000 (20:32 +0200)]
migration: Export rdma.c functions in its own file

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Export tls.c functions in its own file
Juan Quintela [Wed, 5 Apr 2017 15:45:16 +0000 (17:45 +0200)]
migration: Export tls.c functions in its own file

Just for the functions exported from tls.c.  Notice that we can't
remove the migration/migration.h include from tls.c because it access
directly MigrationState for the tls params.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Export socket.c functions in its own file
Juan Quintela [Wed, 5 Apr 2017 15:40:11 +0000 (17:40 +0200)]
migration: Export socket.c functions in its own file

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Export fd.c functions in its own file
Juan Quintela [Wed, 5 Apr 2017 13:58:29 +0000 (15:58 +0200)]
migration: Export fd.c functions in its own file

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Export exec.c functions in its own file
Juan Quintela [Wed, 5 Apr 2017 13:54:10 +0000 (15:54 +0200)]
migration: Export exec.c functions in its own file

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Split qemu-file.h
Juan Quintela [Thu, 20 Apr 2017 16:52:18 +0000 (18:52 +0200)]
migration: Split qemu-file.h

Split the file into public and internal interfaces.  I have to rename
the external one because we can't have two include files with the same
name in the same directory.  Build system gets confused.  The only
exported functions are the ones that handle basic types.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Remove unneeded includes of migration/vmstate.h
Juan Quintela [Fri, 21 Apr 2017 08:45:29 +0000 (10:45 +0200)]
migration: Remove unneeded includes of migration/vmstate.h

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: shut src return path unconditionally
Peter Xu [Fri, 19 May 2017 06:43:30 +0000 (14:43 +0800)]
migration: shut src return path unconditionally

We were do the shutting off only for postcopy. Now we do this as long as
the source return path is there.

Moving the cleanup of from_src_file there too.

Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agomigration: fix leak of src file on dst
Peter Xu [Fri, 19 May 2017 06:43:29 +0000 (14:43 +0800)]
migration: fix leak of src file on dst

The return path channel is possibly leaked. Fix it.

Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agomigration: Remove section_id parameter from vmstate_load
Juan Quintela [Wed, 24 May 2017 07:28:47 +0000 (09:28 +0200)]
migration: Remove section_id parameter from vmstate_load

Everything else assumes that we always load a device from its own
savevm handler.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: loadvm handlers are not used
Juan Quintela [Wed, 24 May 2017 07:25:50 +0000 (09:25 +0200)]
migration: loadvm handlers are not used

So we remove all traces of them.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agomigration: Use savevm_handlers instead of loadvm copy
Juan Quintela [Wed, 24 May 2017 07:09:58 +0000 (09:09 +0200)]
migration: Use savevm_handlers instead of loadvm copy

There is no reason for having the loadvm_handlers at all.  There is
only one use, and we can use the savevm handlers.

We will remove the loadvm handlers on a following patch.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
--

- Added load_version_id: version_id read from the stream (laurent)
- Added load_section_id: section_id read from the stream (dave)

7 years agoMerge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2017-05-31' into staging
Peter Maydell [Thu, 1 Jun 2017 15:39:16 +0000 (16:39 +0100)]
Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2017-05-31' into staging

QAPI patches for 2017-05-31

# gpg: Signature made Wed 31 May 2017 18:06:39 BST
# gpg:                using RSA key 0x3870B400EB918653
# 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-qapi-2017-05-31:
  qapi: Reject alternates that can't work with keyval_parse()
  tests/qapi-schema: Avoid 'str' in alternate test cases
  qapi: Document visit_type_any() issues with keyval input
  qobject-input-visitor: Reject non-finite numbers with keyval

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/riku/tags/pull-linux-user-20170531' into staging
Peter Maydell [Thu, 1 Jun 2017 14:50:40 +0000 (15:50 +0100)]
Merge remote-tracking branch 'remotes/riku/tags/pull-linux-user-20170531' into staging

Misc linux-user updates

# gpg: Signature made Wed 31 May 2017 12:33:17 BST
# gpg:                using RSA key 0xB44890DEDE3C9BC0
# gpg: Good signature from "Riku Voipio <riku.voipio@iki.fi>"
# gpg:                 aka "Riku Voipio <riku.voipio@linaro.org>"
# Primary key fingerprint: FF82 03C8 C391 98AE 0581  41EF B448 90DE DE3C 9BC0

* remotes/riku/tags/pull-linux-user-20170531:
  linux-user: add strace support for uinfo structure of rt_sigqueueinfo() and rt_tgsigqueueinfo()
  linux-user: fix inconsistent spaces in print_siginfo() output
  linux-user: add rt_tgsigqueueinfo() strace
  linux-user: add support for rt_tgsigqueueinfo() system call
  linux-user: fix argument type declaration of rt_sigqueinfo() syscall
  linux-user: fix mismatch of lock/unlock_user() invocations in rt_sigqueinfo() syscall
  linux-user: fix ssetmask() system call
  linux-user: add tkill(), tgkill() and rt_sigqueueinfo() strace
  linux-user: add strace for getuid(), gettid(), getppid(), geteuid()
  linux-user: remove all traces of qemu from /proc/self/cmdline
  linux-user: allocate heap memory for execve arguments
  linux-user: fix inotify
  linux-user: fix fadvise64_64() on ppc
  linux-user: fix eventfd
  linux-user: call fd_trans_target_to_host_data() for write()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/juanquintela/tags/migration/20170531' into...
Peter Maydell [Thu, 1 Jun 2017 14:01:59 +0000 (15:01 +0100)]
Merge remote-tracking branch 'remotes/juanquintela/tags/migration/20170531' into staging

migration/next for 20170531

# gpg: Signature made Wed 31 May 2017 08:53:06 BST
# gpg:                using RSA key 0xF487EF185872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>"
# gpg:                 aka "Juan Quintela <quintela@trasno.org>"
# Primary key fingerprint: 1899 FF8E DEBF 58CC EE03  4B82 F487 EF18 5872 D723

* remotes/juanquintela/tags/migration/20170531:
  migration: use dirty_rate_high_cnt more aggressively
  migration: set bytes_xfer_* outside of autoconverge logic
  migration: set dirty_pages_rate before autoconverge logic
  migration: keep bytes_xfer_prev init'd to zero
  migration: Create savevm.h for functions exported from savevm.c

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/aurel/tags/pull-target-sh4-20170530' into staging
Peter Maydell [Thu, 1 Jun 2017 12:12:20 +0000 (13:12 +0100)]
Merge remote-tracking branch 'remotes/aurel/tags/pull-target-sh4-20170530' into staging

Queued target/sh4 patches

# gpg: Signature made Tue 30 May 2017 20:12:10 BST
# gpg:                using RSA key 0xBA9C78061DDD8C9B
# gpg: Good signature from "Aurelien Jarno <aurelien@aurel32.net>"
# gpg:                 aka "Aurelien Jarno <aurelien@jarno.fr>"
# gpg:                 aka "Aurelien Jarno <aurel32@debian.org>"
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 7746 2642 A9EF 94FD 0F77  196D BA9C 7806 1DDD 8C9B

* remotes/aurel/tags/pull-target-sh4-20170530:
  target/sh4: fix RTE instruction delay slot
  target/sh4: ignore interrupts in a delay slot
  target/sh4: introduce DELAY_SLOT_MASK
  target/sh4: fix reset when using a kernel and an initrd
  target/sh4: log unauthorized accesses using qemu_log_mask

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/gkurz/tags/for-upstream' into staging
Peter Maydell [Thu, 1 Jun 2017 11:06:58 +0000 (12:06 +0100)]
Merge remote-tracking branch 'remotes/gkurz/tags/for-upstream' into staging

Various bugfixes and code cleanups. Most notably, it fixes metadata handling in
mapped-file security mode (especially for the virtfs root).

# gpg: Signature made Tue 30 May 2017 14:36:22 BST
# gpg:                using DSA key 0x02FC3AEB0101DBC2
# gpg: Good signature from "Greg Kurz <groug@kaod.org>"
# gpg:                 aka "Greg Kurz <groug@free.fr>"
# gpg:                 aka "Greg Kurz <gkurz@linux.vnet.ibm.com>"
# gpg:                 aka "Gregory Kurz (Groug) <groug@free.fr>"
# gpg:                 aka "[jpeg image of size 3330]"
# 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: 2BD4 3B44 535E C0A7 9894  DBA2 02FC 3AEB 0101 DBC2

* remotes/gkurz/tags/for-upstream:
  9pfs: local: metadata file for the VirtFS root
  9pfs: local: simplify file opening
  9pfs: local: resolve special directories in paths
  9pfs: check return value of v9fs_co_name_to_path()
  util: drop old utimensat() compat code
  9pfs: assume utimensat() and futimens() are present
  fsdev: fix virtfs-proxy-helper cwd
  9pfs: local: fix unlink of alien files in mapped-file mode
  9pfs: drop pdu_push_and_notify()
  fsdev: don't allow unknown format in marshal/unmarshal
  virtio-9p/xen-9p: move 9p specific bits to core 9p code

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/ehabkost/tags/numa-pull-request' into staging
Peter Maydell [Thu, 1 Jun 2017 09:58:47 +0000 (10:58 +0100)]
Merge remote-tracking branch 'remotes/ehabkost/tags/numa-pull-request' into staging

NUMA fixes, 2017-05-30

# gpg: Signature made Tue 30 May 2017 20:10:44 BST
# gpg:                using RSA key 0x2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>"
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* remotes/ehabkost/tags/numa-pull-request:
  numa: Fix format string for "Invalid node" message
  numa-test: fix query-cpus leaks

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoqapi: Reject alternates that can't work with keyval_parse()
Markus Armbruster [Mon, 22 May 2017 16:42:15 +0000 (18:42 +0200)]
qapi: Reject alternates that can't work with keyval_parse()

Alternates are sum types like unions, but use the JSON type on the
wire / QType in QObject instead of an explicit tag.  That's why we
require alternate members to have distinct QTypes.

The recently introduced keyval_parse() (commit d454dbe) can only
produce string scalars.  The qobject_input_visitor_new_keyval() input
visitor mostly hides the difference, so code using a QObject input
visitor doesn't have to care whether its input was parsed from JSON or
KEY=VALUE,...  The difference leaks for alternates, as noted in commit
0ee9ae7: a non-string, non-enum scalar alternate value can't currently
be expressed.

In part, this is just our insufficiently sophisticated implementation.
Consider alternate type 'GuestFileWhence'.  It has an integer member
and a 'QGASeek' member.  The latter is an enumeration with values
'set', 'cur', 'end'.  The meaning of b=set, b=cur, b=end, b=0, b=1 and
so forth is perfectly obvious.  However, our current implementation
falls apart at run time for b=0, b=1, and so forth.  Fixable, but not
today; add a test case and a TODO comment.

Now consider an alternate type with a string and an integer member.
What's the meaning of a=42?  Is it the string "42" or the integer 42?
Whichever meaning you pick makes the other inexpressible.  This isn't
just an implementation problem, it's fundamental.  Our current
implementation will pick string.

So far, we haven't needed such alternates.  To make sure we stop and
think before we add one that cannot sanely work with keyval_parse(),
let's require alternate members to have sufficiently distinct
representation in KEY=VALUE,... syntax:

* A string member clashes with any other scalar member

* An enumeration member clashes with bool members when it has value
  'on' or 'off'.

* An enumeration member clashes with numeric members when it has a
  value that starts with '-', '+', or a decimal digit.  This is a
  rather lazy approximation of the actual number syntax accepted by
  the visitor.

  Note that enumeration values starting with '-' and '+' are rejected
  elsewhere already, but better safe than sorry.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1495471335-23707-5-git-send-email-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
7 years agotests/qapi-schema: Avoid 'str' in alternate test cases
Markus Armbruster [Mon, 22 May 2017 16:42:14 +0000 (18:42 +0200)]
tests/qapi-schema: Avoid 'str' in alternate test cases

The next commit is going to make alternate members of type 'str'
conflict with other scalar types.  Would break a few test cases that
don't actually require 'str'.  Flip them from 'str' to 'bool' or
'EnumOne'.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1495471335-23707-4-git-send-email-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
7 years agoqapi: Document visit_type_any() issues with keyval input
Markus Armbruster [Mon, 22 May 2017 16:42:13 +0000 (18:42 +0200)]
qapi: Document visit_type_any() issues with keyval input

It's already documented in keyval.c (commit 0ee9ae7), but visitor.h
can use a note, too.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1495471335-23707-3-git-send-email-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
7 years agoqobject-input-visitor: Reject non-finite numbers with keyval
Markus Armbruster [Mon, 22 May 2017 16:42:12 +0000 (18:42 +0200)]
qobject-input-visitor: Reject non-finite numbers with keyval

The QObject input visitor can produce only finite numbers when its
input comes out of the JSON parser, because the the JSON parser
implements RFC 7159, which provides no syntax for infinity and NaN.

However, it can produce infinity and NaN when its input comes out of
keyval_parse(), because we parse with strtod() then.

The keyval variant should not be able to express things the JSON
variant can't.  Rejecting non-finite numbers there is the conservative
fix.  It's also minimally invasive.

We could instead extend our JSON dialect to provide for infinity and
NaN.  Not today.

Note that the JSON formatter can emit non-finite numbers (marked FIXME
in commit 6e8e5cb).

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1495471335-23707-2-git-send-email-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
7 years agomigration: use dirty_rate_high_cnt more aggressively
Felipe Franciosi [Wed, 24 May 2017 16:10:03 +0000 (17:10 +0100)]
migration: use dirty_rate_high_cnt more aggressively

The commit message from 070afca25 suggests that dirty_rate_high_cnt
should be used more aggressively to start throttling after two
iterations instead of four. The code, however, only changes the auto
convergence behaviour to throttle after three iterations. This makes the
behaviour more aggressive by kicking off throttling after two iterations
as originally intended.

Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agomigration: set bytes_xfer_* outside of autoconverge logic
Felipe Franciosi [Wed, 24 May 2017 16:10:02 +0000 (17:10 +0100)]
migration: set bytes_xfer_* outside of autoconverge logic

The bytes_xfer_now/prev counters are only used by the auto convergence
logic. However, they are used alongside the dirty_pages_rate counter,
which is calculated (and required) outside of this logic. The problem
with this approach is that if the auto convergence capability is changed
while a migration is ongoing, the relationship of the counters will be
broken.

This moves the management of bytes_xfer_now/prev counters outside of the
auto convergence logic to address this issue.

Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agomigration: set dirty_pages_rate before autoconverge logic
Felipe Franciosi [Wed, 24 May 2017 16:10:01 +0000 (17:10 +0100)]
migration: set dirty_pages_rate before autoconverge logic

Currently, a "period" in the RAM migration logic is at least a second
long and accounts for what happened since the last period (or the
beginning of the migration). The dirty_pages_rate counter is calculated
at the end this logic.

If the auto convergence capability is enabled from the start of the
migration, it won't be able to use this counter the first time around.
This calculates dirty_pages_rate as soon as a period is deemed over,
which allows for it to be used immediately.

Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agomigration: keep bytes_xfer_prev init'd to zero
Felipe Franciosi [Wed, 24 May 2017 16:10:00 +0000 (17:10 +0100)]
migration: keep bytes_xfer_prev init'd to zero

The first time migration_bitmap_sync() is called, bytes_xfer_prev is set
to ram_state.bytes_transferred which is, at this point, zero. The next
time migration_bitmap_sync() is called, an iteration has happened and
bytes_xfer_prev is set to 'x' bytes. Most likely, more than one second
has passed, so the auto converge logic will be triggered and
bytes_xfer_now will also be set to 'x' bytes.

This condition is currently masked by dirty_rate_high_cnt, which will
wait for a few iterations before throttling. It would otherwise always
assume zero bytes have been copied and therefore throttle the guest
(possibly) prematurely.

Given bytes_xfer_prev is only used by the auto convergence logic, it
makes sense to only set its value after a check has been made against
bytes_xfer_now.

Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agomigration: Create savevm.h for functions exported from savevm.c
Juan Quintela [Thu, 20 Apr 2017 12:48:46 +0000 (14:48 +0200)]
migration: Create savevm.h for functions exported from savevm.c

This removes last trace of migration functions from sysemu/sysemu.h.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
7 years agonuma: Fix format string for "Invalid node" message
Eduardo Habkost [Tue, 30 May 2017 18:40:13 +0000 (15:40 -0300)]
numa: Fix format string for "Invalid node" message

Some compilers complain about the PRIu16 format string with the
MAX(src, dst) and MAX_NODES arguments.  Example output from Apple LLVM
version 7.3.0 (clang-703.0.31):

  numa.c:236:20: warning: format specifies type 'unsigned short' but the argument has type 'int' [-Wformat]
                     MAX(src, dst), MAX_NODES);
  ~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~
  include/qapi/error.h:163:35: note: expanded from macro 'error_setg'
                          (fmt), ## __VA_ARGS__)
                                    ^~~~~~~~~~~
  glib/2.52.2/include/glib-2.0/glib/gmacros.h:288:20: note: expanded from macro 'MAX'
  #define MAX(a, b)  (((a) > (b)) ? (a) : (b))
                     ^~~~~~~~~~~~~~~~~~~~~~~~~
  numa.c:236:35: warning: format specifies type 'unsigned short' but the argument has type 'int' [-Wformat]
                     MAX(src, dst), MAX_NODES);
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~
  include/qapi/error.h:163:35: note: expanded from macro 'error_setg'
                          (fmt), ## __VA_ARGS__)
                                    ^~~~~~~~~~~
  include/sysemu/sysemu.h:165:19: note: expanded from macro 'MAX_NODES'
  #define MAX_NODES 128
                    ^~~
MAX(src, dst) promotes the src and dst arguments to int, and MAX_NODES
is an int.  Use %d to silence those warnings.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20170530184013.31044-1-ehabkost@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
7 years agonuma-test: fix query-cpus leaks
Marc-André Lureau [Fri, 26 May 2017 11:04:55 +0000 (15:04 +0400)]
numa-test: fix query-cpus leaks

Fix test leaks introduced in commit 2941020a476.

(and small extra space removed)

Spotted by ASAN.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170526110456.32004-1-marcandre.lureau@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
7 years agotarget/sh4: fix RTE instruction delay slot
Aurelien Jarno [Tue, 16 May 2017 22:48:18 +0000 (00:48 +0200)]
target/sh4: fix RTE instruction delay slot

The ReTurn from Exception (RTE) instruction loads the system register
(SR) with the saved system register (SSR). It has a delay slot, and
behaves specially according to the SH4 manual:

  The SR value accessed by the instruction in the RTE delay slot is the
  value restored from SSR by the RTE instruction. The SR and MD values
  defined prior to RTE execution are used to fetch the instruction in
  the RTE delay slot.

The instruction in the delay slot being often a NOP, it doesn't cause
any issue most of the time except in some rare cases where the NOP is
being splitted in a different TB (for example when the TCG op buffer
is full). In that case the NOP is fetched with the user permissions
and causes an instruction TLB protection violation exception.

This patches fixes that by introducing a new delay slot flag for the
RTE instruction. Given it's a privileged instruction, the RTE delay
slot instruction is always fetched in privileged mode. It is therefore
enough to to check for this flag in cpu_mmu_index.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
7 years agotarget/sh4: ignore interrupts in a delay slot
Aurelien Jarno [Tue, 16 May 2017 22:48:18 +0000 (00:48 +0200)]
target/sh4: ignore interrupts in a delay slot

Delay slots are indivisible, therefore avoid scheduling an interrupt in
the delay slot. However exceptions are possible.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
7 years agotarget/sh4: introduce DELAY_SLOT_MASK
Aurelien Jarno [Tue, 16 May 2017 22:48:18 +0000 (00:48 +0200)]
target/sh4: introduce DELAY_SLOT_MASK

This will make easier the introduction of a new flag in the next
patches.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
7 years agotarget/sh4: fix reset when using a kernel and an initrd
Aurelien Jarno [Tue, 16 May 2017 22:48:18 +0000 (00:48 +0200)]
target/sh4: fix reset when using a kernel and an initrd

When a masked exception happens, the SH4 CPU generates a non-masked
reset exception, which then jumps to the reset vector at address
0xA0000000. While this is emulated correctly in QEMU, this does not
work when using a kernel and initrd as this address then contain an
illegal instruction (and there is no guarantee the kernel and initrd
haven't been overwritten).

Therefore call qemu_system_reset_request to reload the kernel and initrd
and load the program counter to the kernel entry point.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
7 years agotarget/sh4: log unauthorized accesses using qemu_log_mask
Aurelien Jarno [Tue, 16 May 2017 22:48:18 +0000 (00:48 +0200)]
target/sh4: log unauthorized accesses using qemu_log_mask

qemu_log_mask() is preferred over fprintf() for logging errors.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
7 years agoMerge remote-tracking branch 'kwolf/tags/for-upstream' into staging
Stefan Hajnoczi [Tue, 30 May 2017 13:15:15 +0000 (14:15 +0100)]
Merge remote-tracking branch 'kwolf/tags/for-upstream' into staging

Block layer patches

# gpg: Signature made Mon 29 May 2017 03:34:59 PM BST
# gpg:                using RSA key 0x7F09B272C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* kwolf/tags/for-upstream:
  block/file-*: *_parse_filename() and colons
  block: Fix backing paths for filenames with colons
  block: Tweak error message related to qemu-img amend
  qemu-img: Fix leakage of options on error
  qemu-img: copy *key-secret opts when opening newly created files
  qemu-img: introduce --target-image-opts for 'convert' command
  qemu-img: fix --image-opts usage with dd command
  qemu-img: add support for --object with 'dd' command
  qemu-img: Fix documentation of convert
  qcow2: remove extra local_error variable
  mirror: Drop permissions on s->target on completion
  nvme: Add support for Controller Memory Buffers
  iotests: 147: Don't test inet6 if not available
  qemu-iotests: Test streaming with missing job ID
  stream: fix crash in stream_start() when block_job_create() fails

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'kraxel/tags/pull-usb-20170529-1' into staging
Stefan Hajnoczi [Tue, 30 May 2017 13:15:09 +0000 (14:15 +0100)]
Merge remote-tracking branch 'kraxel/tags/pull-usb-20170529-1' into staging

usb: depricate legacy options and hmp commands
usb: fixes for ehci and hub, split xhci variants

# gpg: Signature made Mon 29 May 2017 02:07:17 PM BST
# gpg:                using RSA key 0x4CB6D8EED3E87138
# 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

* kraxel/tags/pull-usb-20170529-1:
  ehci: fix frame timer invocation.
  usb: don't wakeup during coldplug
  usb-hub: set PORT_STAT_C_SUSPEND on host-initiated wake-up
  xhci: add CONFIG_USB_XHCI_NEC option
  xhci: split into multiple files
  usb: Simplify the parameter parsing of the legacy usb serial device
  usb: Deprecate HMP commands usb_add and usb_del
  usb: Deprecate the legacy -usbdevice option
  ehci: fix overflow in frame timer code

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'mst/tags/for_upstream' into staging
Stefan Hajnoczi [Tue, 30 May 2017 13:15:04 +0000 (14:15 +0100)]
Merge remote-tracking branch 'mst/tags/for_upstream' into staging

pci, virtio, vhost: fixes

A bunch of fixes all over the place. Most notably this fixes
the new MTU feature when using vhost.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Mon 29 May 2017 01:10:24 AM BST
# gpg:                using RSA key 0x281F0DB8D28D5469
# 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

* mst/tags/for_upstream:
  acpi-test: update expected files
  pc: ACPI BIOS: use highest NUMA node for hotplug mem hole SRAT entry
  vhost-user: pass message as a pointer to process_message_reply()
  virtio_net: Bypass backends for MTU feature negotiation
  intel_iommu: turn off pt before 2.9
  intel_iommu: support passthrough (PT)
  intel_iommu: allow dev-iotlb context entry conditionally
  intel_iommu: use IOMMU_ACCESS_FLAG()
  intel_iommu: provide vtd_ce_get_type()
  intel_iommu: renaming context entry helpers
  x86-iommu: use DeviceClass properties
  memory: remove the last param in memory_region_iommu_replay()
  memory: tune last param of iommu_ops.translate()

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'sthibault/tags/samuel-thibault' into staging
Stefan Hajnoczi [Tue, 30 May 2017 13:14:57 +0000 (14:14 +0100)]
Merge remote-tracking branch 'sthibault/tags/samuel-thibault' into staging

slirp updates

# gpg: Signature made Sat 27 May 2017 10:36:33 PM BST
# gpg:                using RSA key 0xB0A51BF58C9179C5
# gpg: Good signature from "Samuel Thibault <samuel.thibault@aquilenet.fr>"
# gpg:                 aka "Samuel Thibault <sthibault@debian.org>"
# gpg:                 aka "Samuel Thibault <samuel.thibault@gnu.org>"
# gpg:                 aka "Samuel Thibault <samuel.thibault@inria.fr>"
# gpg:                 aka "Samuel Thibault <samuel.thibault@labri.fr>"
# gpg:                 aka "Samuel Thibault <samuel.thibault@ens-lyon.org>"
# gpg:                 aka "Samuel Thibault <samuel.thibault@u-bordeaux.fr>"
# Primary key fingerprint: 900C B024 B679 31D4 0F82  304B D017 8C76 7D06 9EE6
#      Subkey fingerprint: AEBF 7448 FAB9 453A 4552  390E B0A5 1BF5 8C91 79C5

* sthibault/tags/samuel-thibault:
  Fix total IP header length in forwarded TCP packets
  slirp: fix leak
  slirp: Fix wrong mss bug.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'jtc/tags/block-pull-request' into staging
Stefan Hajnoczi [Tue, 30 May 2017 13:14:48 +0000 (14:14 +0100)]
Merge remote-tracking branch 'jtc/tags/block-pull-request' into staging

# gpg: Signature made Fri 26 May 2017 08:22:27 PM BST
# gpg:                using RSA key 0xBDBE7B27C0DE3057
# gpg: Good signature from "Jeffrey Cody <jcody@redhat.com>"
# gpg:                 aka "Jeffrey Cody <jeff@codyprime.org>"
# gpg:                 aka "Jeffrey Cody <codyprime@gmail.com>"
# Primary key fingerprint: 9957 4B4D 3474 90E7 9D98  D624 BDBE 7B27 C0DE 3057

* jtc/tags/block-pull-request:
  block/gluster: glfs_lseek() workaround
  blockjob: use deferred_to_main_loop to indicate the coroutine has ended
  blockjob: reorganize block_job_completed_txn_abort
  blockjob: strengthen a bit test-blockjob-txn
  blockjob: group BlockJob transaction functions together
  blockjob: introduce block_job_cancel_async, check iostatus invariants
  blockjob: move iostatus reset inside block_job_user_resume
  blockjob: separate monitor and blockjob APIs
  blockjob: introduce block_job_pause/resume_all
  blockjob: introduce block_job_early_fail
  blockjob: remove iostatus_reset callback
  blockjob: remove unnecessary check

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'dgibson/tags/ppc-for-2.10-20170525' into staging
Stefan Hajnoczi [Tue, 30 May 2017 08:44:54 +0000 (09:44 +0100)]
Merge remote-tracking branch 'dgibson/tags/ppc-for-2.10-20170525' into staging

ppc patch queue 2017-05-25

Assorted accumulated patches.  These are nearly all bugfixes at one
level or another - some for longstanding problems, others for some
regressions caused by more recent cleanups.

This includes preliminary patches towards fixing migration for Radix
Page Table guests under POWER9 and also fixing some migration
regressions due to the re-organization of the interrupt controller
code.  Not all the pieces are there yet, so those still won't quite
work, but the preliminary changes make sense on their own.

# gpg: Signature made Thu 25 May 2017 04:50:00 AM BST
# gpg:                using RSA key 0x6C38CACA20D9B392
# gpg: Good signature from "David Gibson <david@gibson.dropbear.id.au>"
# gpg:                 aka "David Gibson (kernel.org) <dwg@kernel.org>"
# gpg:                 aka "David Gibson (Red Hat) <dgibson@redhat.com>"
# gpg:                 aka "David Gibson (ozlabs.org) <dgibson@ozlabs.org>"
# Primary key fingerprint: 75F4 6586 AE61 A66C C44E  87DC 6C38 CACA 20D9 B392

* dgibson/tags/ppc-for-2.10-20170525:
  xics: add unrealize handler
  hw/ppc/spapr.c: recover pending LMB unplug info in spapr_lmb_release
  hw/ppc: migrating the DRC state of hotplugged devices
  hw/ppc: removing drc->detach_cb and drc->detach_cb_opaque
  hw/ppc/spapr.c: adding pending_dimm_unplugs to sPAPRMachineState
  spapr: add pre_plug function for memory
  pseries: Restore support for total vcpus not a multiple of threads-per-core for old machine types
  pseries: Split CAS PVR negotiation out into a separate function
  spapr: fix error reporting in xics_system_init()
  spapr_cpu_core: drop reference on ICP object during CPU realization
  hw/ppc/spapr_events.c: removing 'exception' from sPAPREventLogEntry
  spapr: ensure core_slot isn't NULL in spapr_core_unplug()
  xics_kvm: cache already enabled vCPU ids
  spapr: Consolidate HPT freeing code into a routine
  spapr-cpu-core: release ICP object when realization fails
  spapr: sanitize error handling in spapr_ics_create()
  ppc/xics: simplify prototype of xics_spapr_init()
  target/ppc: reset reservation in do_rfi()

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'armbru/tags/pull-qapi-2017-05-23' into staging
Stefan Hajnoczi [Tue, 30 May 2017 08:33:36 +0000 (09:33 +0100)]
Merge remote-tracking branch 'armbru/tags/pull-qapi-2017-05-23' into staging

QAPI patches for 2017-05-23

# gpg: Signature made Tue 23 May 2017 12:33:32 PM BST
# gpg:                using RSA key 0x3870B400EB918653
# 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

* armbru/tags/pull-qapi-2017-05-23:
  qapi-schema: Remove obsolete note from ObjectTypeInfo
  block: Use QDict helpers for --force-share
  shutdown: Expose bool cause in SHUTDOWN and RESET events
  shutdown: Add source information to SHUTDOWN and RESET
  shutdown: Preserve shutdown cause through replay
  shutdown: Prepare for use of an enum in reset/shutdown_request
  shutdown: Simplify shutdown_signal
  sockets: Plug memory leak in socket_address_flatten()
  scripts/qmp/qom-set: fix the value argument passed to srv.command()

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'ehabkost/tags/numa-pull-request' into staging
Stefan Hajnoczi [Tue, 30 May 2017 08:31:05 +0000 (09:31 +0100)]
Merge remote-tracking branch 'ehabkost/tags/numa-pull-request' into staging

Silence "make check" warnings on NUMA test

# gpg: Signature made Tue 23 May 2017 11:44:24 AM BST
# gpg:                using RSA key 0x2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>"
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* ehabkost/tags/numa-pull-request:
  numa: Silence incomplete mapping warning under qtest

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'mreitz/tags/pull-block-2017-05-29-v3' into queue-block
Kevin Wolf [Mon, 29 May 2017 14:34:27 +0000 (16:34 +0200)]
Merge remote-tracking branch 'mreitz/tags/pull-block-2017-05-29-v3' into queue-block

Block patches for the block queue

# gpg: Signature made Mon May 29 16:32:16 2017 CEST
# gpg:                using RSA key 0xF407DB0061D5CF40
# gpg: Good signature from "Max Reitz <mreitz@redhat.com>"
# Primary key fingerprint: 91BE B60A 30DB 3E88 57D1  1829 F407 DB00 61D5 CF40

* mreitz/tags/pull-block-2017-05-29-v3:
  block/file-*: *_parse_filename() and colons
  block: Fix backing paths for filenames with colons
  block: Tweak error message related to qemu-img amend
  qemu-img: Fix leakage of options on error
  qemu-img: copy *key-secret opts when opening newly created files
  qemu-img: introduce --target-image-opts for 'convert' command
  qemu-img: fix --image-opts usage with dd command
  qemu-img: add support for --object with 'dd' command
  qemu-img: Fix documentation of convert
  qcow2: remove extra local_error variable

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
7 years agoblock/file-*: *_parse_filename() and colons
Max Reitz [Mon, 22 May 2017 19:52:16 +0000 (21:52 +0200)]
block/file-*: *_parse_filename() and colons

The file drivers' *_parse_filename() implementations just strip the
optional protocol prefix off the filename. However, for e.g.
"file:foo:bar", this would lead to "foo:bar" being stored as the BDS's
filename which looks like it should be managed using the "foo" protocol.
This is especially troublesome if you then try to resolve a backing
filename based on "foo:bar".

This issue can only occur if the stripped part is a relative filename
("file:/foo:bar" will be shortened to "/foo:bar" and having a slash
before the first colon means that "/foo" is not recognized as a protocol
part). Therefore, we can easily fix it by prepending "./" to such
filenames.

Before this patch:
$ ./qemu-img create -f qcow2 backing.qcow2 64M
Formatting 'backing.qcow2', fmt=qcow2 size=67108864 encryption=off
    cluster_size=65536 lazy_refcounts=off refcount_bits=16
$ ./qemu-img create -f qcow2 -b backing.qcow2 file:top:image.qcow2
Formatting 'file:top:image.qcow2', fmt=qcow2 size=67108864
    backing_file=backing.qcow2 encryption=off cluster_size=65536
    lazy_refcounts=off refcount_bits=16
$ ./qemu-io file:top:image.qcow2
can't open device file:top:image.qcow2: Could not open backing file:
    Unknown protocol 'top'

After this patch:
$ ./qemu-io file:top:image.qcow2
[no error]

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170522195217.12991-3-mreitz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agoblock: Fix backing paths for filenames with colons
Max Reitz [Mon, 22 May 2017 19:52:15 +0000 (21:52 +0200)]
block: Fix backing paths for filenames with colons

path_combine() naturally tries to preserve a protocol prefix. However,
it recognizes such a prefix by scanning for the first colon; which is
different from what path_has_protocol() does: There only is a protocol
prefix if there is a colon before the first slash.

A protocol prefix that is not recognized by path_has_protocol() is none,
and should thus not be taken as one.

Case in point, before this patch:
$ ./qemu-img create -f qcow2 -b backing.qcow2 ./top:image.qcow2
qemu-img: ./top:image.qcow2: Could not open './top:backing.qcow2':
    No such file or directory

Afterwards:
$ ./qemu-img create -f qcow2 -b backing.qcow2 ./top:image.qcow2
qemu-img: ./top:image.qcow2: Could not open './backing.qcow2':
    No such file or directory

Reported-by: yangyang <yangyang@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20170522195217.12991-2-mreitz@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agoblock: Tweak error message related to qemu-img amend
Eric Blake [Mon, 8 May 2017 17:13:02 +0000 (12:13 -0500)]
block: Tweak error message related to qemu-img amend

When converting a 1.1 image down to 0.10, qemu-iotests 060 forces
a contrived failure where allocating a cluster used to replace a
zero cluster reads unaligned data.  Since it is a zero cluster
rather than a data cluster being converted, changing the error
message to match our earlier change in 'qcow2: Make distinction
between zero cluster types obvious' is worthwhile.

Suggested-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-id: 20170508171302.17805-1-eblake@redhat.com
[mreitz: Commit message fixes]
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agoqemu-img: Fix leakage of options on error
Fam Zheng [Mon, 15 May 2017 14:10:14 +0000 (22:10 +0800)]
qemu-img: Fix leakage of options on error

Reported by Coverity.

Signed-off-by: Fam Zheng <famz@redhat.com>
Message-id: 20170515141014.25793-1-famz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agoqemu-img: copy *key-secret opts when opening newly created files
Daniel P. Berrange [Mon, 15 May 2017 16:47:12 +0000 (17:47 +0100)]
qemu-img: copy *key-secret opts when opening newly created files

The qemu-img dd/convert commands will create an image file and
then try to open it. Historically it has been possible to open
new files without passing any options. With encrypted files
though, the *key-secret options are mandatory, so we need to
provide those options when opening the newly created file.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170515164712.6643-5-berrange@redhat.com
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agoqemu-img: introduce --target-image-opts for 'convert' command
Daniel P. Berrange [Mon, 15 May 2017 16:47:11 +0000 (17:47 +0100)]
qemu-img: introduce --target-image-opts for 'convert' command

The '--image-opts' flag indicates whether the source filename
includes options. The target filename has to remain in the
plain filename format though, since it needs to be passed to
bdrv_create().  When using --skip-create though, it would be
possible to use image-opts syntax. This adds --target-image-opts
to indicate that the target filename includes options. Currently
this mandates use of the --skip-create flag too.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170515164712.6643-4-berrange@redhat.com
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agoqemu-img: fix --image-opts usage with dd command
Daniel P. Berrange [Mon, 15 May 2017 16:47:10 +0000 (17:47 +0100)]
qemu-img: fix --image-opts usage with dd command

The --image-opts flag can only be used to affect the parsing
of the source image. The target image has to be specified in
the traditional style regardless, since it needs to be passed
to the bdrv_create() API which does not support the new style
opts.

Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170515164712.6643-3-berrange@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agoqemu-img: add support for --object with 'dd' command
Daniel P. Berrange [Mon, 15 May 2017 16:47:09 +0000 (17:47 +0100)]
qemu-img: add support for --object with 'dd' command

The qemu-img dd command added --image-opts support, but missed
the corresponding --object support. This prevented passing
secrets (eg auth passwords) needed by certain disk images.

Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170515164712.6643-2-berrange@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agoqemu-img: Fix documentation of convert
Fam Zheng [Mon, 15 May 2017 10:35:51 +0000 (18:35 +0800)]
qemu-img: Fix documentation of convert

It got lost in commit a8d16f9ca "qemu-img: Update documentation for -U".

Reported-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Message-id: 20170515103551.31313-1-famz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agoqcow2: remove extra local_error variable
Alberto Garcia [Thu, 11 May 2017 15:03:37 +0000 (18:03 +0300)]
qcow2: remove extra local_error variable

Commit d7086422b1c1e75e320519cfe26176db6ec97a37 added a local_err
variable global to the qcow2_amend_options() function, so there's no
need to have this other one.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-id: 20170511150337.21470-1-berto@igalia.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
7 years agomirror: Drop permissions on s->target on completion
Kevin Wolf [Mon, 29 May 2017 12:08:32 +0000 (14:08 +0200)]
mirror: Drop permissions on s->target on completion

This fixes an assertion failure that was triggered by qemu-iotests 129
on some CI host, while the same test case didn't seem to fail on other
hosts.

Essentially the problem is that the blk_unref(s->target) in
mirror_exit() doesn't necessarily mean that the BlockBackend goes away
immediately. It is possible that the job completion was triggered nested
in mirror_drain(), which looks like this:

    BlockBackend *target = s->target;
    blk_ref(target);
    blk_drain(target);
    blk_unref(target);

In this case, the write permissions for s->target are retained until
after blk_drain(), which makes removing mirror_top_bs fail for the
active commit case (can't have a writable backing file in the chain
without the filter driver).

Explicitly dropping the permissions first means that the additional
reference doesn't hurt and the job can complete successfully even if
called from the nested blk_drain().

Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
7 years agoehci: fix frame timer invocation.
Gerd Hoffmann [Fri, 19 May 2017 12:04:28 +0000 (14:04 +0200)]
ehci: fix frame timer invocation.

ehci registers ehci_frame_timer as both timer and bottom half, which
turned out to be a bad idea as it can be called as bottom half then
while it is running as timer, and it isn't prepared to handle recursive
calls.

Change the timer func to just schedule the bottom half to avoid this.

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1449609
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20170519120428.25981-1-kraxel@redhat.com

7 years agousb: don't wakeup during coldplug
Gerd Hoffmann [Tue, 23 May 2017 08:46:35 +0000 (10:46 +0200)]
usb: don't wakeup during coldplug

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1452512
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20170523084635.20062-1-kraxel@redhat.com

7 years agousb-hub: set PORT_STAT_C_SUSPEND on host-initiated wake-up
Ladi Prosek [Mon, 22 May 2017 12:33:25 +0000 (14:33 +0200)]
usb-hub: set PORT_STAT_C_SUSPEND on host-initiated wake-up

PORT_STAT_C_SUSPEND should be set even on host-initiated wake-up,
i.e. on ClearPortFeature(PORT_SUSPEND). Windows is known to not
work properly otherwise.

Side note, since PORT_ENABLE looks similar and might appear to
have the same issue: According to 11.24.2.7.2.2 C_PORT_ENABLE:

  "This bit is set when the PORT_ENABLE bit changes from one to
  zero as a result of a Port Error condition (see Section 11.8.1).
  This bit is not set on any other changes to PORT_ENABLE."

Signed-off-by: Ladi Prosek <lprosek@redhat.com>
Message-id: 20170522123325.2199-1-lprosek@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoxhci: add CONFIG_USB_XHCI_NEC option
Gerd Hoffmann [Wed, 17 May 2017 10:33:13 +0000 (12:33 +0200)]
xhci: add CONFIG_USB_XHCI_NEC option

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1451189
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20170517103313.8459-2-kraxel@redhat.com

7 years agoxhci: split into multiple files
Gerd Hoffmann [Wed, 17 May 2017 10:33:12 +0000 (12:33 +0200)]
xhci: split into multiple files

Moved structs and defines to hcd-xhci.h.
Move nec controller variant to hcd-xhci-nec.c.
No functional changes.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20170517103313.8459-1-kraxel@redhat.com

7 years agousb: Simplify the parameter parsing of the legacy usb serial device
Thomas Huth [Fri, 19 May 2017 07:00:04 +0000 (09:00 +0200)]
usb: Simplify the parameter parsing of the legacy usb serial device

Coverity complains about the current code, so let's get rid of
the now unneeded while loop and simply always emit "unrecognized
serial USB option" for all unsupported options.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1495177204-16808-1-git-send-email-thuth@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agousb: Deprecate HMP commands usb_add and usb_del
Thomas Huth [Fri, 19 May 2017 06:36:43 +0000 (08:36 +0200)]
usb: Deprecate HMP commands usb_add and usb_del

The commands 'device_add' and 'device_del' should be used
nowadays instead.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-id: 1495175803-12830-1-git-send-email-thuth@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agousb: Deprecate the legacy -usbdevice option
Thomas Huth [Fri, 19 May 2017 06:35:16 +0000 (08:35 +0200)]
usb: Deprecate the legacy -usbdevice option

The '-usbdevice' option is considered as deprecated nowadays and
we might want to remove these options in a future version of QEMU.
So mark this options as deprecated in the documenation and print out
a warning if it is used to tell the user what to use instead.
While we're at it, improve also some other minor USB-related spots
in qemu-options.hx that were not up to date anymore.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1495175716-12735-1-git-send-email-thuth@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoehci: fix overflow in frame timer code
Gerd Hoffmann [Mon, 15 May 2017 10:45:43 +0000 (12:45 +0200)]
ehci: fix overflow in frame timer code

In case the frame timer doesn't run for a while due to the host being
busy skipped_uframes can become big enough that UFRAME_TIMER_NS *
skipped_uframes overflows.  Which in turn throws off all subsequent
ehci frame timer calculations.

Reported-by: 李林 <8610_28@163.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20170515104543.32044-1-kraxel@redhat.com

7 years agolinux-user: add strace support for uinfo structure of rt_sigqueueinfo() and rt_tgsigq...
Miloš Stojanović [Mon, 15 May 2017 14:59:49 +0000 (16:59 +0200)]
linux-user: add strace support for uinfo structure of rt_sigqueueinfo() and rt_tgsigqueueinfo()

This commit adds support for printing the content of the target_siginfo_t
structure in a similar way to how it is printed by the host strace. The
pointer to this structure is sent as the last argument of the
rt_sigqueueinfo() and rt_tgsigqueueinfo() system calls.
For this purpose, print_siginfo() is used and the get_target_siginfo()
function is implemented in order to get the information obtained from
the pointer into the form that print_siginfo() expects.

The get_target_siginfo() function is based on
host_to_target_siginfo_noswap() in linux-user mode, but here both
arguments are pointers to target_siginfo_t, so instead of converting
the information to siginfo_t it just extracts and copies it to a
target_siginfo_t structure.

Prior to this commit, typical strace output used to look like this:
8307 rt_sigqueueinfo(8307,50,0x00000040007ff6b0) = 0

After this commit, it looks like this:
8307 rt_sigqueueinfo(8307,50,{si_signo=50, si_code=SI_QUEUE, si_pid=8307,
si_uid=1000, si_sigval=17716762128}) = 0

Signed-off-by: Miloš Stojanović <Milos.Stojanovic@rt-rk.com>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
7 years agolinux-user: fix inconsistent spaces in print_siginfo() output
Miloš Stojanović [Mon, 15 May 2017 14:59:48 +0000 (16:59 +0200)]
linux-user: fix inconsistent spaces in print_siginfo() output

This patch improves the consistentcy of the output from print_siginfo()
by removing spaces around the equal sign of si_pid, si_uid, si_timer1,
si_timer2, si_band, si_fd, si_addr, si_status and si_sigval. This way
they match si_signo and ci_code. Host strace was used as a reference
for this chage.

Prior to this commit, typical strace output used to look like this:

Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
7 years agolinux-user: add rt_tgsigqueueinfo() strace
Miloš Stojanović [Mon, 15 May 2017 14:59:47 +0000 (16:59 +0200)]
linux-user: add rt_tgsigqueueinfo() strace

This commit improves strace support for syscall rt_tgsigqueueinfo().

Prior to this commit, typical strace output used to look like this:
7775 rt_tgsigqueueinfo(7775,7775,50,1996483164,0,0) = 0

After this commit, it looks like this:
7775 rt_tgsigqueueinfo(7775,7775,50,0x76ffea5c) = 0

Signed-off-by: Miloš Stojanović <Milos.Stojanovic@rt-rk.com>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
7 years agolinux-user: add support for rt_tgsigqueueinfo() system call
Miloš Stojanović [Mon, 15 May 2017 14:59:46 +0000 (16:59 +0200)]
linux-user: add support for rt_tgsigqueueinfo() system call

Add a new system call: rt_tgsigqueueinfo().

This system call is similar to rt_sigqueueinfo(), but instead of
sending the signal and data to the whole thread group with the ID
equal to the argument tgid, it sends it to a single thread within
that thread group. The ID of the thread is specified by the tid
argument.

The implementation is based on the rt_sigqueueinfo() in linux-user
mode, where the tid is added as the second argument and the
previous second and third argument become arguments three and four,
respectively.

Signed-off-by: Miloš Stojanović <Milos.Stojanovic@rt-rk.com>
Conflicts:
linux-user/syscall.c
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
7 years agolinux-user: fix argument type declaration of rt_sigqueinfo() syscall
Miloš Stojanović [Mon, 15 May 2017 14:59:45 +0000 (16:59 +0200)]
linux-user: fix argument type declaration of rt_sigqueinfo() syscall

Change the type of the first argument of rt_sigqueinfo() from int to pid_t
in the syscall declaration to match specifications of the system call.

Proper spacing is added to satisfy checkpatch.pl.

Signed-off-by: Miloš Stojanović <Milos.Stojanovic@rt-rk.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>