bsd-user: Implement chown, fchown, lchown and fchownat

Signed-off-by: Stacey Son <sson@FreeBSD.org>
Signed-off-by: Warner Losh <imp@bsdimp.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

bsd-user: Implement freebsd11_mknod, freebsd11_mknodat and mknodat

These implement both the old-pre INO64 mknod variations, as well as the
now current INO64 variant. Make direct syscall calls for these older
syscalls to avloid too many dependencies.

Signed-off-by: Stacey Son <sson@FreeBSD.org>
Signed-off-by: Michal Meloun <mmel@FreeBSD.org>
Signed-off-by: Warner Losh <imp@bsdimp.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

bsd-user: implement chmod, fchmod, lchmod and fchmodat

Signed-off-by: Stacey Son <sson@FreeBSD.org>
Signed-off-by: Warner Losh <imp@bsdimp.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

bsd-user: Implement symlink, symlinkat, readlink and readlinkat

Signed-off-by: Stacey Son <sson@FreeBSD.org>
Signed-off-by: Jung-uk Kim <jkim@FreeBSD.org>
Signed-off-by: Warner Losh <imp@bsdimp.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

bsd-user: Implement mount, umount and nmount

Signed-off-by: Stacey Son <sson@FreeBSD.org>
Signed-off-by: Jung-uk Kim <jkim@FreeBSD.org>
Signed-off-by: Warner Losh <imp@bsdimp.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

Merge tag 'pull-aspeed-20220630' of https://github.com/legoater/qemu into staging

aspeed queue:

* m25p80 improvements (Iris)
* Code cleanup in preparation of multi SoC machine (Peter)
* New MAX31785 model (Mahesh)
* New Qualcomm machines (Jae and Graeme)
* Core I2C slave mode (Klaus)
* Aspeed I2C slave mode for old and new register interface (Peter and Klaus)
* New Aspeed PECI model (Peter)
* Various small fixes

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEoPZlSPBIlev+awtgUaNDx8/77KEFAmK9UfQACgkQUaNDx8/7
# 7KFYWhAAtlx3aaEacrn/ONNHjk6G9Fxku56gAbaIiuiaIWNOj3/T2frPsnmbO8x8
# EKrgUYB8i8PFve/fJYA5vZUzIddPTaHkULZ12JQoGVg0L9hDBbizslN5lJWRXoSv
# 9r3DF9nahzLKRNvzoBfuKjHDQ2cwHoFgYmKmlYpDcgfmBcl16uzZy8jvxg/Tghur
# umH4IJMjeDNz/kLfINoO/m+kuFPVXmbTJNwl8uK5MUVDTgVSqharywWlUizugBVH
# StLE+GmBPylTuYXyiOzLTkoGJeeHp3sQ1DmyI4DD83odjnfxa0BGMGDVhD35exXi
# 9tLY9FgQ4smATuyN0UGAKZTBmzpI+ov0HMzvH4lUMR8i8daBuEet3RVr/DqkOP4h
# LEVTRWTaTJip24ohgw4K/b86pI9nTJWVPGV56eZGYmnqufnvf/upNU65/nCsF/xD
# i1TdS+zJWxhjgGEepg9cTmxxUlA4jVNNbl6dvAgS5Jr6Igrd1BlCSXjmyhO3NRPZ
# bgOuvCb3RyxAY4+/9wphx2/t5X2VIU6R8EAjnh+7nIgBhOQU5SZ6uefFVYZq8xx+
# IYEDHj3saiRa4FHmyOgeRxRaQj/Vvs83PPti2rPmJuieqiClJmbE+XfTIamoxVIv
# 5USlKmMRRVI69MjsjwFi/gOaV/N1EUgcFoYbnvwZ+Md3fg5+70M=
# =oUKu
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 30 Jun 2022 01:04:12 PM +0530
# gpg:                using RSA key A0F66548F04895EBFE6B0B6051A343C7CFFBECA1
# gpg: Good signature from "Cédric Le Goater <clg@kaod.org>" [undefined]
# 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: A0F6 6548 F048 95EB FE6B  0B60 51A3 43C7 CFFB ECA1

* tag 'pull-aspeed-20220630' of https://github.com/legoater/qemu: (27 commits)
  hw/misc/aspeed: Add PECI controller
  hw/i2c/aspeed: Add new-registers DMA slave mode RX support
  hw/i2c/aspeed: add slave device in old register mode
  hw/i2c: add asynchronous send
  hw/i2c: support multiple masters
  hw/i2c/aspeed: Fix MASTER_EN missing error message
  hw/i2c/aspeed: Fix DMA len write-enable bit handling
  hw/i2c/aspeed: Fix R_I2CD_FUN_CTRL reference
  hw/arm/aspeed: firework: add I2C MUXes for VR channels
  hw/arm/aspeed: firework: Add Thermal Diodes
  hw/arm/aspeed: Add MAX31785 Fan controllers
  hw/sensor: add Maxim MAX31785 device
  hw/i2c: pmbus: Page #255 is valid page for read requests.
  hw/arm/aspeed: add Qualcomm Firework BMC machine
  hw/arm/aspeed: add support for the Qualcomm DC-SCM v1 board
  aspeed: Remove use of qemu_get_cpu
  aspeed: Map unimplemented devices in SoC memory
  aspeed: Remove usage of sysbus_mmio_map
  aspeed: Add memory property to Aspeed SoC
  aspeed: Set CPU memory property explicitly
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

hw/misc/aspeed: Add PECI controller

This introduces a really basic PECI controller that responses to
commands by always setting the response code to success and then raising
an interrupt to indicate the command is done. This helps avoid getting
hit with constant errors if the driver continuously attempts to send a
command and keeps timing out.

The AST2400 and AST2500 only included registers up to 0x5C, not 0xFC.
They supported PECI 1.1, 2.0, and 3.0. The AST2600 and AST1030 support
PECI 4.0, which includes more read/write buffer registers from 0x80 to
0xFC to support 64-byte mode.

This patch doesn't attempt to handle that, or to create a different
version of the controller for the different generations, since it's only
implementing functionality that is common to all generations.

The basic sequence of events is that the firmware will read and write to
various registers and then trigger a command by setting the FIRE bit in
the command register (similar to the I2C controller).

Then the firmware waits for an interrupt from the PECI controller,
expecting the interrupt status register to be filled in with info on
what happened. If the command was transmitted and received successfully,
then response codes from the host CPU will be found in the data buffer
registers.

Signed-off-by: Peter Delevoryas <pdel@fb.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220630045133.32251-12-me@pjd.dev>
[ clg: s/sysbus_mmio_map/aspeed_mmio_map/ ]
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/i2c/aspeed: Add new-registers DMA slave mode RX support

This commit adds support for DMA RX in slave mode while using the new
register set in the AST2600 and AST1030. This patch also pretty much
assumes packet mode is enabled, I'm not sure if this will work in DMA
step mode.

This is particularly useful for testing IPMB exchanges between Zephyr
and external devices, which requires multi-master I2C support and DMA in
the new register mode, because the Zephyr drivers from Aspeed use DMA in
the new mode by default. The Zephyr drivers are also using packet mode.

The typical sequence of events for receiving data in DMA slave + packet
mode is that the Zephyr firmware will configure the slave address
register with an address to receive on and configure the bus's function
control register to enable master mode and slave mode simultaneously at
startup, before any transfers are initiated.

RX DMA is enabled in the slave mode command register, and the slave RX
DMA buffer address and slave RX DMA buffer length are set. TX DMA is not
covered in this patch.

When the Aspeed I2C controller receives data from some other I2C master,
it will reset the I2CS_DMA_LEN RX_LEN value to zero, then buffer
incoming data in the RX DMA buffer while incrementing the I2CC_DMA_ADDR
address counter and decrementing the I2CC_DMA_LEN counter. It will also
update the I2CS_DMA_LEN RX_LEN value along the way.

Once all the data has been received, the bus controller will raise an
interrupt indicating a packet command was completed, the slave address
matched, a normal stop condition was seen, and the transfer was an RX
operation.

If the master sent a NACK instead of a normal stop condition, or the
transfer timed out, then a slightly different set of interrupt status
values would be set. Those conditions are not handled in this commit.

The Zephyr firmware then collects data from the RX DMA buffer and clears
the status register by writing the PKT_MODE_EN bit to the status
register. In packet mode, clearing the packet mode interrupt enable bit
also clears most of the other interrupt bits automatically (except for a
few bits above it).

Note: if the master transmit or receive functions were in use
simultaneously with the slave mode receive functionality, then the
master mode functions may have raised the interrupt line for the bus
before the DMA slave transfer is complete. It's important to have the
slave's interrupt status register clear throughout the receive
operation, and if the slave attempts to raise the interrupt before the
master interrupt status is cleared, then it needs to re-raise the
interrupt once the master interrupt status is cleared. (And vice-versa).
That's why in this commit, when the master interrupt status is cleared
and the interrupt line is lowered, we call the slave interrupt _raise_
function, to see if the interrupt was pending. (And again, vice-versa).

Signed-off-by: Peter Delevoryas <pdel@fb.com>
Message-Id: <20220630045133.32251-8-me@pjd.dev>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/i2c/aspeed: add slave device in old register mode

Add slave mode functionality for the Aspeed I2C controller in old
register mode. This is implemented by realizing an I2C slave device
owned by the I2C controller and attached to its own bus.

The I2C slave device only implements asynchronous sends on the bus, so
slaves not supporting that will not be able to communicate with it.

Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
[ clg: checkpatch fixes ]
Message-Id: <20220601210831.67259-6-its@irrelevant.dk>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220630045133.32251-7-me@pjd.dev>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/i2c: add asynchronous send

Add an asynchronous version of i2c_send() that requires the slave to
explicitly acknowledge on the bus with i2c_ack().

The current master must use the new i2c_start_send_async() to indicate
that it wants to do an asynchronous transfer. This allows the i2c core
to check if the target slave supports this or not. This approach relies
on adding a new enum i2c_event member, which is why a bunch of other
devices needs changes in their event handling switches.

Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Message-Id: <20220601210831.67259-5-its@irrelevant.dk>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220630045133.32251-6-me@pjd.dev>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/i2c: support multiple masters

Allow slaves to master the bus by registering a bottom halve. If the bus
is busy, the bottom half is queued up. When a slave has succesfully
mastered the bus, the bottom half is scheduled.

Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
[ clg : - fixed typos in commit log ]
Message-Id: <20220601210831.67259-4-its@irrelevant.dk>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220630045133.32251-5-me@pjd.dev>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/i2c/aspeed: Fix MASTER_EN missing error message

aspeed_i2c_bus_is_master is checking if master mode is enabled in the I2C
bus controller's function-control register, not that slave mode is enabled
or something.  The error here is that the guest is trying to trigger an I2C
master mode command while master mode is not enabled.

Fixes: ba2cccd64e90f342 ("aspeed: i2c: Add new mode support")
Signed-off-by: Peter Delevoryas <pdel@fb.com>
Message-Id: <20220630045133.32251-4-me@pjd.dev>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/i2c/aspeed: Fix DMA len write-enable bit handling

I noticed i2c rx transfers were getting shortened to "1" on Zephyr. It
seems to be because the Zephyr i2c driver sets the RX DMA len with the
RX field write-enable bit set (bit 31) to avoid a read-modify-write. [1]

/* 0x1C : I2CM Master DMA Transfer Length Register   */

I think we should be checking the write-enable bits on the incoming
value, not checking the register array. I'm not sure we're even writing
the write-enable bits to the register array, actually.

[1] https://github.com/AspeedTech-BMC/zephyr/blob/db3dbcc9c52e67a47180890ac938ed380b33f91c/drivers/i2c/i2c_aspeed.c#L145-L148

Fixes: ba2cccd64e90f34 ("aspeed: i2c: Add new mode support")
Signed-off-by: Peter Delevoryas <pdel@fb.com>
Message-Id: <20220630045133.32251-3-me@pjd.dev>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/i2c/aspeed: Fix R_I2CD_FUN_CTRL reference

Very minor, doesn't effect functionality, but this is supposed to be
R_I2CC_FUN_CTRL (new-mode, not old-mode).

Fixes: ba2cccd64e9 ("aspeed: i2c: Add new mode support")
Signed-off-by: Peter Delevoryas <pdel@fb.com>
Message-Id: <20220630045133.32251-2-me@pjd.dev>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/arm/aspeed: firework: add I2C MUXes for VR channels

Add 2-level cascaded I2C MUXes for SOC VR channels into the Firework
machine.

Signed-off-by: Jae Hyun Yoo <quic_jaehyoo@quicinc.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220627154703.148943-8-quic_jaehyoo@quicinc.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/arm/aspeed: firework: Add Thermal Diodes

Add Thermal Diodes for Firework machine.

Signed-off-by: Maheswara Kurapati <quic_mkurapat@quicinc.com>
Signed-off-by: Jae Hyun Yoo <quic_jaehyoo@quicinc.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220627154703.148943-7-quic_jaehyoo@quicinc.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/arm/aspeed: Add MAX31785 Fan controllers

Add MAX31785 fan controllers in machines so that the Linux driver
populates the sysfs interface.

Firework has two MAX31785 Fan controllers at 0x52, and 0x54 on bus 9.
Witherspoon has one at 0x52 on bus 3.
Rainier has one at 0x52 on bus 7.

Signed-off-by: Maheswara Kurapati <quic_mkurapat@quicinc.com>
Signed-off-by: Jae Hyun Yoo <quic_jaehyoo@quicinc.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220627154703.148943-6-quic_jaehyoo@quicinc.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/sensor: add Maxim MAX31785 device

MAX31785 is a PMBus compliant 6-Channel fan controller. It supports 6 fan
channels, 11 temperature sensors, and 6-Channel ADC to measure the remote
voltages. Datasheet can be found here:
https://datasheets.maximintegrated.com/en/ds/MAX31785.pdf

This initial version of the driver has skeleton and support for the
fan channels. Requests for temperature sensors, and ADC Channels the
are serviced with the default values as per the datasheet.  No additional
instrumentation is done. NV Log feature is not supported.

Signed-off-by: Maheswara Kurapati <quic_mkurapat@quicinc.com>
Signed-off-by: Jae Hyun Yoo <quic_jaehyoo@quicinc.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220627154703.148943-5-quic_jaehyoo@quicinc.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/i2c: pmbus: Page #255 is valid page for read requests.

Current implementation of the pmbus core driver treats the read request
for page 255 as invalid request and sets the invalid command bit (bit 7)
in the STATUS_CML register. As per the PMBus specification it is a valid
request.

Refer to the PMBus specification, revision 1.3.1, section 11.10 PAGE,
on the page 58:
  "Setting the PAGE to FFh means that all subsequent comands are to be
   applied to all outputs.

   Some commands, such as READ_TEMPERATURE, may use a common sensor but
   be available on all pages of a device. Such implementations are the
   decision of each device manufacturer or are specified in a PMBus
   Application Profile. Consult the manufacturer's documents or the
   Application Profile Specification as needed."

For e.g.,
The VOUT_MODE is a valid command for page 255 for maxim 31785 device.
refer to Table 1. PMBus Command Codes on page 14 in the datasheet.
https://datasheets.maximintegrated.com/en/ds/MAX31785.pdf

Fixes: 38870253f1d1 ("hw/i2c: pmbus: fix error returns and guard against out of range accesses")
Signed-off-by: Maheswara Kurapati <quic_mkurapat@quicinc.com>
Signed-off-by: Jae Hyun Yoo <quic_jaehyoo@quicinc.com>
Reviewed-by: Titus Rwantare <titusr@google.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220627154703.148943-4-quic_jaehyoo@quicinc.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/arm/aspeed: add Qualcomm Firework BMC machine

Add base for Qualcomm Firework BMC machine.

Signed-off-by: Graeme Gregory <quic_ggregory@quicinc.com>
Signed-off-by: Jae Hyun Yoo <quic_jaehyoo@quicinc.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220627154703.148943-3-quic_jaehyoo@quicinc.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw/arm/aspeed: add support for the Qualcomm DC-SCM v1 board

Add qcom-dc-scm-v1 board support.

Signed-off-by: Jae Hyun Yoo <quic_jaehyoo@quicinc.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220627154703.148943-2-quic_jaehyoo@quicinc.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed: Remove use of qemu_get_cpu

Signed-off-by: Peter Delevoryas <pdel@fb.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220624003701.1363500-6-pdel@fb.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed: Map unimplemented devices in SoC memory

Signed-off-by: Peter Delevoryas <pdel@fb.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220624003701.1363500-5-pdel@fb.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed: Remove usage of sysbus_mmio_map

sysbus_mmio_map maps devices into "get_system_memory()".

With the new SoC memory attribute, we want to make sure that each device is
mapped into the SoC memory.

In single SoC machines, the SoC memory is the same as "get_system_memory()",
but in multi SoC machines it will be different.

Signed-off-by: Peter Delevoryas <pdel@fb.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220624003701.1363500-4-pdel@fb.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed: Add memory property to Aspeed SoC

Multi-SoC machines can use this property to specify a memory container
for each SoC. Single SoC machines will just specify get_system_memory().

Signed-off-by: Peter Delevoryas <pdel@fb.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220624003701.1363500-3-pdel@fb.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed: Set CPU memory property explicitly

Signed-off-by: Peter Delevoryas <pdel@fb.com>
Message-Id: <20220624003701.1363500-2-pdel@fb.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed/smc: Fix potential overflow

Coverity warns that "ssi_transfer(s->spi, 0U) << 8 * i" might overflow
because the expression is evaluated using 32-bit arithmetic and then
used in a context expecting a uint64_t.

Fixes: Coverity CID 1487244
Message-Id: <20220628165512.1133590-1-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed/hace: Accumulative mode supported

While the HMAC mode is not modelled, the accumulative mode is.

Accumulative mode is enabled by setting one of the bits in the HMAC
engine command mode part of the register, so fix the unimplemented check
to only look at the upper of the two bits.

Fixes: 5cd7d8564a8b ("aspeed/hace: Support AST2600 HACE")
Signed-off-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220627100816.125956-1-joel@jms.id.au>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed/i2c: Change trace event for NORMAL_STOP states

Using a 'stop' string seems more appropriate than 'normal'.

Reviewed-by: Peter Delevoryas <pdel@fb.com>
Message-Id: <20220628154740.1117349-3-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed/scu: Add trace events for read ops

Reviewed-by: Peter Delevoryas <pdel@fb.com>
Message-Id: <20220628154740.1117349-2-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

aspeed: Set the dram container at the SoC level

Currently, the Aspeed machines allocate a ram container region in
which the machine ram region is mapped. See commit ad1a9782186d
("aspeed: add a RAM memory region container"). An extra region is
mapped after ram in the ram container to catch invalid access done by
FW. That's how FW determines the size of ram. See commit ebe31c0a8ef7
("aspeed: add a max_ram_size property to the memory controller").

Let's move all the logic under the SoC where it should be. It will
also ease the work on multi SoC support.

Reviewed-by: Peter Delevoryas <pdel@fb.com>
Message-Id: <20220623202123.3972977-1-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw: m25p80: add tests for write protect (WP# and SRWD bit)

Signed-off-by: Iris Chen <irischenlj@fb.com>
Message-Id: <20220624183016.2125264-1-irischenlj@fb.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

hw: m25p80: add WP# pin and SRWD bit for write protection

Signed-off-by: Iris Chen <irischenlj@gmail.com>
Reviewed-by: Francisco Iglesias <frasse.iglesias@gmail.com>
Message-Id: <20220621202427.2680413-1-irischenlj@fb.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

Merge tag 'trivial-branch-for-7.1-pull-request' of https://gitlab.com/laurent_vivier/qemu into staging

trivial patches pull request 20220629

# -----BEGIN PGP SIGNATURE-----
#
# iQJGBAABCAAwFiEEzS913cjjpNwuT1Fz8ww4vT8vvjwFAmK8FmsSHGxhdXJlbnRA
# dml2aWVyLmV1AAoJEPMMOL0/L7482EkP/19M/AAUkGqIdU9Dj7H46r+LEKtrT7Xu
# jNRDDrkhVQvx42mklSB+fO/ptMKUDgxvLs4mnuZFxM7SrTOb4h5jfZzyYjk73ENQ
# YZ/TLxRtxAfRCcGwso7NGyk85mwt+sBFKZXfW6qsfc9AjDphLUOblfSieeFegz69
# BUtzbMOPSMR7e54y6azJX3gCkxLytSXYgk4otSLTrL233sT7pnwPRdxKGzCTA5vs
# fRxKb4p/R05lWepcjrL2d2lB1TabsV0kqmNkHDvubVWlgyoK3Vt/1dzD1UP7CrvF
# WghlZWmxCHrmLlBb+VSDUa22kpfv5fi/feauuug+dya+s1Mlq8HZTL8VtjUJHwLL
# 92xRPeP/RfEJdoQDuMKXP9DWAAYM03HGgR37cE5NMDCyHG0XRKOJ+i2P7DQLVDjW
# QyWX6bX1WV6FovdwwMnZR9OclvKtsZnb1jlfj+G2DdKXpLliDH6DkFm8mPQTM1L7
# w53iMtK88erEc+NP6+fPbbZmySvDVUcLmcTiBceZK6Vjo4oTGNrAWP+VgjBTJaz+
# 71ulkJ6vo39ZnEQOUlWrL/yW+8sQNaeO1tO67HZZ8dgTvAnPwyvKq88jSMzGCNpz
# Wpcf4yVAEvU+fP3KkEaqQqmQeK/Vc+H6044O00tcLVICkpCdN/FwRjgfZanX9CIJ
# xQjxW5mkb1Z3
# =fgtJ
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 29 Jun 2022 02:37:55 PM +0530
# gpg:                using RSA key CD2F75DDC8E3A4DC2E4F5173F30C38BD3F2FBE3C
# gpg:                issuer "laurent@vivier.eu"
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>" [undefined]
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>" [undefined]
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>" [undefined]
# 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: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* tag 'trivial-branch-for-7.1-pull-request' of https://gitlab.com/laurent_vivier/qemu:
  hw/i386/xen/xen-hvm: Inline xen_piix_pci_write_config_client() and remove it
  hw/i386/xen/xen-hvm: Allow for stubbing xen_set_pci_link_route()
  hw/ide/atapi.c: Correct typos (CD-CDROM -> CD-ROM)
  common-user: Only compile the common user code if have_user is set
  hw/pci-host/i440fx: Remove unused parameter from i440fx_init()
  MAINTAINERS: Add softmmu/runstate.c to "Main loop"
  trivial typos: namesapce
  Trivial: 3 char repeat typos
  util: Return void on iova_tree_remove
  qom/object: Remove circular include dependency
  vga: avoid crash if no default vga card

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge tag 'pull-block-2022-06-14-v2' of https://gitlab.com/vsementsov/qemu into staging

Block jobs & NBD patches

v2: - add arguments to QEMUMachine constructor in test, to make it work
      on arm in gitlab pipeline
    - use bdrv_inc_in_flight() / bdrv_dec_in_flight() instead of direct
      manipulation with bs->in_flight

- add new options for copy-before-write filter
- new trace points for NBD
- prefer unsigned type for some 'in_flight' fields

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCgAdFiEEi5wmzbL9FHyIDoahVh8kwfGfefsFAmK8BqkACgkQVh8kwfGf
# efuiMw/+P9FFLfGFSjVP+LYeT0Ah6kN1ypCMQzIk3Qq/J6qgMZhtRqpQoZOfIFQL
# U9fGmEtQZ7gvEMD/nJToL6uOYlnQfPxDcA6GrRwWWE3rcFiPK4J0q2LlnPLn4QaU
# W/qag5l/QnZfLlj/iV6neWOEvqdnvY1L8IS+T8xV6N0iBYlwgMC/6FGshQwehzcV
# T5F1qPGB0vjFDjf92LFPEsvsFFHjHIVPwOyJMvF64QtSk57utikq/la9PI/yA9AH
# Ll4mNQuZKx6rSI5wE6b21jc8iOUvaoHdPSEDQZfNILSdgGdiKvFwE51y+baGnIAD
# TpjxfG59q0jyGxMjQVxMRSFaxAC4+Mqi82diSPv4xbiUdsE4byJH0oENn4z7+wAv
# EvjuU9yx4FfHHltoUNwfn2pv00o/ELdZIoBNmW36rPxSGZMvfVfRtuBL7XWNUFbW
# Fr4BwsjC4KfIxb16QTBGhXVv6grxdlwoU9N23npdi0YpW1ftZzXGDa85+gINQ807
# eK/gP/OtYPwIql0bgmLiuaRNzC9psmQOO6vQbdvd/e4BEVWkxiez37+e+zFMStmL
# OAL8rS6jckUoxVZjCYFEWg97XOobLUIhQxt9Fwh2omMDGKTwv861ghUAivxSWs93
# IRNxfwqNPxnpDDXjnK1ayZgU08IL98AUYVKcPN1y8JvEhB4Hr1k=
# =ndKk
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 29 Jun 2022 01:30:41 PM +0530
# gpg:                using RSA key 8B9C26CDB2FD147C880E86A1561F24C1F19F79FB
# gpg: Good signature from "Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>" [unknown]
# gpg:                 aka "Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.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: 8B9C 26CD B2FD 147C 880E  86A1 561F 24C1 F19F 79FB

* tag 'pull-block-2022-06-14-v2' of https://gitlab.com/vsementsov/qemu:
  block: use 'unsigned' for in_flight field on driver state
  nbd: trace long NBD operations
  iotests: copy-before-write: add cases for cbw-timeout option
  block/copy-before-write: implement cbw-timeout option
  block/block-copy: block_copy(): add timeout_ns parameter
  util: add qemu-co-timeout
  iotests: add copy-before-write: on-cbw-error tests
  block/copy-before-write: add on-cbw-error open parameter
  block/copy-before-write: refactor option parsing

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

block: use 'unsigned' for in_flight field on driver state

This patch makes in_flight field 'unsigned' for BDRVNBDState and
MirrorBlockJob. This matches the definition of this field on BDS
and is generically correct - we should never get negative value here.

Signed-off-by: Denis V. Lunev <den@openvz.org>
CC: John Snow <jsnow@redhat.com>
CC: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
CC: Kevin Wolf <kwolf@redhat.com>
CC: Hanna Reitz <hreitz@redhat.com>
CC: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

nbd: trace long NBD operations

At the moment there are 2 sources of lengthy operations if configured:
* open connection, which could retry inside and
* reconnect of already opened connection
These operations could be quite lengthy and cumbersome to catch thus
it would be quite natural to add trace points for them.

This patch is based on the original downstream work made by Vladimir.

Signed-off-by: Denis V. Lunev <den@openvz.org>
CC: Eric Blake <eblake@redhat.com>
CC: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
CC: Kevin Wolf <kwolf@redhat.com>
CC: Hanna Reitz <hreitz@redhat.com>
CC: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

iotests: copy-before-write: add cases for cbw-timeout option

Add two simple test-cases: timeout failure with
break-snapshot-on-cbw-error behavior and similar with
break-guest-write-on-cbw-error behavior.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

block/copy-before-write: implement cbw-timeout option

In some scenarios, when copy-before-write operations lasts too long
time, it's better to cancel it.

Most useful would be to use the new option together with
on-cbw-error=break-snapshot: this way if cbw operation takes too long
time we'll just cancel backup process but do not disturb the guest too
much.

Note the tricky point of realization: we keep additional point in
bs->in_flight during block_copy operation even if it's timed-out.
Background "cancelled" block_copy operations will finish at some point
and will want to access state. We should care to not free the state in
.bdrv_close() earlier.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
  [vsementsov: use bdrv_inc_in_flight()/bdrv_dec_in_flight() instead of
   direct manipulation on bs->in_flight]
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

block/block-copy: block_copy(): add timeout_ns parameter

Add possibility to limit block_copy() call in time. To be used in the
next commit.

As timed-out block_copy() call will continue in background anyway (we
can't immediately cancel IO operation), it's important also give user a
possibility to pass a callback, to do some additional actions on
block-copy call finish.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

util: add qemu-co-timeout

Add new API, to make a time limited call of the coroutine.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

iotests: add copy-before-write: on-cbw-error tests

Add tests for new option of copy-before-write filter: on-cbw-error.

Note that we use QEMUMachine instead of VM class, because in further
commit we'll want to use throttling which doesn't work with -accel
qtest used by VM.

We also touch pylintrc to not break iotest 297.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
  [vsementsov: add arguments to QEMUMachine constructor]
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

hw/i386/xen/xen-hvm: Inline xen_piix_pci_write_config_client() and remove it

xen_piix_pci_write_config_client() is implemented in the xen sub tree and
uses PIIX constants internally, thus creating a direct dependency on
PIIX. Now that xen_set_pci_link_route() is stubbable, the logic of
xen_piix_pci_write_config_client() can be moved to PIIX which resolves
the dependency.

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Paul Durrant <paul@xen.org>
Message-Id: <20220626094656.15673-3-shentey@gmail.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/i386/xen/xen-hvm: Allow for stubbing xen_set_pci_link_route()

The only user of xen_set_pci_link_route() is
xen_piix_pci_write_config_client() which implements PIIX-specific logic in
the xen namespace. This makes xen-hvm depend on PIIX which could be
avoided if xen_piix_pci_write_config_client() was implemented in PIIX. In
order to do this, xen_set_pci_link_route() needs to be stubbable which
this patch addresses.

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Paul Durrant <paul@xen.org>
Message-Id: <20220626094656.15673-2-shentey@gmail.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/ide/atapi.c: Correct typos (CD-CDROM -> CD-ROM)

Signed-off-by: Lev Kujawski <lkujaw@member.fsf.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220528204702.167912-1-lkujaw@member.fsf.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

common-user: Only compile the common user code if have_user is set

There is no need to waste cycles here if we only compile the system
binaries or tools. Additionally, this change is even a hard requirement
for building the tools on systems that do not have an entry in the
common-user/host/ folder (since common-user/meson.build is trying
to add such a path via the include_directories() command).

Reported-by: Michael Tokarev <mjt@tls.msk.ru>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Zhang Chen <chen.zhang@intel.com>
Message-Id: <20220622140328.383961-1-thuth@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/pci-host/i440fx: Remove unused parameter from i440fx_init()

pi440fx_state is an out-parameter which is never read by the caller.

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20220612192800.40813-1-shentey@gmail.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

MAINTAINERS: Add softmmu/runstate.c to "Main loop"

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220615122338.340426-1-armbru@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

trivial typos: namesapce

'namespace' is misspelled in a bunch of places.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Message-Id: <20220614104045.85728-3-dgilbert@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

Trivial: 3 char repeat typos

Inspired by Julia Lawall's fixing of Linux
kernel comments, I looked at qemu, although I did it manually.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Message-Id: <20220614104045.85728-2-dgilbert@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

util: Return void on iova_tree_remove

It always returns IOVA_OK so nobody uses it.

Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Message-Id: <20220427154931.3166388-1-eperezma@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

qom/object: Remove circular include dependency

"qom/object.h" doesn't need to include itself.

Fixes: db1015e92e04 ("Move QOM typedefs and add missing includes")
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20220509084659.52076-1-philippe.mathieu.daude@gmail.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

vga: avoid crash if no default vga card

QEMU in some arch will crash when executing -vga help command, because
there is no default vga model.  Add check to this case and avoid crash.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/978

Signed-off-by: Guo Zhi <qtxuning1999@sjtu.edu.cn>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220503091724.970009-1-qtxuning1999@sjtu.edu.cn>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

block/copy-before-write: add on-cbw-error open parameter

Currently, behavior on copy-before-write operation failure is simple:
report error to the guest.

Let's implement alternative behavior: break the whole copy-before-write
process (and corresponding backup job or NBD client) but keep guest
working. It's needed if we consider guest stability as more important.

The realisation is simple: on copy-before-write failure we set
s->snapshot_ret and continue guest operations. s->snapshot_ret being
set will lead to all further snapshot API requests. Note that all
in-flight snapshot-API requests may still success: we do wait for them
on BREAK_SNAPSHOT-failure path in cbw_do_copy_before_write().

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

block/copy-before-write: refactor option parsing

We are going to add one more option of enum type. Let's refactor option
parsing so that we can simply work with BlockdevOptionsCbw object.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

Merge tag 'for_upstream' of git://git./virt/kvm/mst/qemu into staging

virtio: fixes

fixes all over the place

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# -----BEGIN PGP SIGNATURE-----
#
# iQFDBAABCAAtFiEEXQn9CHHI+FuUyooNKB8NuNKNVGkFAmK6NQoPHG1zdEByZWRo
# YXQuY29tAAoJECgfDbjSjVRp/sQIAJGiYliUHElJapM/4KSsXKWCFtk9B8wJuUie
# yeMKOdD6QSk9tk/HkYSCnMB7G6Fe+MtoE+sPm/6l5nOFqvqVVJw9vOKteWHSpQ0E
# 9CgbR7s7K1MoLG9J613iB2OtAfhPrWEvSOJ6mvTAxGgxhCQw6UzC88cYfBHJ/efn
# GAhVlriSfSCRANmivjY+g4h4JFWWSMTH6m9u4wKBJF8GRkNgN+C50Z+bp8aE7wRT
# KiMoaaYUDOjxzD+8nGYggg/t+UIM7jG2t8M5BMbC0NMP+ovVZeesWK6ZOzoda2tI
# ZONV0dLikLCicyOvfMH6YDqzGtchCDmS0hpfuorhlzsntm42RBM=
# =T+gr
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 28 Jun 2022 04:24:02 AM +0530
# gpg:                using RSA key 5D09FD0871C8F85B94CA8A0D281F0DB8D28D5469
# gpg:                issuer "mst@redhat.com"
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>" [undefined]
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>" [undefined]
# 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: 0270 606B 6F3C DF3D 0B17  0970 C350 3912 AFBE 8E67
#      Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA  8A0D 281F 0DB8 D28D 5469

* tag 'for_upstream' of git://git.kernel.org/pub/scm/virt/kvm/mst/qemu:
  include/hw/virtio: document vhost_ack_features
  include/hw/virtio: document vhost_get_features
  contrib/vhost-user-blk: fix 32 bit build and enable
  MAINTAINERS: Collect memory device files in "Memory devices"
  libvhost-user: Fix VHOST_USER_ADD_MEM_REG reply
  libvhost-user: Fix VHOST_USER_GET_MAX_MEM_SLOTS reply
  docs/vhost-user: Fix mismerge
  virtio-iommu: Fix migration regression
  vhost: setup error eventfd and dump errors
  vhost: add method vhost_set_vring_err
  msi: fix MSI vector limit check in msi_set_mask()
  virtio-iommu: Fix the partial copy of probe request

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge tag 'pull-semi-20220628' of https://gitlab.com/rth7680/qemu into staging

Semihosting syscall reorg:
  * Split out semihosting/syscalls.c with common implementations.
  * Reorg arm-compat-semi.c to use syscalls.c.
  * Minor prep cleanups to m68k, mips, nios2.

# -----BEGIN PGP SIGNATURE-----
#
# iQFRBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAmK6iSodHHJpY2hhcmQu
# aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV8SEwgAmmowW2oeFA9uCrwz
# gUJo17AJ+RmRF/zXHyu5CPswylvfwH0zJXAm5BV7P/pVdyaL36b8YcgSEf+EWLsf
# rLFHxCshTYEnZSk6yFtWk5bn5azfevHm9/ObPeS9XGL4seQqGy7C/FReoTQ7/zI0
# W3zUDd3bWah3fXw8XYgSzh/RCrC5E2gFFc1G1g+6SIVZ7pbgkre2rRk5WMmylCLd
# jf9pmyswrheaKumCoBxU/S4XDgxVpaf3khiIqdbo8A20MDGnK/SZUWsBwJLK3QB8
# SKKv8o1ovbnl/HykABaszCIkO/LIu6SX3LoK7pF2CujkgSuwEN3WW0DOml6+b3fU
# J7YeZg==
# =sTbM
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 28 Jun 2022 10:22:58 AM +0530
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [ultimate]

* tag 'pull-semi-20220628' of https://gitlab.com/rth7680/qemu: (60 commits)
  target/nios2: Move nios2-semi.c to nios2_softmmu_ss
  target/nios2: Eliminate nios2_semi_is_lseek
  target/mips: Drop pread and pwrite syscalls from semihosting
  target/mips: Add UHI errno values
  target/mips: Use an exception for semihosting
  target/m68k: Make semihosting system only
  target/m68k: Eliminate m68k_semi_is_fseek
  semihosting: Create semihost_sys_poll_one
  semihosting: Remove qemu_semihosting_console_outs
  semihosting: Use console_out_gf for SYS_WRITE0
  semihosting: Remove qemu_semihosting_console_outc
  semihosting: Use console_out_gf for SYS_WRITEC
  semihosting: Use console_in_gf for SYS_READC
  semihosting: Create qemu_semihosting_guestfd_init
  semihosting: Add GuestFDConsole
  semihosting: Create qemu_semihosting_console_write
  semihosting: Cleanup chardev init
  semihosting: Expand qemu_semihosting_console_inc to read
  semihosting: Pass CPUState to qemu_semihosting_console_inc
  semihosting: Fix docs comment for qemu_semihosting_console_inc
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/nios2: Move nios2-semi.c to nios2_softmmu_ss

Semihosting is not enabled for nios2-linux-user.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/nios2: Eliminate nios2_semi_is_lseek

Reorg nios2_semi_return_* to gdb_syscall_complete_cb.
Use the 32-bit version normally, and the 64-bit version
for HOSTED_LSEEK.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/mips: Drop pread and pwrite syscalls from semihosting

We don't implement it with _WIN32 hosts, and the syscalls
are missing from the gdb remote file i/o interface.
Since we can't implement them universally, drop them.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/mips: Add UHI errno values

From the Unified Hosting Interface, MD01069 Reference Manual,
version 1.1.6, 06 July 2015.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/mips: Use an exception for semihosting

Within do_interrupt, we hold the iothread lock, which
is required for Chardev access for the console, and for
the round trip for use_gdb_syscalls().

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/m68k: Make semihosting system only

While we had a call to do_m68k_semihosting in linux-user, it
wasn't actually reachable.  We don't include DISAS_INSN(halt)
as an instruction unless system mode.

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

target/m68k: Eliminate m68k_semi_is_fseek

Reorg m68k_semi_return_* to gdb_syscall_complete_cb.
Use the 32-bit version normally, and the 64-bit version
for HOSTED_LSEEK.

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

semihosting: Create semihost_sys_poll_one

This will be used for implementing the xtensa select_one
system call.  Choose "poll" over "select" so that we can
reuse Glib's g_poll constants and to avoid struct timeval.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Remove qemu_semihosting_console_outs

This function has been replaced by *_write.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Use console_out_gf for SYS_WRITE0

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Remove qemu_semihosting_console_outc

This function has been replaced by *_write.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Use console_out_gf for SYS_WRITEC

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Use console_in_gf for SYS_READC

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Create qemu_semihosting_guestfd_init

For arm-compat, initialize console_{in,out}_gf;
otherwise, initialize stdio file descriptors.

This will go some way to cleaning up arm-compat, and
will allow other semihosting to use normal stdio.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Add GuestFDConsole

Add a GuestFDType for connecting to the semihosting console.
Hook up to read, write, isatty, and fstat syscalls.

Note that the arm-specific syscall flen cannot be applied
to the console, because the console is not a descriptor
exposed to the guest.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Create qemu_semihosting_console_write

Will replace qemu_semihosting_console_{outs,outc},
but we need more plumbing first.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Cleanup chardev init

Rename qemu_semihosting_connect_chardevs to
qemu_semihosting_chardev_init; pass the result
directly to qemu_semihosting_console_init.

Store the chardev in SemihostingConsole instead
of SemihostingConfig, which lets us drop
semihosting_get_chardev.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Expand qemu_semihosting_console_inc to read

Allow more than one character to be read at one time.
Will be used by m68k and nios2 semihosting for stdio.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Pass CPUState to qemu_semihosting_console_inc

We don't need CPUArchState, and we do want the CPUState of the
thread performing the operation -- use this instead of current_cpu.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Fix docs comment for qemu_semihosting_console_inc

The implementation of qemu_semihosting_console_inc does not
defer to gdbstub, but only reads from the fifo in console.c.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

gdbstub: Adjust gdb_syscall_complete_cb declaration

Change 'ret' to uint64_t.  This resolves a FIXME in the
m68k and nios2 semihosting that we've lost data.
Change 'err' to int.  There is nothing target-specific
about the width of the errno value.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Create semihost_sys_gettimeofday

This syscall will be used by m68k and nios2 semihosting.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Create semihost_sys_{stat,fstat}

These syscalls will be used by m68k and nios2 semihosting.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_system

Split out the non-ARM specific portions of SYS_SYSTEM to a
reusable function.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_rename

Split out the non-ARM specific portions of SYS_RENAME to a
reusable function.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_remove

Split out the non-ARM specific portions of SYS_REMOVE to a
reusable function.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_flen

The ARM-specific SYS_FLEN isn't really something that can be
reused by other semihosting apis, but there are parts that can
reused for the implementation of semihost_sys_fstat.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_isatty

Split out the non-ARM specific portions of SYS_ISTTY to a
reusable function.  This handles all GuestFD.

Add a common_semi_istty_cb helper to translate the Posix
error return, 0+ENOTTY, to the Arm semihosting not-a-file
success result.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_lseek

Split out the non-ARM specific portions of SYS_SEEK to a
reusable function.  This handles all GuestFD.  Isolate the
curious ARM-specific return value processing to a new
callback, common_semi_seek_cb.

Expand the internal type of the offset to int64_t, and
provide the whence argument, which will be required by
m68k and nios2 semihosting.

Note that gdb_do_syscall %x reads target_ulong, not int.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Bound length for semihost_sys_{read,write}

Fixes a minor bug in which a 64-bit guest on a 32-bit host could
truncate the length.  This would only ever cause a problem if
there were no bits set in the low 32, so that it truncates to 0.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_write

Split out the non-ARM specific portions of SYS_WRITE to a
reusable function.  This handles all GuestFD.  This removes
the last use of common_semi_syscall_len.

Note that gdb_do_syscall %x reads target_ulong, not int.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_read

Split out the non-ARM specific portions of SYS_READ to a
reusable function.  This handles all GuestFD.  Isolate the
curious ARM-specific return value processing to a new
callback, common_semi_rw_cb.

Note that gdb_do_syscall %x reads target_ulong, not int.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_close

Split out the non-ARM specific portions of SYS_CLOSE to a
reusable function.  This handles all GuestFD.

Note that gdb_do_syscall %x reads target_ulong, not int.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out semihost_sys_open

Split out the non-ARM specific portions of SYS_OPEN to a
reusable function.  This handles gdb and host file i/o.

Add helpers to validate the length of the filename string.
Prepare for usage by other semihosting by allowing the
filename length parameter to be 0, and calling strlen.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Move GET_ARG/SET_ARG earlier in the file

Moving this to be useful for another function
besides do_common_semihosting.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Use env more often in do_common_semihosting

We've already loaded cs->env_ptr into a local variable; use it.
Since env is unconditionally used, we don't need a dummy use.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out common-semi-target.h

Move the ARM and RISCV specific helpers into
their own header file.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

semihosting: Split out common_semi_has_synccache

We already have some larger ifdef blocks for ARM and RISCV;
split out a boolean test for SYS_SYNCCACHE.

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

semihosting: Split common_semi_flen_buf per target

We already have some larger ifdef blocks for ARM and RISCV;
split out common_semi_stack_bottom per target.

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

semihosting: Split is_64bit_semihosting per target

We already have some larger ifdef blocks for ARM and RISCV;
split the function into multiple implementations per arch.

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

semihosting: Use struct gdb_stat in common_semi_flen_cb

Load the entire 64-bit size value.  While we're at it,
use offsetof instead of an integer constant.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

gdbstub: Convert GDB error numbers to host error numbers

Provide the callback with consistent state -- always use
host error numbers.  The individual callback can then
decide if the errno requires conversion for the guest.

Reviewed-by: Luc Michel <lmichel@kalray.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

include/exec: Define errno values in gdbstub.h

Define constants for the errno values defined by the
gdb remote fileio protocol.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>