Under KVM, the kernel gets the HVC call and handle the PSCI requests.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-20-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This code is specific to the SoftFloat floating-point
implementation, which is only used by TCG.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-18-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The vfp_set_fpscr() helper contains code specific to the host
floating point implementation (here the SoftFloat library).
Extract this code to vfp_set_fpscr_from_host().
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-17-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The vfp_set_fpscr() helper contains code specific to the host
floating point implementation (here the SoftFloat library).
Extract this code to vfp_set_fpscr_to_host().
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-16-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
To ease the review of the next commit,
move the vfp_exceptbits_to_host() function directly after
vfp_exceptbits_from_host(). Amusingly the diff shows we
are moving vfp_get_fpscr().
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-15-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
These routines are TCG specific.
The arm_deliver_fault() function is only used within the new
helper. Make it static.
Suggested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-13-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
In the next commit we will split the TLB related routines of
this file, and this function will also be called in the new
file. Declare it in the "internals.h" header.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-12-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Suggested-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-11-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Those helpers are a software implementation of the ARM v8 memory zeroing
op code. They should be moved to the op helper file, which is going to
eventually be built only when TCG is enabled.
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Robert Bradford <robert.bradford@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-10-philmd@redhat.com
[PMD: Rebased]
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Since we'll move this code around, fix its style first.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-9-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Since commit 8c06fbdf36 checkpatch.pl enforce a new multiline
comment syntax. Since we'll move this code around, fix its style
first.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-8-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-7-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Robert Bradford <robert.bradford@intel.com>
Reviewed-by: Samuel Ortiz <sameo@linux.intel.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-6-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Group SOFTMMU objects together.
Since PSCI is TCG specific, keep it separate.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-5-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Group KVM rules together.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-4-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Group ARM objects together, TCG related ones at the bottom.
This will help when restricting TCG-only objects.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-3-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Group Aarch64 rules together, TCG related ones at the bottom.
This will help when restricting TCG-only objects.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190701132516.26392-2-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Following the previous patch, this patch adds peripheral devices to the
newly introduced SBSA-ref machine.
Signed-off-by: Hongbo Zhang <hongbo.zhang@linaro.org>
Message-id: 1561890034-15921-3-git-send-email-hongbo.zhang@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
For AArch64, the existing "virt" machine is primarily meant to
run on KVM and execute virtualization workloads, but we need an
environment as faithful as possible to physical hardware, for supporting
firmware and OS development for physical Aarch64 machines.
This patch introduces new machine type 'sbsa-ref' with main features:
- Based on 'virt' machine type.
- A new memory map.
- CPU type cortex-a57.
- EL2 and EL3 are enabled.
- GIC version 3.
- System bus AHCI controller.
- System bus EHCI controller.
- CDROM and hard disc on AHCI bus.
- E1000E ethernet card on PCIE bus.
- VGA display adaptor on PCIE bus.
- No virtio devices.
- No fw_cfg device.
- No ACPI table supplied.
- Only minimal device tree nodes.
Arm Trusted Firmware and UEFI porting to this are done accordingly,
and the firmware should supply ACPI tables to the guest OS. The
minimal device tree nodes supplied by QEMU for this platform are only
to pass the dynamic info reflecting command line input to firmware,
not for loading the guest OS.
To make the review easier, this task is split into two patches, the
fundamental skeleton part and the peripheral devices part; this patch is
the first part.
Signed-off-by: Hongbo Zhang <hongbo.zhang@linaro.org>
Message-id: 1561890034-15921-2-git-send-email-hongbo.zhang@linaro.org
[PMM: commit message tweaks; moved some bits between patch 1 and 2
to ensure patch 1 builds cleanly; removed unneeded lines from
Kconfig stanza; only provide board for qemu-system-aarch64, not
qemu-system-arm; added MAINTAINERS entry]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The ast2500 uses the watchdog to reset the SDRAM controller. This
operation is usually performed by u-boot's memory training procedure,
and it is enabled by setting a bit in the SCU and then causing the
watchdog to expire. Therefore, we need the watchdog to be able to
access the SCU's register space.
This causes the watchdog to not perform a system reset when the bit is
set. In the future it could perform a reset of the SDMC model.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190621065242.32535-1-joel@jms.id.au
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The legacy interface only supported up to 32 IRQs, which became
restrictive around the AST2400 generation. QEMU support for the SoCs
started with the AST2400 along with an effort to reimplement and
upstream drivers for Linux, so up until this point the consumers of the
QEMU ASPEED support only required the 64 IRQ register interface.
In an effort to support older BMC firmware, add support for the 32 IRQ
interface.
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 20190618165311.27066-22-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The XDMA engine embedded in the Aspeed SOCs performs PCI DMA operations
between the SOC (acting as a BMC) and a host processor in a server.
The XDMA engine exists on the AST2400, AST2500, and AST2600 SOCs, so
enable it for all of those. Add trace events on the important register
writes in the XDMA engine.
Signed-off-by: Eddie James <eajames@linux.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-id: 20190618165311.27066-21-clg@kaod.org
[clg: - changed title ]
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The Swift board is an OpenPOWER system hosting POWER processors.
Add support for their BMC including the I2C devices as found on HW.
Signed-off-by: Adriana Kobylak <anoo@us.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 20190618165311.27066-20-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The DRAM address of a DMA transaction depends on the DRAM base address
of the SoC. Inform the SMC controller model with this value.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190618165311.27066-15-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The RAM memory region is defined after the SoC is realized when the
SDMC controller has checked that the defined RAM size for the machine
is correct. This is problematic for controller models requiring a link
on the RAM region, for DMA support in the SMC controller for instance.
Introduce a container memory region for the RAM that we can link into
the controllers early, before the SoC is realized. It will be
populated with the RAM region after the checks have be done.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 20190618165311.27066-14-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
It has never been used as far as I can tell from the git history.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 20190618165311.27066-13-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
If the host decrements the counter register that results in a negative
delta. This is then passed to muldiv64 which only handles unsigned
numbers resulting in bogus results.
This fix ensures the delta being operated on is positive.
Test case: kexec a kernel using aspeed_timer and it will freeze on the
second bootup when the kernel initializes the timer. With this patch
that no longer happens and the timer appears to run OK.
Signed-off-by: Christian Svensson <bluecmd@google.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Andrew Jeffery <andrew@aj.id.au>
Message-id: 20190618165311.27066-12-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
If the match value exceeds reload then we don't want to include it in
calculations for the next event.
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-id: 20190618165311.27066-10-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
From the datasheet:
This register stores the current status of counter #N. When timer
enable bit TMC30[N * b] is disabled, the reload register will be
loaded into this counter. When timer bit TMC30[N * b] is set, the
counter will start to decrement. CPU can update this register value
when enable bit is set.
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 20190618165311.27066-9-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The Linux kernel driver was updated in commit 4451d3f59f2a
("clocksource/drivers/fttmr010: Fix set_next_event handler) to fix an
issue observed on hardware:
> RELOAD register is loaded into COUNT register when the aspeed timer
> is enabled, which means the next event may be delayed because timer
> interrupt won't be generated until <0xFFFFFFFF - current_count +
> cycles>.
When running under Qemu, the system appeared "laggy". The guest is now
scheduling timer events too regularly, starving the host of CPU time.
This patch modifies the timer model to attempt to schedule the timer
expiry as the guest requests, but if we have missed the deadline we
re interrupt and try again, which allows the guest to catch up.
Provides expected behaviour with old and new guest code.
Fixes: c04bd47db6 ("hw/timer: Add ASPEED timer device model")
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-id: 20190618165311.27066-8-clg@kaod.org
[clg: - merged a fix from Andrew Jeffery <andrew@aj.id.au>
"Fire interrupt on failure to meet deadline"
https://lists.ozlabs.org/pipermail/openbmc/2019-January/014641.html
- adapted commit log
- checkpatch fixes ]
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The Aspeed SoCs have two MACs. Extend the Aspeed model to support a
second NIC.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 20190618165311.27066-7-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The current models of the Aspeed SoCs only have one CPU but future
ones will support SMP. Introduce a new num_cpus field at the SoC class
level to define the number of available CPUs per SoC and also
introduce a 'num-cpus' property to activate the CPUs configured for
the machine.
The max_cpus limit of the machine should depend on the SoC definition
but, unfortunately, these values are not available when the machine
class is initialized. This is the reason why we add a check on
num_cpus in the AspeedSoC realize handler.
SMP support will be activated when models for such SoCs are implemented.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 20190618165311.27066-6-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
All systems have an RTC.
The IRQ is hooked up but the model does not use it at this stage. There
is no guest code that uses it, so this limitation is acceptable.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20190618165311.27066-5-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The RTC is modeled to provide time and date functionality. It is
initialised at zero to match the hardware.
There is no modelling of the alarm functionality, which includes the IRQ
line. As there is no guest code to exercise this function that is
acceptable for now.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20190618165311.27066-4-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This will simplify the definition of new SoCs, like the AST2600 which
should use a slightly different address space and have a different set
of controllers.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 20190618165311.27066-3-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This will simplify the definition of new SoCs, like the AST2600 which
should use a different CPU and a different IRQ number layout.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 20190618165311.27066-2-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Datasheet for i.MX7 is incorrect and i.MX7's PCI IRQ mapping matches
that of i.MX6:
* INTD/MSI 122
* INTC 123
* INTB 124
* INTA 125
Fix all of the relevant code to reflect that fact. Needed by latest
Linux kernels.
(Reference: Linux kernel commit 538d6e9d597584e80 from an
NXP employee confirming that the datasheet is incorrect and
with a report of a test against hardware.)
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: qemu-devel@nongnu.org
Cc: qemu-arm@nongnu.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
[PMM: added ref to kernel commit confirming the datasheet error]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
MSI mapping needs to be update when MSI address changes, so add the
code to do so.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: qemu-devel@nongnu.org
Cc: qemu-arm@nongnu.org
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Expression to calculate update_msi_mapping in code handling writes to
DESIGNWARE_PCIE_MSI_INTR0_ENABLE is missing an ! operator and should
be:
!!root->msi.intr[0].enable ^ !!val;
so that MSI mapping is updated when enabled transitions from either
"none" -> "any" or "any" -> "none". Since that register shouldn't be
written to very often, change the code to update MSI mapping
unconditionally instead of trying to fix the update_msi_mapping logic.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: qemu-devel@nongnu.org
Cc: qemu-arm@nongnu.org
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Add no-op/unimplemented PCIE PHY IP block. Needed by new kernels to
use PCIE.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: qemu-devel@nongnu.org
Cc: qemu-arm@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Allow cortex-a7 to be used with the virt board; it supports
the v7VE features and there is no reason to deny this type.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: fc5404f7-4d1d-c28f-6e48-d8799c82acc0@web.de
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This machine correctly defines its default_cpu_type to cortex-m3
and report an error if the user requested another cpu_type,
however it does not exit, and this can confuse users trying
to use another core:
$ qemu-system-arm -M emcraft-sf2 -cpu cortex-m4 -kernel test-m4.elf
qemu-system-arm: This board can only be used with CPU cortex-m3-arm-cpu
[output related to M3 core ...]
The CPU is indeed a M3 core:
(qemu) info qom-tree
/machine (emcraft-sf2-machine)
/unattached (container)
/device[0] (msf2-soc)
/armv7m (armv7m)
/cpu (cortex-m3-arm-cpu)
Add the missing exit() call to return to the shell.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-id: 20190617160136.29930-1-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Fix the condition used to check whether the initrd fits
into RAM; in some cases if an initrd was also passed on
the command line we would get an error stating that it
was too big to fit into RAM after the kernel. Despite the
error the loader continued anyway, though, so also add an
exit(1) when the initrd is actually too big.
Fixes: 852dc64d66 ("hw/arm/boot: Diagnose layouts that put initrd or
DTB off the end of RAM")
Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190618125844.4863-1-drjones@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Rename function arguments to make intent clearer.
Better documentation for slot control logic.
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
During boot, linux guests tend to clear all bits in pcie slot status
register which is used for hotplug.
If they clear bits that weren't set this is racy and will lose events:
not a big problem for manual hotplug on bare-metal, but a problem for us.
For example, the following is broken ATM:
/x86_64-softmmu/qemu-system-x86_64 -enable-kvm -S -machine q35 \
-device pcie-root-port,id=pcie_root_port_0,slot=2,chassis=2,addr=0x2,bus=pcie.0 \
-device virtio-balloon-pci,id=balloon,bus=pcie_root_port_0 \
-monitor stdio disk.qcow2
(qemu)device_del balloon
(qemu)cont
Balloon isn't deleted as it should.
As a work-around, detect this attempt to clear slot status and revert
status to what it was before the write.
Note: in theory this can be detected as a duplicate button press
which cancels the previous press. Does not seem to happen in
practice as guests seem to only have this bug during init.
Note2: the right thing to do is probably to fix Linux to
read status before clearing it, and act on the bits that are set.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Tested-by: Igor Mammedov <imammedo@redhat.com>
During boot, linux would sometimes overwrites control of a powered off
slot before powering it on. Unfortunately QEMU interprets that as a
power off request and ejects the device.
For example:
/x86_64-softmmu/qemu-system-x86_64 -enable-kvm -S -machine q35 \
-device pcie-root-port,id=pcie_root_port_0,slot=2,chassis=2,addr=0x2,bus=pcie.0 \
-monitor stdio disk.qcow2
(qemu)device_add virtio-balloon-pci,id=balloon,bus=pcie_root_port_0
(qemu)cont
Balloon is deleted during guest boot.
To fix, save control beforehand and check that power
or led state actually change before ejecting.
Note: this is more a hack than a solution, ideally we'd
find a better way to detect ejects, or move away
from ejects completely and instead monitor whether
it's safe to delete device due to e.g. its power state.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Tested-by: Igor Mammedov <imammedo@redhat.com>
