IEEE 758 does not define a fixed rule for which NaN to pick as the
result if both operands of a 2-operand operation are NaNs. As a
result different architectures have ended up with different rules for
propagating NaNs.
QEMU currently hardcodes the NaN propagation logic into the binary
because pickNaN() has an ifdef ladder for different targets. We want
to make the propagation rule instead be selectable at runtime,
because:
* this will let us have multiple targets in one QEMU binary
* the Arm FEAT_AFP architectural feature includes letting
the guest select a NaN propagation rule at runtime
* x86 specifies different propagation rules for x87 FPU ops
and for SSE ops, and specifying the rule in the float_status
would let us emulate this, instead of wrongly using the
x87 rules everywhere
In this commit we add an enum for the propagation rule, the field in
float_status, and the corresponding getters and setters. We change
pickNaN to honour this, but because all targets still leave this
field at its default 0 value, the fallback logic will pick the rule
type with the old ifdef ladder.
It's valid not to set a propagation rule if default_nan_mode is
enabled, because in that case there's no need to pick a NaN; all the
callers of pickNaN() catch this case and skip calling it. So we can
already assert that we don't get into the "no rule defined" codepath
for our four targets which always set default_nan_mode: Hexagon,
RiscV, SH4 and Tricore, and for the one target which does not have FP
at all: avr. These targets will not need to be updated to call
set_float_2nan_prop_rule().
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20241025141254.2141506-2-peter.maydell@linaro.org
* Fixed eMMC size calculation
* Fixed IRQ definitions on AST2700
* Added RTC support to AST2700
* Fixed timer IRQ status on AST2600
* Improved SDHCI model with new registers
* Added -nodefaults support to AST1030
* Provided a way to use an eMMC device without boot partitions
-----BEGIN PGP SIGNATURE-----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=ms8H
-----END PGP SIGNATURE-----
Merge tag 'pull-aspeed-20241104' of https://github.com/legoater/qemu into staging
aspeed queue:
* Fixed eMMC size calculation
* Fixed IRQ definitions on AST2700
* Added RTC support to AST2700
* Fixed timer IRQ status on AST2600
* Improved SDHCI model with new registers
* Added -nodefaults support to AST1030
* Provided a way to use an eMMC device without boot partitions
# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEoPZlSPBIlev+awtgUaNDx8/77KEFAmcoo4oACgkQUaNDx8/7
# 7KGYAQ/9GWiwM7SHFD/WTEo6iClQCk+Do3pzGXZPQq7WLqYhBU8mYwSaqMDUtXj+
# MQVywyLxSYaKdCKessN0haATyzEDVRtxKwIRnbrSDWWnxG8NGj2esOTsU6/wgfD4
# FqARaMH91FQB6rY8QbmbGmqTJ1QbWEPXj7v2piJol5dvI2Oe8iqn/6z1Cv4NMXwh
# aYHwSVwcHLD9tfmyXP0DKN/XHLC4pTAOoU96ajcN6RRW+D6vuQEsQq0caZt8CHQc
# I2oSptU+RZF2DPbSeEB42y9I138/kQzTIaVnbBN//NLRwbzRsLlXhA92F2CJyDrD
# FGNQyynteil8F7M5Oab47fFia1QF/v4G45VOAsHpT1tLBsZPKJdRwfLLqDPZbVbG
# 2lAVuukqW0gKoEHsXfVsDzcIxpX81SlUsccHY4kCxsRNnwSzCWaDK9OOTx3CAxjG
# CzzDgQszNr/12dzkWExIhLpMhQNeiUXX1veAH/jzbjyRAKxzjkDYaX2lUC3MfmqX
# irjmzOU0AbtComv4ybeBqtqNmQvUx5/y993Hgakc9mqqCoAm/Fn4qtx6uW5vSSZJ
# w4heyWbzcLp5RIzSYZypWlmgI+3bJgJq2aX276MYqAe3m8PnUCkuW9NTsfb+ARMl
# XGExHPNrAsw7eiiQsTa7Byt/jkEf3KmEp8ye+3cAvJwPgxlDyys=
# =ms8H
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 04 Nov 2024 10:35:54 GMT
# gpg: using RSA key A0F66548F04895EBFE6B0B6051A343C7CFFBECA1
# gpg: Good signature from "Cédric Le Goater <clg@redhat.com>" [full]
# gpg: aka "Cédric Le Goater <clg@kaod.org>" [full]
# Primary key fingerprint: A0F6 6548 F048 95EB FE6B 0B60 51A3 43C7 CFFB ECA1
* tag 'pull-aspeed-20241104' of https://github.com/legoater/qemu:
aspeed: Don't set always boot properties of the emmc device
aspeed: Support create flash devices via command line for AST1030
hw/sd/aspeed_sdhci: Introduce Capabilities Register 2 for SD slot 0 and 1
hw/timer/aspeed: Fix interrupt status does not be cleared for AST2600
hw/timer/aspeed: Fix coding style
aspeed/soc: Support RTC for AST2700
hw/arm/aspeed_ast27x0: Avoid hardcoded '256' in IRQ calculation
hw/arm/aspeed_ast27x0: Use bsa.h for PPI definitions
hw/sd/sdcard: Fix calculation of size when using eMMC boot partitions
hw/arm: enable at24c with aspeed
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
* Big cleanup of deprecated machines
* Power11 support for spapr
* XIVE improvements
* Goodbye to Cedric and David as ppc reviewers, thank you both o7
-----BEGIN PGP SIGNATURE-----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=vmIG
-----END PGP SIGNATURE-----
Merge tag 'pull-ppc-for-9.2-1-20241104' of https://gitlab.com/npiggin/qemu into staging
* Various bug fixes
* Big cleanup of deprecated machines
* Power11 support for spapr
* XIVE improvements
* Goodbye to Cedric and David as ppc reviewers, thank you both o7
# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCgAdFiEETkN92lZhb0MpsKeVZ7MCdqhiHK4FAmcoEicACgkQZ7MCdqhi
# HK5M8Q//fz+ZkJndXkBjb1Oinx+q+eVtNm2JrvcWIsXyhG3K+6VxYPp69H+SRv/Z
# TWuUqMQPxq8mhQvBJlDAttp/oaUEiOcCRvs/iUoBN12L4mVxXfdoT88TZ4frN3eP
# 8bePq+DW2N/7gpmsJm5CyEZPpcf9AjVHgLRp3KYFkOJ/14uzvuwnocU39gl+2IUh
# MXHTedQgMNXaKorJXk1NVdM6NxMuVhOvwxAs6ya2gwhxyA5tteo5PiQOnDJWkejf
# xg3RRsNzGYcs1Qg/3kFIf3RfEB0aYbPxROM8IfPaJWKN5KnMggj/JAkHyK1x/V3J
# wml7+cB0doMt/yRiuYJhXpyrtOqpvjRWPA6RhxECWW2kwrovv8NAF8IrFnw9NvOQ
# QC66ZaaFcbAcFrVT1e/iggU76d01II6m4OAgKcXw+FRHgps4VU9y83j7ApNnNUWN
# IXp9hkzoHi5VwX0FrG4ELUr2iEf1HASMvM8EZ/0AxzWj5iNtQB8lFsrEdaGVXyIS
# M5JaJeNjCn4koCyYaFSctH5eKtbzIwnGWnDcdTwaOuQ+9itBvY8O+HZalE6sAc5S
# kLFZ7i/Ut/qxbY5pMumt8LKD4pR1SsOxFB8dJCmn/f/tvRGtIVsoY6btNe4M0+24
# 42MxZbWO6W379C32bwbtsPiGA+aLSgShjP4cWm9cgRjz4RJFnwg=
# =vmIG
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 04 Nov 2024 00:15:35 GMT
# gpg: using RSA key 4E437DDA56616F4329B0A79567B30276A8621CAE
# gpg: Good signature from "Nicholas Piggin <npiggin@gmail.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: 4E43 7DDA 5661 6F43 29B0 A795 67B3 0276 A862 1CAE
* tag 'pull-ppc-for-9.2-1-20241104' of https://gitlab.com/npiggin/qemu: (67 commits)
MAINTAINERS: Remove myself as reviewer
MAINTAINERS: Remove myself from XIVE
MAINTAINERS: Remove myself from the PowerNV machines
hw/ppc: Consolidate ppc440 initial mapping creation functions
hw/ppc: Consolidate e500 initial mapping creation functions
tests/qtest: Add XIVE tests for the powernv10 machine
pnv/xive2: TIMA CI ops using alternative offsets or byte lengths
pnv/xive2: TIMA support for 8-byte OS context push for PHYP
pnv/xive: Update PIPR when updating CPPR
pnv/xive: Add special handling for pool targets
ppc/xive2: Support "Pull Thread Context to Odd Thread Reporting Line"
ppc/xive2: Change context/ring specific functions to be generic
ppc/xive2: Support "Pull Thread Context to Register" operation
ppc/xive2: Allow 1-byte write of Target field in TIMA
ppc/xive2: Dump the VP-group and crowd tables with 'info pic'
ppc/xive2: Dump more NVP state with 'info pic'
pnv/xive2: Support for "OS LGS Push" TIMA operation
ppc/xive2: Support TIMA "Pull OS Context to Odd Thread Reporting Line"
pnv/xive2: Define OGEN field in the TIMA
pnv/xive: TIMA patch sets pre-req alignment and formatting changes
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Now replaced by meson build.
Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Message-Id: <20241023212812.1376972-4-pierrick.bouvier@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tried to unify this meson.build with tests/tcg/plugins/meson.build but
the resulting modules are not output in the right directory.
Originally proposed by Anton Kochkov, thank you!
Solves: https://gitlab.com/qemu-project/qemu/-/issues/1710
Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Message-Id: <20241023212812.1376972-3-pierrick.bouvier@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
contrib/plugins/cflow.c: In function ‘plugin_exit’:
contrib/plugins/cflow.c:167:19: error: declaration of ‘n’ shadows a previous local [-Werror=shadow=local]
167 | NodeData *n = l->data;
| ^
contrib/plugins/cflow.c:139:9: note: shadowed declaration is here
139 | int n = 0;
| ^
Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Message-Id: <20241023212812.1376972-2-pierrick.bouvier@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Currently we only set it on fork.
Note: Upstream (blitz) commit also did new threads, but that code isn't
in qemu project repo yet.
Signed-off-by: Jessica Clarke <jrtc27@jrtc27.com>
Pull-Request: https://github.com/qemu-bsd-user/qemu-bsd-user/pull/52
Reviewed-by: Warner Losh <imp@bsdimp.com>
Signed-off-by: Warner Losh <imp@bsdimp.com>
While qemu-system can set tb-size using -accel tcg,tb-size=n, there
is no similar knob for qemu-bsd-user. Add one in a way similar to how
one-insn-per-tb is already handled.
Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Warner Losh <imp@bsdimp.com>
Signed-off-by: Warner Losh <imp@bsdimp.com>
bsd-user qemu-x86_64 almost immediately dies with:
qemu: 0x4002201a68: unhandled CPU exception 0xd - aborting
on FreeBSD 14.1-RELEASE. This is an instruction that requires
alignment:
(gdb) x/i 0x4002201a68
0x4002201a68: movaps %xmm0,-0x40(%rbp)
and the argument is not aligned:
(gdb) p/x env->regs[5]
$1 = 0x822443b58
A quick experiment shows that the userspace entry point expects
misaligned rsp:
(gdb) starti
(gdb) p/x $rsp
$1 = 0x7fffffffeaa8
Emulate this behavior in bsd-user.
[[ applied Richard's suggestion ]]
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Warner Losh <imp@bsdimp.com>
Signed-off-by: Warner Losh <imp@bsdimp.com>
IEC descriptor is 128-bit invalidation descriptor, must be padded with
128-bits of 0s in the upper bytes to create a 256-bit descriptor when
the invalidation queue is configured for 256-bit descriptors (IQA_REG.DW=1).
Fixes: 02a2cbc872 ("x86-iommu: introduce IEC notifiers")
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Message-Id: <20241104125536.1236118-4-zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
According to VTD spec, a 256-bit descriptor will result in an invalid
descriptor error if submitted in an IQ that is setup to provide hardware
with 128-bit descriptors (IQA_REG.DW=0). Meanwhile, there are old inv desc
types (e.g. iotlb_inv_desc) that can be either 128bits or 256bits. If a
128-bit version of this descriptor is submitted into an IQ that is setup
to provide hardware with 256-bit descriptors will also result in an invalid
descriptor error.
The 2nd will be captured by the tail register update. So we only need to
focus on the 1st.
Because the reserved bit check between different types of invalidation desc
are common, so introduce a common function vtd_inv_desc_reserved_check()
to do all the checks and pass the differences as parameters.
With this change, need to replace error_report_once() call with error_report()
to catch different call sites. This isn't an issue as error_report_once()
here is mainly used to help debug guest error, but it only dumps once in
qemu life cycle and doesn't help much, we need error_report() instead.
Fixes: c0c1d35184 ("intel_iommu: add 256 bits qi_desc support")
Suggested-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Message-Id: <20241104125536.1236118-3-zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
According to VTD spec, Figure 11-22, Invalidation Queue Tail Register,
"When Descriptor Width (DW) field in Invalidation Queue Address Register
(IQA_REG) is Set (256-bit descriptors), hardware treats bit-4 as reserved
and a value of 1 in the bit will result in invalidation queue error."
Current code missed to send IQE event to guest, fix it.
Fixes: c0c1d35184 ("intel_iommu: add 256 bits qi_desc support")
Suggested-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Message-Id: <20241104125536.1236118-2-zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The ACPI CPU hotplug states must be migrated along with other vCPU
hotplug states to the destination VM. Update the GED's VM State
Description (VMSD) table subsection to conditionally include the CPU
Hotplug VM State Description (VMSD).
Excerpt of GED VMSD State Dump at Source:
"acpi-ged (16)": {
"ged_state": {
"sel": "0x00000000"
},
[...]
"acpi-ged/cpuhp": {
"cpuhp_state": {
"selector": "0x00000005",
"command": "0x00",
"devs": [
{
"is_inserting": false,
"is_removing": false,
"ost_event": "0x00000000",
"ost_status": "0x00000000"
},
[...]
{
"is_inserting": false,
"is_removing": false,
"ost_event": "0x00000000",
"ost_status": "0x00000000"
}
]
}
}
},
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Message-Id: <20241103102419.202225-6-salil.mehta@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Update DSDT golden master files for x86/pc and x86/q35 platforms to
accommodate changes made in the architecture-agnostic CPU AML. These
updates notify the guest OS of vCPU hot-plug and hot-unplug status
using the ACPI `_STA.Enabled` bit.
The following is a diff of the changes in the .dsl file generated with
IASL:
@@ -1480,6 +1480,7 @@
CRMV, 1,
CEJ0, 1,
CEJF, 1,
+ CPRS, 1,
Offset (0x05),
CCMD, 8
}
@@ -1514,9 +1515,16 @@
Acquire (\_SB.PCI0.PRES.CPLK, 0xFFFF)
\_SB.PCI0.PRES.CSEL = Arg0
Local0 = Zero
- If ((\_SB.PCI0.PRES.CPEN == One))
- {
- Local0 = 0x0F
+ If ((\_SB.PCI0.PRES.CPRS == One))
+ {
+ If ((\_SB.PCI0.PRES.CPEN == One))
+ {
+ Local0 = 0x0F
+ }
+ Else
+ {
+ Local0 = 0x0D
+ }
}
Release (\_SB.PCI0.PRES.CPLK)
Reported-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reflect the QOM vCPUs ACPI CPU hotplug states in the `_STA.Present` and
and `_STA.Enabled` bits when the guest kernel evaluates the ACPI
`_STA` method during initialization, as well as when vCPUs are
hot-plugged or hot-unplugged. If the CPU is present then the its
`enabled` status can be fetched using architecture-specific code [1].
Reference:
[1] Example implementation of architecture-specific hook to fetch CPU
`enabled status
Link: c0b416b11e
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Message-Id: <20241103102419.202225-4-salil.mehta@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
list changed files in tests/qtest/bios-tables-test-allowed-diff.h
Reported-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Message-Id: <20241103102419.202225-3-salil.mehta@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
On most architectures, during vCPU hot-plug and hot-unplug actions, the
firmware or VMM/QEMU can update the OS on vCPU status by toggling the
ACPI method `_STA.Present` bit. However, certain CPU architectures
prohibit [1] modifications to a CPU’s `presence` status after the kernel
has booted.
This limitation [2][3] exists because many per-CPU components, such as
interrupt controllers and various per-CPU features tightly integrated
with CPUs, may not support reconfiguration once the kernel is
initialized. Often, these components cannot be powered down, as they may
belong to an `always-on` power domain. As a result, some architectures
require all CPUs to remain `_STA.Present` after system initialization.
Therefore, it is essential to mirror the exact QOM vCPU status through
ACPI for the Guest kernel. For this, we should determine—via
architecture-specific code[4]—whether vCPUs must always remain present
and whether the associated `AcpiCpuStatus::cpu` object should remain
valid, even following a vCPU hot-unplug operation.
References:
[1] Check comment 5 in the bugzilla entry
Link: https://bugzilla.tianocore.org/show_bug.cgi?id=4481#c5
[2] KVMForum 2023 Presentation: Challenges Revisited in Supporting Virt CPU Hotplug on
architectures that don’t Support CPU Hotplug (like ARM64)
a. Kernel Link: https://kvm-forum.qemu.org/2023/KVM-forum-cpu-hotplug_7OJ1YyJ.pdf
b. Qemu Link: https://kvm-forum.qemu.org/2023/Challenges_Revisited_in_Supporting_Virt_CPU_Hotplug_-__ii0iNb3.pdf
[3] KVMForum 2020 Presentation: Challenges in Supporting Virtual CPU Hotplug on
SoC Based Systems (like ARM64)
Link: https://kvmforum2020.sched.com/event/eE4m
[4] Example implementation of architecture-specific CPU persistence hook
Link: c0b416b11e
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Message-Id: <20241103102419.202225-2-salil.mehta@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The bus parameter in the macro PCI_BUILD_BDF is not surrounded by
parenthesis. This can create a compile error when warnings are
treated as errors or can potentially create runtime errors due to the
operator precedence.
For instance:
file.c❌32: error: suggest parentheses around '-' inside '<<'
[-Werror=parentheses]
171 | uint16_t bdf = PCI_BUILD_BDF(a - b, sdev->devfn);
| ~~^~~
include/hw/pci/pci.h:19:41: note: in definition of macro
'PCI_BUILD_BDF'
19 | #define PCI_BUILD_BDF(bus, devfn) ((bus << 8) | (devfn))
| ^~~
cc1: all warnings being treated as errors
Signed-off-by: Roque Arcudia Hernandez <roqueh@google.com>
Reviewed-by: Nabih Estefan <nabihestefan@google.com>
Message-Id: <20241101215923.3399311-1-roqueh@google.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
If len_in is smaller than the header length then the accessing the
number of ports will result in an out of bounds access.
Add a check to avoid this.
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-11-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The properties of the requested set command cannot be established if
len_in is less than the size of the header.
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-10-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In cmd_features_set_feature() the an offset + data size schemed
is used to allow for large features. Ensure this does not write
beyond the end fo the buffers used to accumulate the full feature
attribute set.
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-9-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
If len_in is less than the minimum spec allowed value, then return
CXL_MBOX_INVALID_PAYLOAD_LENGTH
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-8-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add a check that the requested offset + length does not go beyond the end
of the cel_log.
Whilst the cci->cel_log is large enough to include all possible CEL
entries, the guest might still ask for entries beyond the end of it.
Move the comment to this new check rather than before the check on the
type of log requested.
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-7-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Checking offset + length is of no relevance when verifying the CEL
data will fit in the mailbox payload. Only the length is is relevant.
Note that this removes a potential overflow.
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-6-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Buggy guest can write a message that advertises more data that
is provided. As QEMU internally duplicates the reported message
size, this may result in an out of bounds access.
Add sanity checks on the size to avoid this.
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-5-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Buggy software might write a message that is too short for
either the header, or the header + the event data that is specified
in the header. This may result in accesses beyond the range of the
message allocated as a duplicate of the incoming message buffer.
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-4-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
A buggy guest might write an insufficiently large message.
Check the header is present. Whilst zero data after the header is very
odd it will just result in failure to copy any data.
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-3-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
cxl_cmd_dcd_release_dyn_cap() and cmd_dcd_add_dyn_cap_rsp() are missing
input message size checks. These must be done in the individual
commands when the command has a variable length input payload.
A buggy or malicious guest might send undersized messages via the mailbox.
As that size is used to take a copy of the mailbox content, each command
must check there is sufficient data. In this case the first check is that
there is enough data to read how many extents there are, and the second
that there is enough for those elements to be accessed.
Reported-by: Esifiel <esifiel@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101133917.27634-2-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In the function of retrieving DC extents (cmd_dcd_get_dyn_cap_ext_list),
the output buffer index was not correctly updated while iterating the
extent list on the device, leaving the extents returned incorrect except for
the first one.
Fixes: 1c9221f19e ("hw/mem/cxl_type3: Add DC extent list representative and get DC extent list mailbox support")
Signed-off-by: Fan Ni <fan.ni@samsung.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101132005.26633-3-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In the function cmd_firmware_update_get_info for handling Get FW info
command (0x0200h), the vmem, pmem and DC capacity size check were
incorrect. The size should be aligned to 256MiB, not smaller than
256MiB.
Signed-off-by: Fan Ni <fan.ni@samsung.com>
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241101132005.26633-2-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>From what I read PCI has 32 transactions, PCI Express devices can handle
256 with Extended tag enabled (spec mentions also larger values but I
lack PCIe knowledge).
QEMU leaves 'Extended tag field' with 0 as value:
Capabilities: [e0] Express (v1) Root Complex Integrated Endpoint, IntMsgNum 0
DevCap: MaxPayload 128 bytes, PhantFunc 0
ExtTag- RBE+ FLReset- TEE-IO-
SBSA ACS has test 824 which checks for PCIe device capabilities. BSA
specification [1] (SBSA is on top of BSA) in section F.3.2 lists
expected values for Device Capabilities Register:
Device Capabilities Register Requirement
Role based error reporting RCEC and RCiEP: Hardwired to 1
Endpoint L0s acceptable latency RCEC and RCiEP: Hardwired to 0
L1 acceptable latency RCEC and RCiEP: Hardwired to 0
Captured slot power limit scale RCEC and RCiEP: Hardwired to 0
Captured slot power limit value RCEC and RCiEP: Hardwired to 0
Max payload size value must be compliant with PCIe spec
Phantom functions RCEC and RCiEP: Recommendation is to
hardwire this bit to 0.
Extended tag field Hardwired to 1
1. https://developer.arm.com/documentation/den0094/c/
This change enables Extended tag field. All versioned platforms should
have it disabled for older versions (tested with Arm/virt).
Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Message-Id: <20241023113820.486017-1-marcin.juszkiewicz@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
VT-d spec removed Transient Mapping (TM) field from second-level page-tables
and treat the field as Reserved(0) since revision 3.2.
Changing the field as reserved(0) will break backward compatibility, so
introduce a property "stale-tm" to allow user to control the setting.
Use pc_compat_9_1 to handle the compatibility for machines before 9.2 which
allow guest to set the field. Starting from 9.2, this field is reserved(0)
by default to match spec. Of course, user can force it on command line.
This doesn't impact function of vIOMMU as there was no logic to emulate
Transient Mapping.
Suggested-by: Yi Liu <yi.l.liu@intel.com>
Suggested-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Yi Liu <yi.l.liu@intel.com>
Message-Id: <20241028022514.806657-1-zhenzhong.duan@intel.com>
Reviewed-by: Clément Mathieu--Drif<clement.mathieu--drif@eviden.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
VHOST_USER_BACKEND_SHARED_OBJECT_ADD and
VHOST_USER_BACKEND_SHARED_OBJECT_REMOVE state
in the spec that they return 0 for successful
operations, non-zero otherwise. However,
implementation relies on the return types
of the virtio-dmabuf library, with opposite
semantics (true if everything is correct,
false otherwise). Therefore, current
implementation violates the specification.
Revert the logic so that the implementation
of the vhost-user handling methods matches
the specification.
Fixes: 043e127a12
Fixes: 1609476662
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Albert Esteve <aesteve@redhat.com>
Message-Id: <20241022124615.585596-1-aesteve@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
For the CXL PXB there is additional code after pxb_dev_realize_common()
is called. If that realize failed (e.g. due to an out of range numa_node)
we will get a segfault. Return a bool so the caller can check if the
pxb_dev_realize_common() succeeded or not without having to poke around
in the errp.
Fixes: 4f8db8711c ("hw/pxb: Allow creation of a CXL PXB (host bridge)")
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241014121902.2146424-8-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add missing 4 spaces of indent to structure element.
Reported-by: Davidlohr Bueso <dave@stgolabs.net>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241014121902.2146424-7-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
CXL spec 3.1 section 8.2.9.9.11.2 describes the DDR5 Error Check Scrub (ECS)
control feature.
ECS log capabilities field in following ECS tables, which is common for all
memory media FRUs in a CXL device.
Fix struct CXLMemECSReadAttrs and struct CXLMemECSWriteAttrs to make
log entry type field common.
Fixes: 2d41ce38fb ("hw/cxl/cxl-mailbox-utils: Add device DDR5 ECS control feature")
Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241014121902.2146424-6-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Per cxl spec r3.1, for multiple dynamic capacity event records grouped via
the More flag, the last record in the sequence should clear the More flag.
Before the change, the More flag of the event record is cleared before
the loop of inserting records into the event log, which will leave the flag
always set once it is set in the loop.
Fixes: d0b9b28a5b ("hw/cxl/events: Add qmp interfaces to add/release dynamic capacity extents")
Signed-off-by: Fan Ni <fan.ni@samsung.com>
Link: https://lore.kernel.org/r/20240827164304.88876-2-nifan.cxl@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241014121902.2146424-5-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
When injecting a new poisoned region through qmp_cxl_inject_poison(),
the newly injected region should not overlap with existing poisoned
regions.
The current validation method does not consider the following
overlapping region:
┌───┬───────┬───┐
│a │ b(a) │a │
└───┴───────┴───┘
(a is a newly added region, b is an existing region, and b is a
subregion of a)
Fixes: 9547754f40 ("hw/cxl: QMP based poison injection support")
Signed-off-by: Yao Xingtao <yaoxt.fnst@fujitsu.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241014121902.2146424-4-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The current completion percentage calculation does not account for the
relative time since the start of the background activity, this leads to
showing incorrect start percentage vs what has actually been completed.
This patch calculates the percentage based on the actual elapsed time since
the start of the operation.
Fixes: 221d2cfbdb ("hw/cxl/mbox: Add support for background operations")
Signed-off-by: Ajay Joshi <ajay.opensrc@micron.com>
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Link: https://lore.kernel.org/r/20240729102338.22337-1-ajay.opensrc@micron.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241014121902.2146424-3-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The sum offset + length may overflow uint32. Since this sum is
compared with uint64_t return value of get_lsa_size(), it makes
sense to choose uint64_t type for offset and length.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: 3ebe676a34 ("hw/cxl/device: Implement get/set Label Storage Area (LSA)")
Signed-off-by: Dmitry Frolov <frolov@swemel.ru>
Link: https://lore.kernel.org/r/20240917080925.270597-2-frolov@swemel.ru
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20241014121902.2146424-2-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
During the hot-unplugging of vhost-user-net type network cards,
the vhost_user_cleanup function may add the same rcu node to
the rcu linked list. The function call in this case is as follows:
vhost_user_cleanup
->vhost_user_host_notifier_remove
->call_rcu(n, vhost_user_host_notifier_free, rcu);
->g_free_rcu(n, rcu);
When this happens, QEMU will abort in try_dequeue:
if (head == &dummy && qatomic_mb_read(&tail) == &dummy.next) {
abort();
}
backtrace is as follows:
0 __pthread_kill_implementation () at /usr/lib64/libc.so.6
1 raise () at /usr/lib64/libc.so.6
2 abort () at /usr/lib64/libc.so.6
3 try_dequeue () at ../util/rcu.c:235
4 call_rcu_thread (0) at ../util/rcu.c:288
5 qemu_thread_start (0) at ../util/qemu-thread-posix.c:541
6 start_thread () at /usr/lib64/libc.so.6
7 clone3 () at /usr/lib64/libc.so.6
The reason for the abort is that adding two identical nodes to
the rcu linked list will cause the rcu linked list to become a ring,
but when the dummy node is added after the two identical nodes,
the ring is opened. But only one node is added to list with
rcu_call_count added twice. This will cause rcu try_dequeue abort.
This happens when n->addr != 0. In some scenarios, this does happen.
For example, this situation will occur when using a 32-queue DPU
vhost-user-net type network card for hot-unplug testing, because
VhostUserHostNotifier->addr will be cleared during the processing of
VHOST_USER_BACKEND_VRING_HOST_NOTIFIER_MSG. However,it is asynchronous,
so we cannot guarantee that VhostUserHostNotifier->addr is zero in
vhost_user_cleanup. Therefore, it is necessary to merge g_free_rcu
and vhost_user_host_notifier_free into one rcu node.
Fixes: 503e355465 ("virtio/vhost-user: dynamically assign VhostUserHostNotifiers")
Signed-off-by: yaozhenguo <yaozhenguo@jd.com>
Message-Id: <20241011102913.45582-1-yaozhenguo@jd.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The XTSup mode enables x2APIC support for AMD IOMMU, which is needed
to support vcpu w/ APIC ID > 255.
Reviewed-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Signed-off-by: Santosh Shukla <santosh.shukla@amd.com>
Message-Id: <20240927172913.121477-6-santosh.shukla@amd.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In order to support AMD IOMMU interrupt remapping emulation with PCI
pass-through devices, QEMU needs to notify VFIO when guest IOMMU driver
updates and invalidate the guest interrupt remapping table (IRT), and
communicate information so that the host IOMMU driver can update
the shadowed interrupt remapping table in the host IOMMU.
Therefore, send notification when guest IOMMU emulates the IRT
invalidation commands.
Reviewed-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Signed-off-by: Santosh Shukla <santosh.shukla@amd.com>
Message-Id: <20240927172913.121477-5-santosh.shukla@amd.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Use shared memory region for interrupt remapping which can be
aliased by all devices.
Reviewed-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Signed-off-by: Santosh Shukla <santosh.shukla@amd.com>
Message-Id: <20240927172913.121477-4-santosh.shukla@amd.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Introduce 'nodma' shared memory region to support PT mode
so that for each device, we only create an alias to shared memory
region when DMA-remapping is disabled.
Reviewed-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Signed-off-by: Santosh Shukla <santosh.shukla@amd.com>
Message-Id: <20240927172913.121477-3-santosh.shukla@amd.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Rename the MMIO memory region variable 'mmio' to 'mr_mmio'
so to correctly name align with struct AMDVIState::variable type.
No functional change intended.
Reviewed-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Signed-off-by: Santosh Shukla <santosh.shukla@amd.com>
Message-Id: <20240927172913.121477-2-santosh.shukla@amd.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Note: since all we did is replace VarPackageOp with PackageOP,
and both are represented by Package() in ASL, the AML is
different but ASL is the same.
Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
Message-Id: <20240924132417.739809-4-ribalda@chromium.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
Message-Id: <20240924132417.739809-2-ribalda@chromium.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Igor Mammedov <imammedo@redhat.com>
