tcg: Re-order tcg_region_init vs tcg_prologue_init
Instead of delaying tcg_region_init until after tcg_prologue_init
is complete, do tcg_region_init first and let tcg_prologue_init
shrink the first region by the size of the generated prologue.
Peter Maydell [Fri, 11 Jun 2021 08:21:48 +0000 (09:21 +0100)]
Merge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into staging
# gpg: Signature made Fri 11 Jun 2021 03:54:51 BST
# gpg: using RSA key EF04965B398D6211
# gpg: Good signature from "Jason Wang (Jason Wang on RedHat) <[email protected]>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg: It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 215D 46F4 8246 689E C77F 3562 EF04 965B 398D 6211
* remotes/jasowang/tags/net-pull-request:
Fixed calculation error of pkt->header_size in fill_pkt_tcp_info()
Add the function of colo_compare_cleanup
Add a function named packet_new_nocopy for COLO.
Remove migrate_set_block_enabled in checkpoint
Optimize the function of filter_send
Fix the qemu crash when guest shutdown during checkpoint
Remove some duplicate trace code.
netdev: add more commands to preconfig mode
vhost-vdpa: remove the unused vhost_vdpa_get_acked_features()
vhost-vdpa: don't initialize backend_features
vhost-vdpa: map virtqueue notification area if possible
vhost-vdpa: skip ram device from the IOTLB mapping
Rao, Lei [Tue, 8 Jun 2021 08:23:31 +0000 (16:23 +0800)]
Fixed calculation error of pkt->header_size in fill_pkt_tcp_info()
The data pointer has skipped vnet_hdr_len in the function of
parse_packet_early().So, we can not subtract vnet_hdr_len again
when calculating pkt->header_size in fill_pkt_tcp_info(). Otherwise,
it will cause network packet comparsion errors and greatly increase
the frequency of checkpoints.
Rao, Lei [Tue, 8 Jun 2021 08:23:30 +0000 (16:23 +0800)]
Add the function of colo_compare_cleanup
This patch fixes the following:
#0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
#1 0x00007f6ae4559859 in __GI_abort () at abort.c:79
#2 0x0000559aaa386720 in error_exit (err=16, msg=0x559aaa5973d0 <__func__.16227> "qemu_mutex_destroy") at util/qemu-thread-posix.c:36
#3 0x0000559aaa3868c5 in qemu_mutex_destroy (mutex=0x559aabffe828) at util/qemu-thread-posix.c:69
#4 0x0000559aaa2f93a8 in char_finalize (obj=0x559aabffe800) at chardev/char.c:285
#5 0x0000559aaa23318a in object_deinit (obj=0x559aabffe800, type=0x559aabfd7d20) at qom/object.c:606
#6 0x0000559aaa2331b8 in object_deinit (obj=0x559aabffe800, type=0x559aabfd9060) at qom/object.c:610
#7 0x0000559aaa233200 in object_finalize (data=0x559aabffe800) at qom/object.c:620
#8 0x0000559aaa234202 in object_unref (obj=0x559aabffe800) at qom/object.c:1074
#9 0x0000559aaa2356b6 in object_finalize_child_property (obj=0x559aac0dac10, name=0x559aac778760 "compare0-0", opaque=0x559aabffe800) at qom/object.c:1584
#10 0x0000559aaa232f70 in object_property_del_all (obj=0x559aac0dac10) at qom/object.c:557
#11 0x0000559aaa2331ed in object_finalize (data=0x559aac0dac10) at qom/object.c:619
#12 0x0000559aaa234202 in object_unref (obj=0x559aac0dac10) at qom/object.c:1074
#13 0x0000559aaa2356b6 in object_finalize_child_property (obj=0x559aac0c75c0, name=0x559aac0dadc0 "chardevs", opaque=0x559aac0dac10) at qom/object.c:1584
#14 0x0000559aaa233071 in object_property_del_child (obj=0x559aac0c75c0, child=0x559aac0dac10, errp=0x0) at qom/object.c:580
#15 0x0000559aaa233155 in object_unparent (obj=0x559aac0dac10) at qom/object.c:599
#16 0x0000559aaa2fb721 in qemu_chr_cleanup () at chardev/char.c:1159
#17 0x0000559aa9f9b110 in main (argc=54, argv=0x7ffeb62fa998, envp=0x7ffeb62fab50) at vl.c:4539
When chardev is cleaned up, chr_write_lock needs to be destroyed. But
the colo-compare module is not cleaned up normally before it when the
guest poweroff. It is holding chr_write_lock at this time. This will
cause qemu crash.So we add the function of colo_compare_cleanup() before
qemu_chr_cleanup() to fix the bug.
Rao, Lei [Tue, 8 Jun 2021 08:23:29 +0000 (16:23 +0800)]
Add a function named packet_new_nocopy for COLO.
Use the packet_new_nocopy instead of packet_new in the
filter-rewriter module. There will be one less memory
copy in the processing of each network packet.
Rao, Lei [Tue, 8 Jun 2021 08:23:28 +0000 (16:23 +0800)]
Remove migrate_set_block_enabled in checkpoint
We can detect disk migration in migrate_prepare, if disk migration
is enabled in COLO mode, we can directly report an error.and there
is no need to disable block migration at every checkpoint.
Rao, Lei [Tue, 8 Jun 2021 08:23:27 +0000 (16:23 +0800)]
Optimize the function of filter_send
The iov_size has been calculated in filter_send(). we can directly
return the size.In this way, this is no need to repeat calculations
in filter_redirector_receive_iov();
Paolo Bonzini [Tue, 11 May 2021 15:39:55 +0000 (11:39 -0400)]
netdev: add more commands to preconfig mode
Creating and destroying network backend does not require a fully
constructed machine. Allow the related monitor commands to run before
machine initialization has concluded.
Jason Wang [Wed, 2 Jun 2021 03:31:26 +0000 (11:31 +0800)]
vhost-vdpa: don't initialize backend_features
We used to initialize backend_features during vhost_vdpa_init()
regardless whether or not it was supported by vhost. This will lead
the unsupported features like VIRTIO_F_IN_ORDER to be included and set
to the vhost-vdpa during vhost_dev_start. Because the
VIRTIO_F_IN_ORDER is not supported by vhost-vdpa so it won't be
advertised to guest which will break the datapath.
Fix this by not initializing the backend_features, so the
acked_features could be built only from guest features via
vhost_net_ack_features().
Jason Wang [Thu, 15 Apr 2021 07:33:56 +0000 (15:33 +0800)]
vhost-vdpa: map virtqueue notification area if possible
This patch implements the vq notification mapping support for
vhost-vDPA. This is simply done by using mmap()/munmap() for the
vhost-vDPA fd during device start/stop. For the device without
notification mapping support, we fall back to eventfd based
notification gracefully.
Jason Wang [Wed, 2 Jun 2021 08:41:05 +0000 (16:41 +0800)]
vhost-vdpa: skip ram device from the IOTLB mapping
vDPA is not tie to any specific hardware, for safety and simplicity,
vhost-vDPA doesn't allow MMIO area to be mapped via IOTLB. Only the
doorbell could be mapped via mmap(). So this patch exclude skip the
ram device from the IOTLB mapping.
Peter Maydell [Wed, 9 Jun 2021 15:40:21 +0000 (16:40 +0100)]
Merge remote-tracking branch 'remotes/dgilbert-gitlab/tags/pull-migration-20210609a' into staging
Migration pull for 2021-06-09
Yank crash fix from Leo
RDMA fix from Li
mptcp support from me
dirty-rate changes from Hyman and Peter
(Note I've switched to the gitlab I've been using for virtiofs pulls)
Signed-off-by: Dr. David Alan Gilbert <[email protected]>
# gpg: Signature made Wed 09 Jun 2021 15:40:01 BST
# gpg: using RSA key 45F5C71B4A0CB7FB977A9FA90516331EBC5BFDE7
# gpg: Good signature from "Dr. David Alan Gilbert (RH2) <[email protected]>" [full]
# Primary key fingerprint: 45F5 C71B 4A0C B7FB 977A 9FA9 0516 331E BC5B FDE7
* remotes/dgilbert-gitlab/tags/pull-migration-20210609a:
hmp: Add "calc_dirty_rate" and "info dirty_rate" cmds
migration/dirtyrate: make sample page count configurable
sockets: Support multipath TCP
migration/socket: Close the listener at the end
migration: Add cleanup hook for inwards migration
io/net-listener: Call the notifier during finalize
channel-socket: Only set CLOEXEC if we have space for fds
migration/rdma: Fix cm event use after free
yank: Unregister function when using TLS migration
Hyman Huang(黄勇) [Mon, 7 Jun 2021 01:11:34 +0000 (09:11 +0800)]
migration/dirtyrate: make sample page count configurable
introduce optional sample-pages argument in calc-dirty-rate,
making sample page count per GB configurable so that more
accurate dirtyrate can be calculated.
channel-socket: Only set CLOEXEC if we have space for fds
MSG_CMSG_CLOEXEC cleans up received fd's; it's really only for Unix
sockets, but currently we enable it for everything; some socket types
(IP_MPTCP) don't like this.
Only enable it when we're giving the recvmsg room to receive fd's
anyway.
Leonardo Bras [Tue, 1 Jun 2021 05:40:31 +0000 (02:40 -0300)]
yank: Unregister function when using TLS migration
After yank feature was introduced in migration, whenever migration
is started using TLS, the following error happens in both source and
destination hosts:
This happens because of a missing yank_unregister_function() when using
qio-channel-tls.
Fix this by also allowing TYPE_QIO_CHANNEL_TLS object type to perform
yank_unregister_function() in channel_close() and multifd_load_cleanup().
Also, inside migration_channel_connect() and
migration_channel_process_incoming() move yank_register_function() so
it only runs once on a TLS migration.
Fixes: b5eea99ec2f ("migration: Add yank feature", 2021-01-13) Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1964326 Signed-off-by: Leonardo Bras <[email protected]> Reviewed-by: Lukas Straub <[email protected]> Reviewed-by: Peter Xu <[email protected]>
--
Changes since v2:
- Dropped all references to ioc->master
- yank_register_function() and yank_unregister_function() now only run
once in a TLS migration.
Peter Maydell [Tue, 8 Jun 2021 12:54:23 +0000 (13:54 +0100)]
Merge remote-tracking branch 'remotes/alistair/tags/pull-riscv-to-apply-20210608-1' into staging
Second RISC-V PR for QEMU 6.1
- Update the PLIC and CLINT DT bindings
- Improve documentation for RISC-V machines
- Support direct kernel boot for microchip_pfsoc
- Fix WFI exception behaviour
- Improve CSR printing
- Initial support for the experimental Bit Manip extension
# gpg: Signature made Tue 08 Jun 2021 01:28:27 BST
# gpg: using RSA key F6C4AC46D4934868D3B8CE8F21E10D29DF977054
# gpg: Good signature from "Alistair Francis <[email protected]>" [full]
# Primary key fingerprint: F6C4 AC46 D493 4868 D3B8 CE8F 21E1 0D29 DF97 7054
* remotes/alistair/tags/pull-riscv-to-apply-20210608-1: (32 commits)
target/riscv: rvb: add b-ext version cpu option
target/riscv: rvb: support and turn on B-extension from command line
target/riscv: rvb: add/shift with prefix zero-extend
target/riscv: rvb: address calculation
target/riscv: rvb: generalized or-combine
target/riscv: rvb: generalized reverse
target/riscv: rvb: rotate (left/right)
target/riscv: rvb: shift ones
target/riscv: rvb: single-bit instructions
target/riscv: add gen_shifti() and gen_shiftiw() helper functions
target/riscv: rvb: sign-extend instructions
target/riscv: rvb: min/max instructions
target/riscv: rvb: pack two words into one register
target/riscv: rvb: logic-with-negate
target/riscv: rvb: count bits set
target/riscv: rvb: count leading/trailing zeros
target/riscv: reformat @sh format encoding for B-extension
target/riscv: Pass the same value to oprsz and maxsz.
target/riscv/pmp: Add assert for ePMP operations
target/riscv: Dump CSR mscratch/sscratch/satp
...
Alistair Francis [Thu, 20 May 2021 22:55:53 +0000 (06:55 +0800)]
target/riscv/pmp: Add assert for ePMP operations
Although we construct epmp_operation in such a way that it can only be
between 0 and 15 Coverity complains that we don't handle the other
possible cases. To fix Coverity and make it easier for humans to read
add a default case to the switch statement that calls
g_assert_not_reached().
docs/system: Move the RISC-V -bios information to removed
QEMU 5.1 changed the behaviour of the default boot for the RISC-V virt
and sifive_u machines. This patch moves that change from the
deprecated.rst file to the removed-features.rst file and the
target-riscv.rst.
Jose Martins [Tue, 20 Apr 2021 21:36:56 +0000 (22:36 +0100)]
target/riscv: fix wfi exception behavior
The wfi exception trigger behavior should take into account user mode,
hstatus.vtw, and the fact the an wfi might raise different types of
exceptions depending on various factors:
If supervisor mode is not present:
- an illegal instruction exception should be generated if user mode
executes and wfi instruction and mstatus.tw = 1.
If supervisor mode is present:
- when a wfi instruction is executed, an illegal exception should be triggered
if either the current mode is user or the mode is supervisor and mstatus.tw is
set.
Plus, if the hypervisor extensions are enabled:
- a virtual instruction exception should be raised when a wfi is executed from
virtual-user or virtual-supervisor and hstatus.vtw is set.
Bin Meng [Fri, 30 Apr 2021 07:13:02 +0000 (15:13 +0800)]
hw/riscv: microchip_pfsoc: Support direct kernel boot
At present the Microchip Icicle Kit machine only supports using
'-bios' to load the HSS, and does not support '-kernel' for direct
kernel booting just like other RISC-V machines do. One has to use
U-Boot which is chain-loaded by HSS, to load a kernel for testing.
This is not so convenient.
Adding '-kernel' support together with the existing '-bios', we
follow the following table to select which payload we execute:
-bios | -kernel | payload
------+------------+--------
N | N | HSS
Y | don't care | HSS
N | Y | kernel
This ensures backwards compatibility with how we used to expose
'-bios' to users. When '-kernel' is used for direct boot, '-dtb'
must be present to provide a valid device tree for the board,
as we don't generate device tree.
When direct kernel boot is used, the OpenSBI fw_dynamic BIOS image
is used to boot a payload like U-Boot or OS kernel directly.
Documentation is updated to describe the direct kernel boot. Note
as of today there is still no PolarFire SoC support in the upstream
Linux kernel hence the document does not include instructions for
that. It will be updated in the future.
Bin Meng [Fri, 30 Apr 2021 07:12:58 +0000 (15:12 +0800)]
hw/riscv: Support the official PLIC DT bindings
The official DT bindings of PLIC uses "sifive,plic-1.0.0" as the
compatible string in the upstream Linux kernel. "riscv,plic0" is
now legacy and has to be kept for backward compatibility of legacy
systems.
Bin Meng [Fri, 30 Apr 2021 07:12:57 +0000 (15:12 +0800)]
hw/riscv: Support the official CLINT DT bindings
Linux kernel commit a2770b57d083 ("dt-bindings: timer: Add CLINT bindings")
adds the official DT bindings for CLINT, which uses "sifive,clint0"
as the compatible string. "riscv,clint0" is now legacy and has to
be kept for backward compatibility of legacy systems.
Bin Meng [Fri, 30 Apr 2021 07:12:56 +0000 (15:12 +0800)]
hw/riscv: virt: Switch to use qemu_fdt_setprop_string_array() helper
Since commit 78da6a1bca22 ("device_tree: add qemu_fdt_setprop_string_array helper"),
we can use the new helper to set the compatible strings for the
SiFive test device node.
Bin Meng [Fri, 30 Apr 2021 07:12:55 +0000 (15:12 +0800)]
hw/riscv: sifive_u: Switch to use qemu_fdt_setprop_string_array() helper
Since commit 78da6a1bca22 ("device_tree: add qemu_fdt_setprop_string_array helper"),
we can use the new helper to set the clock name for the ethernet
controller node.
* remotes/vivier2/tags/trivial-branch-for-6.1-pull-request:
vhost-vdpa: Remove redundant declaration of address_space_memory
scripts/oss-fuzz: Fix typo in documentation
target/mips: Fix 'Uncoditional' typo
target/hppa: Remove unused 'memory.h' header
hw/display/macfb: Classify the "nubus-macfb" as display device
target/nios2: fix page-fit instruction count
docs: fix broken reference
linux-user/syscall: Constify bitmask_transtbl fcntl/mmap flags_tlb[]
misc: Correct relative include path
i386/kvm: The value passed to strerror should be positive
target/riscv: Do not include 'pmp.h' in user emulation
Peter Maydell [Mon, 7 Jun 2021 14:45:48 +0000 (15:45 +0100)]
Merge remote-tracking branch 'remotes/stsquad/tags/pull-testing-updates-070621-2' into staging
A few testing and configure updates:
- add the multiarch signals stress test
- fix display of multi-word compiler stanzas in meson
- fix quoting of multi-word compiler stazas in configure.sh
- tag some acceptance tests as TCG only
- make checkpatch test work harder to find clean diffs
- split gprof/gconv job to avoid timeouts
- fix centos8 VM build by adding --source-path
- make checkpatch aware of .h.inc and .c.inc paths
# gpg: Signature made Mon 07 Jun 2021 14:51:12 BST
# gpg: using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <[email protected]>" [full]
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8 DF35 FBD0 DB09 5A9E 2A44
* remotes/stsquad/tags/pull-testing-updates-070621-2:
scripts/checkpatch.pl: process .c.inc and .h.inc files as C source
tests/vm: expose --source-path to scripts to find extra files
gitlab-ci: Split gprof-gcov job
gitlab: work harder to avoid false positives in checkpatch
tests/acceptance: tag various arm tests as TCG only
tests/tcg/configure.sh: tweak quoting of target_compiler
meson.build: fix cosmetics of compiler display
tests/tcg: add a multiarch signals test to stress test signal delivery
Alex Bennée [Thu, 27 May 2021 16:03:16 +0000 (17:03 +0100)]
tests/tcg/configure.sh: tweak quoting of target_compiler
If you configure the host compiler with a multi-command stanza like:
--cc="ccache gcc"
then the configure.sh machinery falls over with confusion. Work around
this by ensuring we correctly quote so where we need a complete
evaluation we get it. Of course the has() check needs single variable
so we need to unquote that. This does mean it essentially checks that
just the ccache command exits but if we got past that step we still
check the compiler actually does something.
Alex Bennée [Thu, 27 May 2021 16:03:15 +0000 (17:03 +0100)]
meson.build: fix cosmetics of compiler display
If you specify something like --cc="ccache gcc" on your configure line
the summary output misses the rest of the cmd_array. Do some string
joining to make it complete.
Alex Bennée [Thu, 27 May 2021 16:03:14 +0000 (17:03 +0100)]
tests/tcg: add a multiarch signals test to stress test signal delivery
This adds a simple signal test that combines the POSIX timer_create
with signal delivery across multiple threads. The aim is to provide a
bit more of a stress test to flush out signal handling issues for
easily than the occasional random crash we sometimes see in linux-test
or threadcount.
Xie Yongji [Mon, 17 May 2021 12:32:46 +0000 (20:32 +0800)]
vhost-vdpa: Remove redundant declaration of address_space_memory
The symbol address_space_memory are already declared in
include/exec/address-spaces.h. So let's add this header file
and remove the redundant declaration in include/hw/virtio/vhost-vdpa.h.
Pavel Dovgalyuk [Tue, 11 May 2021 08:40:10 +0000 (11:40 +0300)]
target/nios2: fix page-fit instruction count
This patch fixes calculation of number of the instructions
that fit the current page. It prevents creation of the translation
blocks that cross the page boundaries. It is required for deterministic
exception generation in icount mode.
which does not reference the desired entity, and leaves some extra junk
in the rendered output. See
https://qemu-project.gitlab.io/qemu/devel/memory.html#c.MemoryListener
member log_start for an example of the broken output as it looks today.
The three vector shift by vector operations are all implemented via
expansion. Therefore do not actually set TCG_TARGET_HAS_shv_vec,
as none of shlv_vec, shrv_vec, sarv_vec may actually appear in the
instruction stream, and therefore also do not appear in tcg_target_op_def.
Change the return value to bool, because that's what is should
have been from the start. Pass the ct mask instead of the whole
TCGArgConstraint, as that's the only part that's relevant.
Change the value argument to int64_t. We will need the extra
width for 32-bit hosts wanting to match vector constants.
Peter Maydell [Fri, 4 Jun 2021 16:27:29 +0000 (17:27 +0100)]
Merge remote-tracking branch 'remotes/bonzini-gitlab/tags/for-upstream' into staging
* OpenBSD cleanup (Brad)
* fixes for the i386 accel/cpu refactoring (Claudio)
* unmap test for emulated SCSI (Kit)
* fix for iscsi module (myself)
* fix for -readconfig of objects (myself)
* fixes for x86 16-bit task switching (myself)
* fix for x86 MOV from/to CR8 (Richard)
* remotes/bonzini-gitlab/tags/for-upstream:
vl: plug -object back into -readconfig
vl: plumb keyval-based options into -readconfig
qemu-config: parse configuration files to a QDict
i386: run accel_cpu_instance_init as post_init
i386: reorder call to cpu_exec_realizefn
tests/qtest/virtio-scsi-test: add unmap large LBA with 4k blocks test
target/i386: Fix decode of cr8
target/i386: tcg: fix switching from 16-bit to 32-bit tasks or vice versa
target/i386: tcg: fix loading of registers from 16-bit TSS
target/i386: tcg: fix segment register offsets for 16-bit TSS
oslib-posix: Remove OpenBSD workaround for fcntl("/dev/null", F_SETFL, O_NONBLOCK) failure
iscsi: link libm into the module
meson: allow optional dependencies for block modules
Peter Maydell [Fri, 4 Jun 2021 12:38:48 +0000 (13:38 +0100)]
Merge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into staging
# gpg: Signature made Fri 04 Jun 2021 08:26:16 BST
# gpg: using RSA key EF04965B398D6211
# gpg: Good signature from "Jason Wang (Jason Wang on RedHat) <[email protected]>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg: It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 215D 46F4 8246 689E C77F 3562 EF04 965B 398D 6211
Paolo Bonzini [Mon, 24 May 2021 10:57:52 +0000 (06:57 -0400)]
vl: plug -object back into -readconfig
Commit bc2f4fcb1d ("qom: move user_creatable_add_opts logic to vl.c
and QAPIfy it", 2021-03-19) switched the creation of objects from
qemu_opts_foreach to a bespoke QTAILQ in preparation for supporting JSON
syntax in -object.
Unfortunately in doing so it lost support for [object] stanzas in
configuration files and also for "-set object.ID.KEY=VAL". The latter
is hard to re-establish and probably best solved by deprecating -set.
This patch uses the infrastructure introduced by the previous two
patches in order to parse QOM objects correctly from configuration
files.
projects / qemu.git / log