It can be difficult to debug issues with BHs in production environments.
Although BHs can usually be identified by looking up their ->cb()
function pointer, this requires debug information for the program. It is
also not possible to print human-readable diagnostics about BHs because
they have no identifier.
This patch adds a name to each BH. The name is not unique per instance
but differentiates between cb() functions, which is usually enough. It's
done by changing aio_bh_new() and friends to macros that stringify cb.
The next patch will use the name field when reporting leaked BHs.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210414200247.917496-2-stefanha@redhat.com>
Since the prior calls are successful, in this case a errno doesn't
indicate a real error which would just make us confused.
before:
(qemu) migrate -d rdma:192.168.22.23:8888
source_resolve_host RDMA Device opened: kernel name rxe_eth0 uverbs device name uverbs2, infiniband_verbs class device path /sys/class/infiniband_verbs/uverbs2, infiniband class device path /sys/class/infiniband/rxe_eth0, transport: (2) Ethernet
rdma_get_cm_event != EVENT_ESTABLISHED after rdma_connect: No space left on device
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Message-Id: <20210628071959.23455-1-lizhijian@cn.fujitsu.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
migrate-set-parameters parse "downtime_limit" as integer type when
execute "migrate-set-parameters" before migration, and, the unit
dowtime_limit is milliseconds, fix this two so that test can go
smoothly.
Signed-off-by: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>
Message-Id: <31d82df24cc0c468dbe4d2d86730158ebf248071.1622729934.git.huangy81@chinatelecom.cn>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
thread_id in CpuInfoFast is deprecated, parse thread-id instead
after execute qmp query-cpus-fast. fix this so that test can
go smoothly.
Signed-off-by: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>
Message-Id: <584578c0a0dd781cee45f72ddf517f6e6a41c504.1622729934.git.huangy81@chinatelecom.cn>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
fuse has an option FUSE_POSIX_ACL which needs to be opted in by fuse
server to enable posix acls. As of now we are not opting in for this,
so posix acls are disabled on virtiofs by default.
Add virtiofsd option "-o posix_acl/no_posix_acl" to let users enable/disable
posix acl support. By default it is disabled as of now due to performance
concerns with cache=none.
Currently even if file server has not opted in for FUSE_POSIX_ACL, user can
still query acl and set acl, and system.posix_acl_access and
system.posix_acl_default xattrs show up listxattr response.
Miklos said this is confusing. So he said lets block and filter
system.posix_acl_access and system.posix_acl_default xattrs in
getxattr/setxattr/listxattr if user has explicitly disabled
posix acls using -o no_posix_acl.
As of now continuing to keeping the existing behavior if user did not
specify any option to disable acl support due to concerns about backward
compatibility.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210622150852.1507204-8-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
When posix access acls are set on a file, it can lead to adjusting file
permissions (mode) as well. If caller does not have CAP_FSETID and it
also does not have membership of owner group, this will lead to clearing
SGID bit in mode.
Current fuse code is written in such a way that it expects file server
to take care of chaning file mode (permission), if there is a need.
Right now, host kernel does not clear SGID bit because virtiofsd is
running as root and has CAP_FSETID. For host kernel to clear SGID,
virtiofsd need to switch to gid of caller in guest and also drop
CAP_FSETID (if caller did not have it to begin with).
If SGID needs to be cleared, client will set the flag
FUSE_SETXATTR_ACL_KILL_SGID in setxattr request. In that case server
should kill sgid.
Currently just switch to uid/gid of the caller and drop CAP_FSETID
and that should do it.
This should fix the xfstest generic/375 test case.
We don't have to switch uid for this to work. That could be one optimization
that pass a parameter to lo_change_cred() to only switch gid and not uid.
Also this will not work whenever (if ever) we support idmapped mounts. In
that case it is possible that uid/gid in request are 0/0 but still we
need to clear SGID. So we will have to pick a non-root sgid and switch
to that instead. That's an TODO item for future when idmapped mount
support is introduced.
This patch only adds the capability to switch creds and drop FSETID
when acl xattr is set. This does not take affect yet. It can take
affect when next patch adds the capability to enable posix_acl.
Reported-by: Luis Henriques <lhenriques@suse.de>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210622150852.1507204-7-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
When parent directory has default acl and a file is created in that
directory, then umask is ignored and final file permissions are
determined using default acl instead. (man 2 umask).
Currently, fuse applies the umask and sends modified mode in create
request accordingly. fuse server can set FUSE_DONT_MASK and tell
fuse client to not apply umask and fuse server will take care of
it as needed.
With posix acls enabled, requirement will be that we want umask
to determine final file mode if parent directory does not have
default acl.
So if posix acls are enabled, opt in for FUSE_DONT_MASK. virtiofsd
will set umask of the thread doing file creation. And host kernel
should use that umask if parent directory does not have default
acls, otherwise umask does not take affect.
Miklos mentioned that we already call unshare(CLONE_FS) for
every thread. That means umask has now become property of per
thread and it should be ok to manipulate it in file creation path.
This patch only adds capability to change umask and restore it. It
does not enable it yet. Next few patches will add capability to enable it
based on if user enabled posix_acl or not.
This should fix fstest generic/099.
Reported-by: Luis Henriques <lhenriques@suse.de>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210622150852.1507204-6-vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Patches in this series are going to make use of "umask" syscall.
So allow it.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20210622150852.1507204-5-vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Add the bits to enable support for setxattr_ext if fuse offers it. Do not
enable it by default yet. Let passthrough_ll opt-in. Enabling it by deafult
kind of automatically means that you are taking responsibility of clearing
SGID if ACL is set.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210622150852.1507204-4-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Fixed up double def in fuse_common.h
getxattr/setxattr/removexattr/listxattr operations handle regualar
and non-regular files differently. For the case of non-regular files
we do fchdir(/proc/self/fd) and the xattr operation and then revert
back to original working directory. After this we are saving errno
and that's buggy because fchdir() will overwrite the errno.
FCHDIR_NOFAIL(lo->proc_self_fd);
ret = getxattr(procname, name, value, size);
FCHDIR_NOFAIL(lo->root.fd);
if (ret == -1)
saverr = errno
In above example, if getxattr() failed, we will still return 0 to caller
as errno must have been written by FCHDIR_NOFAIL(lo->root.fd) call.
Fix all such instances and capture "errno" early and save in "saverr"
variable.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210622150852.1507204-3-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
With kernel header updates fuse_setxattr_in struct has grown in size.
But this new struct size only takes affect if user has opted in
for fuse feature FUSE_SETXATTR_EXT otherwise fuse continues to
send "fuse_setxattr_in" of older size. Older size is determined
by FUSE_COMPAT_SETXATTR_IN_SIZE.
Fix this. If we have not opted in for FUSE_SETXATTR_EXT, then
expect that we will get fuse_setxattr_in of size FUSE_COMPAT_SETXATTR_IN_SIZE
and not sizeof(struct fuse_sexattr_in).
Fixes: 278f064e45 ("Update Linux headers to 5.13-rc4")
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210622150852.1507204-2-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
A well behaved FUSE client uses FUSE_CREATE to create files. It isn't
supposed to pass O_CREAT along a FUSE_OPEN request, as documented in
the "fuse_lowlevel.h" header :
/**
* Open a file
*
* Open flags are available in fi->flags. The following rules
* apply.
*
* - Creation (O_CREAT, O_EXCL, O_NOCTTY) flags will be
* filtered out / handled by the kernel.
But if the client happens to do it anyway, the server ends up passing
this flag to open() without the mandatory mode_t 4th argument. Since
open() is a variadic function, glibc will happily pass whatever it
finds on the stack to the syscall. If this file is compiled with
-D_FORTIFY_SOURCE=2, glibc will even detect that and abort:
*** invalid openat64 call: O_CREAT or O_TMPFILE without mode ***: terminated
Specifying O_CREAT with FUSE_OPEN is a protocol violation. Check this
in do_open(), print out a message and return an error to the client,
EINVAL like we already do when fuse_mbuf_iter_advance() fails.
The FUSE filesystem doesn't currently support O_TMPFILE, but the very
same would happen if O_TMPFILE was passed in a FUSE_OPEN request. Check
that as well.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <20210624101809.48032-1-groug@kaod.org>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Different guest xattr prefixes have distinct access control rules applied
by the guest. When remapping a guest xattr care must be taken that the
remapping does not allow the a guest user to bypass guest kernel access
control rules.
For example if 'trusted.*' which requires CAP_SYS_ADMIN is remapped
to 'user.virtiofs.trusted.*', an unprivileged guest user which can
write to 'user.*' can bypass the CAP_SYS_ADMIN control. Thus the
target of any remapping must be explicitly blocked from read/writes
by the guest, to prevent access control bypass.
The examples shown in the virtiofsd man page already do the right
thing and ensure safety, but the security implications of getting
this wrong were not made explicit. This could lead to host admins
and apps unwittingly creating insecure configurations.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20210611120427.49736-1-berrange@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
The GDateTime APIs provided by GLib avoid portability pitfalls, such
as some platforms where 'struct timeval.tv_sec' field is still 'long'
instead of 'time_t'. When combined with automatic cleanup, GDateTime
often results in simpler code too.
Localtime is changed to UTC to avoid the need to grant extra seccomp
permissions for GLib's access of the timezone database.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210611164319.67762-1-berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
If the user cancels the migration in the unplug-wait state,
QEMU will try to plug back the card and this fails because the card
is partially unplugged.
To avoid the problem, continue to wait the card unplug, but to
allow the migration to be canceled if the card never finishes to unplug
use a timeout.
Bug: https://bugzilla.redhat.com/show_bug.cgi?id=1976852
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210629155007.629086-3-lvivier@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
The loop is used in migration_thread() and bg_migration_thread(),
so we can move it to its own function and call it from these both places.
Moreover, in migration_thread() we have a wrong state transition from
SETUP to ACTIVE while state could be WAIT_UNPLUG. This is correctly
managed in bg_migration_thread() so use this code instead.
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20210629155007.629086-2-lvivier@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
It's possible qemu_start_incoming_migration() failed at any point, when it
happens we should reset postcopy_recover_triggered to false so that the user
can still retry with a saner incoming port.
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20210629181356.217312-3-peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Starting from commit b5eea99ec2, qmp_migrate_recover() calls unregister
before calling qemu_start_incoming_migration(). I believe it wanted to mitigate
the next call to yank_register_instance(), but I think that's wrong.
Firstly, if during recover, we should keep the yank instance there, not
"quickly removing and adding it back".
Meanwhile, calling qmp_migrate_recover() twice with b5eea99ec2 will directly
crash the dest qemu (right now it can't; but it'll start to work right after
the next patch) because the 1st call of qmp_migrate_recover() will unregister
permanently when the channel failed to establish, then the 2nd call of
qmp_migrate_recover() crashes at yank_unregister_instance().
This patch fixes it by moving yank ops out of qemu_start_incoming_migration()
into qmp_migrate_incoming. For qmp_migrate_recover(), drop the unregister of
yank instance too since we keep it there during the recovery phase.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210629181356.217312-2-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
When testing migration, a Segmentation fault qemu core is generated.
0 error_free (err=0x1)
1 0x00007f8b862df647 in qemu_fclose (f=f@entry=0x55e06c247640)
2 0x00007f8b8516d59a in migrate_fd_cleanup (s=s@entry=0x55e06c0e1ef0)
3 0x00007f8b8516d66c in migrate_fd_cleanup_bh (opaque=0x55e06c0e1ef0)
4 0x00007f8b8626a47f in aio_bh_poll (ctx=ctx@entry=0x55e06b5a16d0)
5 0x00007f8b8626e71f in aio_dispatch (ctx=0x55e06b5a16d0)
6 0x00007f8b8626a33d in aio_ctx_dispatch (source=<optimized out>, callback=<optimized out>, user_data=<optimized out>)
7 0x00007f8b866bdba4 in g_main_context_dispatch ()
8 0x00007f8b8626cde9 in glib_pollfds_poll ()
9 0x00007f8b8626ce62 in os_host_main_loop_wait (timeout=<optimized out>)
10 0x00007f8b8626cffd in main_loop_wait (nonblocking=nonblocking@entry=0)
11 0x00007f8b862ef01f in main_loop ()
Using gdb print the struct QEMUFile f = {
...,
iovcnt = 65, last_error = 21984,
last_error_obj = 0x1, shutdown = true
}
Well iovcnt is overflow, because the max size of MAX_IOV_SIZE is 64.
struct QEMUFile {
...;
struct iovec iov[MAX_IOV_SIZE];
unsigned int iovcnt;
int last_error;
Error *last_error_obj;
bool shutdown;
};
iovcnt and last_error is overwrited by add_to_iovec().
Right now, add_to_iovec() increase iovcnt before check the limit.
And it seems that add_to_iovec() assumes that iovcnt will set to zero
in qemu_fflush(). But qemu_fflush() will directly return when f->shutdown
is true.
The situation may occur when libvirtd restart during migration, after
f->shutdown is set, before calling qemu_file_set_error() in
qemu_file_shutdown().
So the safiest way is checking the iovcnt before increasing it.
Signed-off-by: Feng Lin <linfeng23@huawei.com>
Message-Id: <20210625062138.1899-1-linfeng23@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Fix typo in 'writeable' which is actually misnamed 'writable'
Add dirty ring test if kernel supports it. Add the dirty ring parameter on
source should be mostly enough, but let's change the dest too to make them
match always.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20210615175523.439830-3-peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Several CVE fixes for the PVRDMA device.
-----BEGIN PGP SIGNATURE-----
iQEcBAABAgAGBQJg4hJVAAoJEDbUwPDPL+RtGEkIAJs7yas9LdaP80scaEQ3UiOJ
Mr9M2+RwoUm19YhrM2MqjIByluB5Jk2bO1ZmpuJmeg03V15I8AV8MozKGAzZaThx
8LVwduqlmdnrprBYKVRRn/SLSqbMkRSXGli9BjX2PoF97VeqvhNOxH1hL6z8Ytn6
zzpsOSq7NmC87hLdvL2UIkpYy1Z4bUjAWr1wn7K6jIhM163pek61rmO9eWwc7W3y
M1lXtGtrLgE0AuLm0oEJwz0xqlOuN8ld5kwIEu33RnL8phLwT0hDjHoLPqIX79bS
UdIL1zNjwXKI0DkR5EoVLRMdF87b3LMfVEvij25PNLPf6gyyh5FHmdqQi0r3sZU=
=MrNO
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/marcel/tags/pvrdma-04-07-2021-v2' into staging
PVRDMA queue
Several CVE fixes for the PVRDMA device.
# gpg: Signature made Sun 04 Jul 2021 20:56:05 BST
# gpg: using RSA key 36D4C0F0CF2FE46D
# gpg: Good signature from "Marcel Apfelbaum <marcel.apfelbaum@zoho.com>" [marginal]
# gpg: aka "Marcel Apfelbaum <marcel@redhat.com>" [marginal]
# gpg: aka "Marcel Apfelbaum <marcel.apfelbaum@gmail.com>" [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: B1C6 3A57 F92E 08F2 640F 31F5 36D4 C0F0 CF2F E46D
* remotes/marcel/tags/pvrdma-04-07-2021-v2:
pvrdma: Fix the ring init error flow (CVE-2021-3608)
pvrdma: Ensure correct input on ring init (CVE-2021-3607)
hw/rdma: Fix possible mremap overflow in the pvrdma device (CVE-2021-3582)
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
- Extract nanoMIPS, microMIPS, Code Compaction from translate.c
- Allow PCI config accesses smaller than 32-bit on Bonito64 device
- Fix migration of g364fb device on Jazz Magnum
- Fix dp8393x PROM checksum on Jazz Magnum and Quadra 800
- Map the UART devices unconditionally on Jazz Magnum
- Add functional test booting Linux on the Fuloong 2E
-----BEGIN PGP SIGNATURE-----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=rR+B
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/philmd/tags/mips-20210702' into staging
MIPS patches queue
- Extract nanoMIPS, microMIPS, Code Compaction from translate.c
- Allow PCI config accesses smaller than 32-bit on Bonito64 device
- Fix migration of g364fb device on Jazz Magnum
- Fix dp8393x PROM checksum on Jazz Magnum and Quadra 800
- Map the UART devices unconditionally on Jazz Magnum
- Add functional test booting Linux on the Fuloong 2E
# gpg: Signature made Fri 02 Jul 2021 16:36:19 BST
# gpg: using RSA key FAABE75E12917221DCFD6BB2E3E32C2CDEADC0DE
# gpg: Good signature from "Philippe Mathieu-Daudé (F4BUG) <f4bug@amsat.org>" [full]
# Primary key fingerprint: FAAB E75E 1291 7221 DCFD 6BB2 E3E3 2C2C DEAD C0DE
* remotes/philmd/tags/mips-20210702:
hw/mips/jazz: Map the UART devices unconditionally
hw/mips/jazz: specify correct endian for dp8393x device
hw/m68k/q800: fix PROM checksum and MAC address storage
qemu/bitops.h: add bitrev8 implementation
dp8393x: remove onboard PROM containing MAC address and checksum
hw/m68k/q800: move PROM and checksum calculation from dp8393x device to board
hw/mips/jazz: move PROM and checksum calculation from dp8393x device to board
dp8393x: convert to trace-events
dp8393x: checkpatch fixes
g364fb: add VMStateDescription for G364SysBusState
g364fb: use RAM memory region for framebuffer
tests/acceptance: Test Linux on the Fuloong 2E machine
hw/pci-host/bonito: Allow PCI config accesses smaller than 32-bit
hw/pci-host/bonito: Trace PCI config accesses smaller than 32-bit
target/mips: Extract nanoMIPS ISA translation routines
target/mips: Extract the microMIPS ISA translation routines
target/mips: Extract Code Compaction ASE translation routines
target/mips: Add declarations for generic TCG helpers
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Without providing a specific slot, QEMU won't be able to create the
second additional PCIe root port with the following error:
$ qemu-system-x86_64 [...] -machine q35 \
> -device pcie-root-port,bus=pcie.0,id=rp1 \
> -device pcie-root-port,bus=pcie.0,id=rp2
qemu-system-x86_64: -device pcie-root-port,bus=pcie.0,id=rp2:
Can't add chassis slot, error -16
This is due to the fact they both try to use slot 0. Update the
documentation to specify a slot for each new PCIe root port.
Signed-off-by: Vincent Bernat <vincent@bernat.ch>
Message-Id: <20210614114357.1146725-1-vincent@bernat.ch>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reset requests should use SHUTDOWN_CAUSE_GUEST_RESET not
SHUTDOWN_CAUSE_GUEST_SHUTDOWN.
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-Id: <20210624110057.2398779-1-kraxel@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Commit [1] moved _SUN variable from only hot-pluggable to
all devices. This made linux kernel enumerate extra slots
that weren't present before. If extra slot happens to be
be enumerated first and there is a device in th same slot
but on other bridge, linux kernel will add -N suffix to
slot name of the later, thus changing NIC name compared to
QEMU 5.2. This in some case confuses systemd, if it is
using SLOT NIC naming scheme and interface name becomes
not the same as it was under QEMU-5.2.
Reproducer QEMU CLI:
-M pc-i440fx-5.2 -nodefaults \
-device pci-bridge,chassis_nr=1,id=pci.1,bus=pci.0,addr=0x3 \
-device virtio-net-pci,id=nic1,bus=pci.1,addr=0x1 \
-device virtio-net-pci,id=nic2,bus=pci.1,addr=0x2 \
-device virtio-net-pci,id=nic3,bus=pci.1,addr=0x3
with RHEL8 guest produces following results:
v5.2:
kernel: virtio_net virtio0 ens1: renamed from eth0
kernel: virtio_net virtio2 ens3: renamed from eth2
kernel: virtio_net virtio1 enp1s2: renamed from eth1
(slot 2 is assigned to empty bus 0 slot and virtio1
is assigned to 2-2 slot, and renaming falls back,
for some reason, to path based naming scheme)
v6.0:
kernel: virtio_net virtio0 ens1: renamed from eth0
kernel: virtio_net virtio2 ens3: renamed from eth2
systemd-udevd[299]: Error changing net interface name 'eth1' to 'ens3': File exists
systemd-udevd[299]: could not rename interface '3' from 'eth1' to 'ens3': File exists
(with commit [1] kernel assigns virtio2 to 3-2 slot
since bridge advertises _SUN=0x3 and kernel assigns
slot 3 to bridge. Still it manages to rename virtio2
correctly to ens3, however systemd gets confused with virtio1
where slot allocation exactly the same (2-2) as in 5.2 case
and tries to rename it to ens3 which is rightfully taken by
virtio2)
I'm not sure what breaks in systemd interface renaming (it probably
should be investigated), but on QEMU side we can safely revert
_SUN to 5.2 behavior (i.e. avoid cold-plugged bridges and non
hot-pluggable device classes), without breaking acpi-index, which uses
slot numbers but it doesn't have to use _SUN, it could use an arbitrary
variable name that has the same slot value).
It will help existing VMs to keep networking with non trivial
configs in working order since systemd will do its interface
renaming magic as it used to do.
1)
Fixes: b7f23f62e4 (pci: acpi: add _DSM method to PCI devices)
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20210624204229.998824-3-imammedo@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: John Sucaet <john.sucaet@ekinops.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20210624204229.998824-2-imammedo@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: John Sucaet <john.sucaet@ekinops.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
When the card is plugged back, reset the partially_hotplugged flag to false
Bug: https://bugzilla.redhat.com/show_bug.cgi?id=1787194
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20210629152937.619193-1-lvivier@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
At some point, after unplugging virtio-pci the virtio device may be unrealised,
but the memory regions may be present in flatview. So, it's a possible situation
when memory region's callbacks are called for "unplugged" device.
Previous two patches made sure this case does not cause QEMU to crash.
This patch adds check for "notify" memory region. Now reads will return "-1" if a virtio
device is not present on a virtio bus.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1938042
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1743098
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Message-Id: <20210609095843.141378-4-andrew@daynix.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Now, if virtio device is not present on virtio-bus - pci config callbacks
will not lead to possible crush. The read will return "-1" which should be
interpreted by a driver that pci device may be unplugged.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Message-Id: <20210609095843.141378-3-andrew@daynix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
During unplug the virtio device is unplugged from virtio-bus on pci. In some cases,
requests to virtio-pci mm may acquire during/after unplug. Added check that virtio
device is on the bus, for "common" memory region.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Message-Id: <20210609095843.141378-2-andrew@daynix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
libFuzzer triggered the following assertion:
cat << EOF | qemu-system-i386 -M pc-q35-5.0 \
-nographic -monitor none -serial none \
-qtest stdio -d guest_errors -trace pci\*
outl 0xcf8 0xf2000060
outl 0xcfc 0x8400056e
EOF
pci_cfg_write mch 00:0 @0x60 <- 0x8400056e
Aborted (core dumped)
This is because guest wrote MCH_HOST_BRIDGE_PCIEXBAR_LENGTH_RVD
(reserved value) to the PCIE XBAR register.
There is no indication on the datasheet about what occurs when
this value is written. Simply ignore it on QEMU (and report an
guest error):
pci_cfg_write mch 00:0 @0x60 <- 0x8400056e
Q35: Reserved PCIEXBAR LENGTH
pci_cfg_read mch 00:0 @0x0 -> 0x8086
pci_cfg_read mch 00:0 @0x0 -> 0x29c08086
...
Cc: qemu-stable@nongnu.org
Reported-by: Alexander Bulekov <alxndr@bu.edu>
BugLink: https://bugs.launchpad.net/qemu/+bug/1878641
Fixes: df2d8b3ed4 ("q35: Introduce q35 pc based chipset emulator")
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210526142438.281477-1-f4bug@amsat.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
When using the Magnum ARC firmware we can see accesses to the
UART1 being rejected, because the device is not mapped:
$ qemu-system-mips64el -M magnum -d guest_errors,unimp -bios NTPROM.RAW
Invalid access at addr 0x80007004, size 1, region '(null)', reason: rejected
Invalid access at addr 0x80007001, size 1, region '(null)', reason: rejected
Invalid access at addr 0x80007002, size 1, region '(null)', reason: rejected
Invalid access at addr 0x80007003, size 1, region '(null)', reason: rejected
Invalid access at addr 0x80007004, size 1, region '(null)', reason: rejected
Since both UARTs are present (soldered on the board) regardless
of whether there are character devices connected, map them
unconditionally.
(This code pre-dated commit 12051d82f0 which made it safe to pass
NULL in as a chardev to serial devices.)
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20210629053704.2584504-1-f4bug@amsat.org>
The MIPS magnum machines are available in both big endian (mips64) and little
endian (mips64el) configurations. Ensure that the dp893x big_endian property
is set accordingly using logic similar to that used for the MIPS malta
machines.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Tested-by: Finn Thain <fthain@linux-m68k.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210625065401.30170-11-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
The checksum used by MacOS to validate the PROM content is an exclusive-OR
rather than a sum over the corresponding bytes. In addition the MAC address
must be stored in bit-reversed format as indicated in comments in Linux's
macsonic.c.
With the PROM contents fixed MacOS starts to probe the device registers
when AppleTalk is enabled in the Control Panel.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Tested-by: Finn Thain <fthain@linux-m68k.org>
Message-Id: <20210625065401.30170-8-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
This will be required for an upcoming checksum calculation.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Tested-by: Finn Thain <fthain@linux-m68k.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210625065401.30170-7-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
According to the datasheet the dp8393x chipset does not contain any NVRAM capable
of storing a MAC address or checksum. Now that both the MIPS jazz and m68k q800
boards generate the PROM region and checksum themselves, remove the generated
PROM from the dp8393x device itself.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Tested-by: Finn Thain <fthain@linux-m68k.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210625065401.30170-6-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
This is in preparation for each board to have its own separate bit storage
format and checksum for storing the MAC address.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Tested-by: Finn Thain <fthain@linux-m68k.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210625065401.30170-5-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
This is in preparation for each board to have its own separate bit storage
format and checksum for storing the MAC address.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Tested-by: Finn Thain <fthain@linux-m68k.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210625065401.30170-4-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Finn Thain <fthain@linux-m68k.org>
Message-Id: <20210625065401.30170-3-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Also fix a simple comment typo of "constrainst" to "constraints".
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Finn Thain <fthain@linux-m68k.org>
Message-Id: <20210625065401.30170-2-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Currently when QEMU attempts to migrate the MIPS magnum machine it crashes due
to a mistake in the g364fb VMStateDescription configuration which expects a
G364SysBusState and not a G364State.
Resolve the issue by adding a new VMStateDescription for G364SysBusState and
embedding the existing vmstate_g364fb VMStateDescription inside it using
VMSTATE_STRUCT.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Fixes: 97a3f6ffbb ("g364fb: convert to qdev")
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210625163554.14879-3-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Since the migration stream is already broken, we can use this opportunity to
change the framebuffer so that it is migrated as a RAM memory region rather
than as an array of bytes.
In particular this helps the output of the analyze-migration.py tool which
no longer contains a huge array representing the framebuffer contents.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210625163554.14879-2-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Test the kernel from Lemote rescue image:
http://dev.lemote.com/files/resource/download/rescue/rescue-yl
Once downloaded, set the RESCUE_YL_PATH environment variable
to point to the downloaded image and test as:
$ RESCUE_YL_PATH=~/images/fuloong2e/rescue-yl \
AVOCADO_ALLOW_UNTRUSTED_CODE=1 \
avocado --show=app,console run tests/acceptance/machine_mips_fuloong2e.py
Fetching asset from tests/acceptance/machine_mips_fuloong2e.py:MipsFuloong2e.test_linux_kernel_isa_serial
(1/1) tests/acceptance/machine_mips_fuloong2e.py:MipsFuloong2e.test_linux_kernel_isa_serial:
console: Linux version 2.6.27.7lemote (root@debian) (gcc version 4.1.3 20080623 (prerelease) (Debian 4.1.2-23)) #6 Fri Dec 12 00:11:25 CST 2008
console: busclock=33000000, cpuclock=-2145008360,memsize=256,highmemsize=0
console: console [early0] enabled
console: CPU revision is: 00006302 (ICT Loongson-2)
PASS (0.16 s)
JOB TIME : 0.51 s
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Message-Id: <20210624202747.1433023-5-f4bug@amsat.org>
When running the official PMON firmware for the Fuloong 2E, we see
8-bit and 16-bit accesses to PCI config space:
$ qemu-system-mips64el -M fuloong2e -bios pmon_2e.bin \
-trace -trace bonito\* -trace pci_cfg\*
pci_cfg_write vt82c686b-pm 05:4 @0x90 <- 0xeee1
bonito_spciconf_small_access PCI config address is smaller then 32-bit, addr: 0x4d2, size: 2
pci_cfg_write vt82c686b-pm 05:4 @0xd2 <- 0x1
pci_cfg_write vt82c686b-pm 05:4 @0x4 <- 0x1
pci_cfg_write vt82c686b-isa 05:0 @0x4 <- 0x7
bonito_spciconf_small_access PCI config address is smaller then 32-bit, addr: 0x81, size: 1
pci_cfg_read vt82c686b-isa 05:0 @0x81 -> 0x0
bonito_spciconf_small_access PCI config address is smaller then 32-bit, addr: 0x81, size: 1
pci_cfg_write vt82c686b-isa 05:0 @0x81 <- 0x80
bonito_spciconf_small_access PCI config address is smaller then 32-bit, addr: 0x83, size: 1
pci_cfg_write vt82c686b-isa 05:0 @0x83 <- 0x89
bonito_spciconf_small_access PCI config address is smaller then 32-bit, addr: 0x85, size: 1
pci_cfg_write vt82c686b-isa 05:0 @0x85 <- 0x3
bonito_spciconf_small_access PCI config address is smaller then 32-bit, addr: 0x5a, size: 1
pci_cfg_write vt82c686b-isa 05:0 @0x5a <- 0x7
bonito_spciconf_small_access PCI config address is smaller then 32-bit, addr: 0x85, size: 1
pci_cfg_write vt82c686b-isa 05:0 @0x85 <- 0x1
Also this is what the Linux kernel does since it supports the Bonito
north bridge:
https://elixir.bootlin.com/linux/v2.6.15/source/arch/mips/pci/ops-bonito64.c#L85
So it seems safe to assume the datasheet is incomplete or outdated
regarding the address constraints.
This problem was exposed by commit 911629e6d3
("vt82c686: Fix SMBus IO base and configuration registers").
Reported-by: BALATON Zoltan <balaton@eik.bme.hu>
Suggested-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210624202747.1433023-4-f4bug@amsat.org>
Tested-by: BALATON Zoltan <balaton@eik.bme.hu>
The device model batching its ioeventfds in a single MR transaction is
an optimization. Clarify this in virtio-scsi, virtio-blk and generic
virtio code. Also clarify that the transaction must commit before
closing ioeventfds so that no one is tempted to merge the loops
in the start functions error path and in the stop functions.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <162125799728.1394228.339855768563326832.stgit@bahia.lan>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
