tests/vm/openbsd: Use the system dtc package

We can use the pre-packaged libfdt from the dtc package to avoid
that we have to compile this code each time again and again.

While we're at it, the "--python=python3" does not seemt to be
necessary anymore, so we can drop it.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20231016154049.37147-1-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20231029145033.592566-2-alex.bennee@linaro.org>

blockdev: mirror: avoid potential deadlock when using iothread

The bdrv_getlength() function is a generated co-wrapper and uses
AIO_WAIT_WHILE() to wait for the spawned coroutine. AIO_WAIT_WHILE()
expects the lock to be acquired exactly once.

Fix a case where it may be acquired twice. This can happen when the
source node is explicitly specified as the @replaces parameter or if the
source node is a filter node.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Message-ID: <20231019131936.414246-4-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: avoid potential deadlock during bdrv_graph_wrlock() in bdrv_close()

by passing the BlockDriverState along, so the held AioContext can be
dropped before polling. See commit 31b2ddfea3 ("graph-lock: Unlock the
AioContext while polling") which introduced this functionality for
more information.

The only way to reach bdrv_close() is via bdrv_unref() and for calling
that the BlockDriverState's AioContext lock is supposed to be held.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Message-ID: <20231019131936.414246-3-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

blockjob: drop AioContext lock before calling bdrv_graph_wrlock()

Same rationale as in 31b2ddfea3 ("graph-lock: Unlock the AioContext
while polling"). Otherwise, a deadlock can happen.

The alternative would be to pass a BlockDriverState along to
bdrv_graph_wrlock(), but there is no BlockDriverState readily
available and it's also better conceptually, because the lock is held
for the job.

The function is always called with the job's AioContext lock held, via
one of the .abort, .clean, .free or .prepare job driver functions.
Thus, it's safe to drop it.

While mirror_exit_common() does hold a second AioContext lock while
calling block_job_remove_all_bdrv(), that is for the main thread's
AioContext and does not need to be dropped (bdrv_graph_wrlock(bs) also
skips dropping the lock if bdrv_get_aio_context(bs) ==
qemu_get_aio_context()).

Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Message-ID: <20231019131936.414246-2-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

iotests: Test media change with iothreads

iotests case 118 already tests all relevant operations for media change
with multiple devices, however never with iothreads. This changes the
test so that the virtio-scsi tests run with an iothread.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20231013153302.39234-3-kwolf@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Fix locking in media change monitor commands

blk_insert_bs() requires that the caller holds the AioContext lock for
the node to be inserted. Since commit c066e808e11, neglecting to do so
causes a crash when the child has to be moved to a different AioContext
to attach it to the BlockBackend.

This fixes qmp_blockdev_insert_anon_medium(), which is called for the
QMP commands 'blockdev-insert-medium' and 'blockdev-change-medium', to
correctly take the lock.

Cc: qemu-stable@nongnu.org
Fixes: https://issues.redhat.com/browse/RHEL-3922
Fixes: c066e808e11a5c181b625537b6c78e0de27a4801
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20231013153302.39234-2-kwolf@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

iotests: add tests for "qemu-img rebase" with compression

The test cases considered so far:

314 (new test suite):

1. Check that compression mode isn't compatible with "-f raw" (raw
   format doesn't support compression).
2. Check that rebasing an image onto no backing file preserves the data
   and writes the copied clusters actually compressed.
3. Same as 2, but with a raw backing file (i.e. the clusters copied from the
   backing are originally uncompressed -- we check they end up compressed
   after being merged).
4. Remove a single delta from a backing chain, perform the same checks
   as in 2.
5. Check that even when backing and overlay are initially uncompressed,
   copied clusters end up compressed when rebase with compression is
   performed.

271:

1. Check that when target image has subclusters, rebase with compression
   will make an entire cluster containing the written subcluster
   compressed.

Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-9-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-img: add compression option to rebase subcommand

If we rebase an image whose backing file has compressed clusters, we
might end up wasting disk space since the copied clusters are now
uncompressed.  In order to have better control over this, let's add
"--compress" option to the "qemu-img rebase" command.

Note that this option affects only the clusters which are actually being
copied from the original backing file.  The clusters which were
uncompressed in the target image will remain so.

Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-8-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

iotests/{024, 271}: add testcases for qemu-img rebase

As the previous commit changes the logic of "qemu-img rebase" (it's using
write alignment now), let's add a couple more test cases which would
ensure it works correctly.  In particular, the following scenarios:

024: add test case for rebase within one backing chain when the overlay
     cluster size > backings cluster size;
271: add test case for rebase images that contain subclusters.  Check
     that no extra allocations are being made.

Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-7-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-img: rebase: avoid unnecessary COW operations

When rebasing an image from one backing file to another, we need to
compare data from old and new backings.  If the diff between that data
happens to be unaligned to the target cluster size, we might end up
doing partial writes, which would lead to copy-on-write and additional IO.

Consider the following simple case (virtual_size == cluster_size == 64K):

base <-- inc1 <-- inc2

qemu-io -c "write -P 0xaa 0 32K" base.qcow2
qemu-io -c "write -P 0xcc 32K 32K" base.qcow2
qemu-io -c "write -P 0xbb 0 32K" inc1.qcow2
qemu-io -c "write -P 0xcc 32K 32K" inc1.qcow2
qemu-img rebase -f qcow2 -b base.qcow2 -F qcow2 inc2.qcow2

While doing rebase, we'll write a half of the cluster to inc2, and block
layer will have to read the 2nd half of the same cluster from the base image
inc1 while doing this write operation, although the whole cluster is already
read earlier to perform data comparison.

In order to avoid these unnecessary IO cycles, let's make sure every
write request is aligned to the overlay subcluster boundaries.  Using
subcluster size is universal as for the images which don't have them
this size equals to the cluster size. so in any case we end up aligning
to the smallest unit of allocation.

Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Message-ID: <20230919165804.439110-6-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-img: add chunk size parameter to compare_buffers()

Add @chsize param to the function which, if non-zero, would represent
the chunk size to be used for comparison.  If it's zero, then
BDRV_SECTOR_SIZE is used as default chunk size, which is the previous
behaviour.

In particular, we're going to use this param in img_rebase() to make the
write requests aligned to a predefined alignment value.

Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-5-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-img: rebase: use backing files' BlockBackend for buffer alignment

Since commit bb1c05973cf ("qemu-img: Use qemu_blockalign"), buffers for
the data read from the old and new backing files are aligned using
BlockDriverState (or BlockBackend later on) referring to the target image.
However, this isn't quite right, because buf_new is only being used for
reading from the new backing, while buf_old is being used for both reading
from the old backing and writing to the target.  Let's take that into account
and use more appropriate values as alignments.

Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Message-ID: <20230919165804.439110-4-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-iotests: 024: add rebasing test case for overlay_size > backing_size

Before previous commit, rebase was getting infitely stuck in case of
rebasing within the same backing chain and when overlay_size > backing_size.
Let's add this case to the rebasing test 024 to make sure it doesn't
break again.

Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-3-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-img: rebase: stop when reaching EOF of old backing file

In case when we're rebasing within one backing chain, and when target image
is larger than old backing file, bdrv_is_allocated_above() ends up setting
*pnum = 0.  As a result, target offset isn't getting incremented, and we
get stuck in an infinite for loop.  Let's detect this case and proceed
further down the loop body, as the offsets beyond the old backing size need
to be explicitly zeroed.

Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-2-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-file: Make qemu_fflush() return errors

This let us simplify code of this shape.

   qemu_fflush(f);
   int ret = qemu_file_get_error(f);
   if (ret) {
      return ret;
   }

into:

   int ret = qemu_fflush(f);
   if (ret) {
      return ret;
   }

I updated all callers where there is any error check.
qemu_fclose() don't need to check for f->last_error because
qemu_fflush() returns it at the beggining of the function.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-13-quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration: Remove transferred atomic counter

After last commit, it is a write only variable.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-12-quintela@redhat.com>

migration: Use migration_transferred_bytes()

There are only two differnces with the old value:

- the amount of QEMUFile that hasn't yet been flushed.  It can be
  discussed what is more exact, the new or the old one.
- the amount of transferred bytes that we forgot to account for (the
  newer is better, i.e. exact).

Notice that this two values are used to:
a - present to the user
b - calculate the rate_limit

So a few KB here and there is not going to make a difference.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-11-quintela@redhat.com>

qemu-file: Simplify qemu_file_get_error()

If we pass a NULL error is the same that returning directly the value.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-10-quintela@redhat.com>

migration: migration_rate_limit_reset() don't need the QEMUFile

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-9-quintela@redhat.com>

migration: migration_transferred_bytes() don't need the QEMUFile

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-8-quintela@redhat.com>

qemu-file: Remove _noflush from qemu_file_transferred_noflush()

qemu_file_transferred() don't exist anymore, so we can reuse the name.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-7-quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

qemu_file: Remove unused qemu_file_transferred()

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-6-quintela@redhat.com>

migration: Use the number of transferred bytes directly

We only use migration_transferred_bytes() to calculate the rate_limit,
for that we don't need to flush whatever is on the qemu_file buffer.
Remember that the buffer is really small (normal case is 32K if we use
iov's can be 64 * TARGET_PAGE_SIZE), so this is not relevant to
calculations.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-5-quintela@redhat.com>

qemu_file: total_transferred is not used anymore

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-4-quintela@redhat.com>

qemu_file: Use a stat64 for qemu_file_transferred

This way we can read it from any thread.
I checked that it gives the same value as the current one.  We never
use two qemu_files at the same time.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-3-quintela@redhat.com>

qemu-file: Don't increment qemu_file_transferred at qemu_file_fill_buffer

We only call qemu_file_transferred_* on the sending side. Remove the
increment at qemu_file_fill_buffer() and add asserts to
qemu_file_transferred* functions.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231025091117.6342-2-quintela@redhat.com>

migration: Stop migration immediately in RDMA error paths

In multiple places, RDMA errors are handled in a strange way, where it only
sets qemu_file_set_error() but not stop the migration immediately.

It's not obvious what will happen later if there is already an error.  Make
all such failures stop migration immediately.

Cc: Zhijian Li (Fujitsu) <lizhijian@fujitsu.com>
Cc: Markus Armbruster <armbru@redhat.com>
Cc: Juan Quintela <quintela@redhat.com>
Cc: Fabiano Rosas <farosas@suse.de>
Reported-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231024163933.516546-1-peterx@redhat.com>

migration: Deprecate old compression method

Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Acked-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231018115513.2163-6-quintela@redhat.com>

migration: Deprecate block migration

It is obsolete.  It is better to use driver-mirror with NBD instead.

CC: Kevin Wolf <kwolf@redhat.com>
CC: Eric Blake <eblake@redhat.com>
CC: Stefan Hajnoczi <stefanha@redhat.com>
CC: Hanna Czenczek <hreitz@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231018115513.2163-5-quintela@redhat.com>

migration: migrate 'blk' command option is deprecated.

Use blocked-mirror with NBD instead.

Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231018115513.2163-4-quintela@redhat.com>

migration: migrate 'inc' command option is deprecated.

Use blockdev-mirror with NBD instead.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231018115513.2163-3-quintela@redhat.com>

qemu-iotests: Filter warnings about block migration being deprecated

Create a new filter that removes the two warnings for test 183.

Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231018115513.2163-2-quintela@redhat.com>

Merge tag 'pull-lu-20231030' of https://gitlab.com/rth7680/qemu into staging

linux-user: Fix guest signal remapping after adjusting SIGABRT
linux-user: Implement VDSOs

* tag 'pull-lu-20231030' of https://gitlab.com/rth7680/qemu: (21 commits)
  build: Add update-linux-vdso makefile rule
  linux-user: Show vdso address in /proc/pid/maps
  linux-user/s390x: Add vdso
  linux-user/s390x: Rename __SIGNAL_FRAMESIZE to STACK_FRAME_OVERHEAD
  linux-user/ppc: Add vdso
  linux-user/loongarch64: Add vdso
  linux-user/riscv: Add vdso
  linux-user/hppa: Add vdso
  linux-user/arm: Add vdso
  linux-user/aarch64: Add vdso
  linux-user/x86_64: Add vdso
  linux-user/i386: Add vdso
  linux-user: Add gen-vdso tool
  linux-user: Load vdso image if available
  linux-user: Replace bprm->fd with bprm->src.fd
  linux-user: Use ImageSource in load_symbols
  linux-user: Use ImageSource in load_elf_image
  linux-user: Do not clobber bprm_buf swapping ehdr
  linux-user: Tidy loader_exec
  linux-user: Introduce imgsrc_read, imgsrc_read_alloc
  ...

Conflicts:
  linux-user/arm/signal.c
  Fix an #include context conflict.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Merge tag 'pull-ufs-20231030' of https://gitlab.com/jeuk20.kim/qemu into staging

ufs-next-pull-request

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCgAdFiEEUBfYMVl8eKPZB+73EuIgTA5dtgIFAmU/DfoACgkQEuIgTA5d
# tgKZ3g/+J38LTaktLPgUb0Kg390anPkIAkqqA1QZC8lC/FRSEWpgsNBqcvAASNTl
# jj1c80k/+Dvf9Ti1lmDNkuYczCFvKNJZQ1iRHmv2wc79A01GV0Ue6xayQjjNjoKK
# SBMIsFpArmFQjR2wGlkRc8PXha1JyWrsD4iPY6ZqedEcyuueLx69XbLL37FfVbQt
# 5IMnDqGkLCmrGowAjwurq2UM5IiYjeB4I5OwUgJC526zlyngXTFJimCWS6b2uUBk
# Yg1PnFffBsh11Pwmq4IZ1DAv3Bv/gFovenuatFqZrgqtfK7tEiARInIEsctu0U0a
# hPK/KojJAPF/cfMssRm1D1GCfsXM4tP2yFY/6q0wTRr9Dod8OSjlvfJR7+ez71/j
# aoY4N/nYYrZ6+pQNsPJcuBqQdtjdNUp4gUHx5qYxwwqZcHK4ubxpIvstmxceoLEX
# 3PG4O1iAapc/aL12ww9bYJ2lrbKGx7ZJU/Ij8bud8tYzLheG3xaYUEhonk7DE6+e
# AXFSad5CJTIF9Duh1uAMe1sV9GxELV8MHZSalqfGOhWYp7LzUBgouEJ1gQdOQbTK
# VsLs48WQ23OjWNKyAMaXQXdFO4FVbsjIg9nQXEHNRPkUownVHNVL8zu6EsXvHfch
# u691ygt5pD100SYdcDv73xTSeqP/rxqyYdxJl4LRkv/hGWU4y78=
# =Oisg
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 30 Oct 2023 10:59:22 JST
# gpg:                using RSA key 5017D831597C78A3D907EEF712E2204C0E5DB602
# gpg: Good signature from "Jeuk Kim <jeuk20.kim@samsung.com>" [unknown]
# gpg:                 aka "Jeuk Kim <jeuk20.kim@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: 5017 D831 597C 78A3 D907  EEF7 12E2 204C 0E5D B602

* tag 'pull-ufs-20231030' of https://gitlab.com/jeuk20.kim/qemu:
  hw/ufs: Modify lu.c to share codes with SCSI subsystem

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Merge tag 'pull-target-arm-20231027' of https://git-us.linaro.org/people/pmaydell/qemu-arm into staging

target-arm queue:
 * Correct minor errors in Cortex-A710 definition
 * Implement Neoverse N2 CPU model
 * Refactor feature test functions out into separate header
 * Fix syndrome for FGT traps on ERET
 * Remove 'hw/arm/boot.h' includes from various header files
 * pxa2xx: Refactoring/cleanup
 * Avoid using 'first_cpu' when first ARM CPU is reachable
 * misc/led: LED state is set opposite of what is expected
 * hw/net/cadence_gen: clean up to use FIELD macros
 * hw/net/cadence_gem: perform PHY access on write only
 * hw/net/cadence_gem: enforce 32 bits variable size for CRC

# -----BEGIN PGP SIGNATURE-----
#
# iQJNBAABCAA3FiEE4aXFk81BneKOgxXPPCUl7RQ2DN4FAmU7yz0ZHHBldGVyLm1h
# eWRlbGxAbGluYXJvLm9yZwAKCRA8JSXtFDYM3n4xEACK4ti+PFSJHVCQ69NzLLBT
# ybFGFMsMhXJTSNS30Pzs+KWCKWPP59knYBD4qO43W1iV6pPUhy+skr+BFCCRvBow
# se74+Fm1l4LmnuHxgukJzTdvRffI3v37alLn6Y/ioWe8bDpf/IJj8WLj8B1IPoNg
# fswJSGDLpPMovaz8NBQRzglUWpfyzxH+uuW779qBS1nuFdPOfIHKrocvvdrfogBP
# aO8AeiBzz5STW9Naeq+BIKho8S9LinSB6FHa+rRPUDkWx03lvRIvkgGPzHpXYy8I
# zAZ8gUQZyXprHAHMpnoBv8Wcw3Bwc2f+8xx8hnRRki3iBroXKfJA9NkeN0StQmL1
# ZHhfYkiKSS5diIFW5pX6ZixKbXHE2a4aH4zPVUNQriNWOevhe7n82mAPNFIYjk97
# ciTtd4I2oew48sDLSodMiirGL987Mit7KC23itVGezcNfQ9FnVTDmuGy8Rq52BZm
# u4TZjVBrtjQOdMBUcD2hKvXhikQNAdOhArPwNfOr0esSQL44MMEe+6Q5/Cbp0BOE
# stAY/xwSP2cY5mIPnAbIBELseEZsV8ySA3M0y1iRCJptjwbyWM+s1TYz0iXcqeOn
# l6LfiI6r1BqUeoWLGP4042R4FLyLNh6gU/TiFNLu7JJQjXl/EkRgqVXWYfzy2n51
# KKY6iGFi5r41sAU6GIXOkQ==
# =szC7
# -----END PGP SIGNATURE-----
# gpg: Signature made Fri 27 Oct 2023 23:37:49 JST
# gpg:                using RSA key E1A5C593CD419DE28E8315CF3C2525ED14360CDE
# gpg:                issuer "peter.maydell@linaro.org"
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>" [full]
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>" [full]
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>" [full]
# gpg:                 aka "Peter Maydell <peter@archaic.org.uk>" [unknown]
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83  15CF 3C25 25ED 1436 0CDE

* tag 'pull-target-arm-20231027' of https://git-us.linaro.org/people/pmaydell/qemu-arm: (41 commits)
  hw/net/cadence_gem: enforce 32 bits variable size for CRC
  hw/net/cadence_gem: perform PHY access on write only
  hw/net/cadence_gem: use FIELD to describe PHYMNTNC register fields
  hw/net/cadence_gem: use FIELD to describe DESCONF6 register fields
  hw/net/cadence_gem: use FIELD to describe IRQ register fields
  hw/net/cadence_gem: use FIELD to describe [TX|RX]STATUS register fields
  hw/net/cadence_gem: use FIELD to describe DMACFG register fields
  hw/net/cadence_gem: use FIELD to describe NWCFG register fields
  hw/net/cadence_gem: use FIELD to describe NWCTRL register fields
  hw/net/cadence_gem: use FIELD for screening registers
  hw/net/cadence_gem: use REG32 macro for register definitions
  misc/led: LED state is set opposite of what is expected
  hw/arm: Avoid using 'first_cpu' when first ARM CPU is reachable
  hw/arm/pxa2xx: Realize PXA2XX_I2C device before accessing it
  hw/intc/pxa2xx: Factor pxa2xx_pic_realize() out of pxa2xx_pic_init()
  hw/intc/pxa2xx: Pass CPU reference using QOM link property
  hw/intc/pxa2xx: Convert to Resettable interface
  hw/pcmcia/pxa2xx: Inline pxa2xx_pcmcia_init()
  hw/pcmcia/pxa2xx: Do not open-code sysbus_create_simple()
  hw/pcmcia/pxa2xx: Realize sysbus device before accessing it
  ...

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

build: Add update-linux-vdso makefile rule

This is not ideal, since it requires all cross-compilers
to be present rather than a simple subset.  But since it
is only run manually, should be good enough for now.

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

linux-user: Show vdso address in /proc/pid/maps

Tested-by: Helge Deller <deller@gmx.de>
Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user/s390x: Add vdso

Acked-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user/s390x: Rename __SIGNAL_FRAMESIZE to STACK_FRAME_OVERHEAD

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

linux-user/ppc: Add vdso

Add support in gen-vdso-elfn.c.inc for the DT_PPC64_OPT
dynamic tag: this is an integer, so does not need relocation.

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

linux-user/loongarch64: Add vdso

Requires a relatively recent binutils version in order to avoid
spurious R_LARCH_NONE relocations.  The presence of these relocs
are diagnosed by our gen-vdso tool.

Tested-by: Song Gao <gaosong@loongson.cn>
Reviewed-by: Song Gao <gaosong@loongson.cn>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user/riscv: Add vdso

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

linux-user/hppa: Add vdso

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

linux-user/arm: Add vdso

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

linux-user/aarch64: Add vdso

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

linux-user/x86_64: Add vdso

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

linux-user/i386: Add vdso

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1267
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Add gen-vdso tool

This tool will be used for post-processing the linked vdso image,
turning it into something that is easy to include into elfload.c.

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

linux-user: Load vdso image if available

The vdso image will be pre-processed into a C data array, with
a simple list of relocations to perform, and identifying the
location of signal trampolines.

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

linux-user: Replace bprm->fd with bprm->src.fd

There are only a couple of uses of bprm->fd remaining.
Migrate to the other field.

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

linux-user: Use ImageSource in load_symbols

Aside from the section headers, we're unlikely to hit the
ImageSource cache on guest executables.  But the interface
for imgsrc_read_* is better.

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

linux-user: Use ImageSource in load_elf_image

Change parse_elf_properties as well, as the bprm_buf argument
ties the two functions closely.

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

linux-user: Do not clobber bprm_buf swapping ehdr

Rearrange the allocation of storage for ehdr between load_elf_image
and load_elf_binary.  The same set of copies are done, but we don't
modify bprm_buf, which will be important later.

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

linux-user: Tidy loader_exec

Reorg the if cases to reduce indentation.
Test for 4 bytes in the file before checking the signatures.

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

linux-user: Introduce imgsrc_read, imgsrc_read_alloc

Introduced and initialized, but not yet really used.
These will tidy the current tests vs BPRM_BUF_SIZE.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Fix guest signal remapping after adjusting SIGABRT

The arithmetic within the loop was not adjusted properly after SIGRTMIN
was stolen for the guest SIGABRT.  The effect was that the guest libc
could not send itself __SIGRTMIN to wake sleeping threads.

Fixes: 38ee0a7dfb4b ("linux-user: Remap guest SIGABRT")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1967
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

migration: set file error on subsection loading

commit 13cde50889237 ("vmstate: Return error in case of error") sets
QemuFile error to stop reading from it and report to the caller (checked
by unit tests). We should do the same on subsection loading error.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231024084043.2926316-8-marcandre.lureau@redhat.com>

migration: rename vmstate_save_needed->vmstate_section_needed

The function is used on save at this point. The following commits will
use it on load.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231024084043.2926316-5-marcandre.lureau@redhat.com>

migration/ram: Fix compilation with -Wshadow=local

Rename the variable here to avoid that it shadows a variable from
the beginning of the function scope. With this change the code now
successfully compiles with -Wshadow=local.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231024092220.55305-1-thuth@redhat.com>

migration: Rename ram_compressed_pages() to compress_ram_pages()

We are moving to have all functions exported from ram-compress.c to
start with compress_.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-12-quintela@redhat.com>

migration: Merge flush_compressed_data() and compress_flush_data()

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-11-quintela@redhat.com>

migration: Move ram_flush_compressed_data() to ram-compress.c

As we export it, rename it compress_flush_data().

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-10-quintela@redhat.com>

migration: Export send_queued_data()

This function is only used for compression.  So we rename it as
compress_send_queued_data().  We put it on ram-compress.h because we
are moving it later to ram-compress.c.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-9-quintela@redhat.com>

migration: Create compress_update_rates()

So we can move more compression_counters stuff to ram-compress.c.
Create compression_counters struct to add the stuff that was on
MigrationState.

Reviewed-by: Lukas Straub <lukasstraub2@web.de>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-8-quintela@redhat.com>

migration: Move busy++ to migrate_with_multithread

And now we can simplify save_compress_page().

Reviewed-by: Lukas Straub <lukasstraub2@web.de>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-7-quintela@redhat.com>

migration: Simplify compress_page_with_multithread()

Move the goto to a while true.

Reviewed-by: Lukas Straub <lukasstraub2@web.de>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-6-quintela@redhat.com>

migration: Make compress_data_with_multithreads return bool

Reviewed-by: Lukas Straub <lukasstraub2@web.de>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-5-quintela@redhat.com>

migration: Remove save_page_use_compression()

After previous patch, we disable the posiblity that we use compression
together with xbzrle.  So we can use directly migrate_compress().

Once there, now we don't need the rs parameter, so remove it.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-4-quintela@redhat.com>

migration: Give one error if trying to set COMPRESSION and XBZRLE

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-3-quintela@redhat.com>

migration: Give one error if trying to set MULTIFD and XBZRLE

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019110724.15324-2-quintela@redhat.com>

migration: Rename ram_handle_compressed() to ram_handle_zero()

Now that we know it only handles zero, we can remove the ch parameter.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019085259.13307-3-quintela@redhat.com>

migration: Receiving a zero page non zero is an error

We don't allow non zero compressed pages since:

commit 3edcd7e6ebae3ef0ac178eed5f4225803159562d
Author: Peter Lieven <pl@kamp.de>
Date:   Tue Mar 26 10:58:35 2013 +0100

    migration: search for zero instead of dup pages

RDMA case is a bit more complicated, but they don't handle it since:

commit a1febc4950f2c6232c002f401d7cd409f6fa6a88
Author: Richard Henderson <rth@twiddle.net>
Date:   Mon Aug 29 11:46:14 2016 -0700

    cutils: Export only buffer_is_zero

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231019085259.13307-2-quintela@redhat.com>

migration/doc: We broke backwards compatibility

When we detect that we have broken backwards compatibility in a
released version, we can't do anything for that version.  But once we
fix that bug on the next released version, we can "mitigate" that
problem when migrating to new versions to give a way out of that
machine until it does a hard reboot.

Acked-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231018112827.1325-5-quintela@redhat.com>

migration/doc: How to migrate when hosts have different features

Sometimes devices have different features depending of things outside
of qemu.  For instance the kernel.  Document how to handle that cases.

Acked-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231018112827.1325-4-quintela@redhat.com>

migration/doc: Add documentation for backwards compatiblity

State what are the requeriments to get migration working between qemu
versions.  And once there explain how one is supposed to implement a
new feature/default value and not break migration.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Acked-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231018112827.1325-3-quintela@redhat.com>

migration/doc: Add contents

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231018112827.1325-2-quintela@redhat.com>

Merge tag 'pull-hex-20231018' of https://github.com/quic/qemu into staging

hexagon: GETPC() fixes, shadowing fixes

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCgAdFiEEPWaq5HRZSCTIjOD4GlSvuOVkbDIFAmUwjhYACgkQGlSvuOVk
# bDK1Kw//fqzt62h+x182uOZz3XCr9co8mJGXAnu9cy8Jkmf7BGT6jNsjpQe5KjyH
# FJNYsfx2dnzC6bqID4ioiBTA3Vgq5XyumpxORzCJTerda3bJr8Q3qt3yK80IIsdf
# HkeYfxHFsOG57KqyRJYZenJQBeCq/CcPWfWrxLspFPa96RF4Ui34srucjS6QZp4R
# l6jlkxmvkkEV0Cq3vwKkTlrOTUDWQ5dZw5GfVrE3D6WfpB6SzprzDrWAi4dBSG3S
# ZNg1WU229A83n4hWw91brHzejE7KHbgjbhfGnlS05AHIqGrhAzDQIWXm4q1yMlca
# /lGTBbKZzg6+RmFj6Rb5MzeaiVIQaQ75L4mJ72C03v+7ZHtV0m7nxTREEbZFsstZ
# 7T8yutG/ljHHkb7xXL5kvJCIA5/UPv9qH3jYQebkVh0vgKSzZOPuPE1OVymTUmqb
# tWKdWcoq0Dj+FWBg/Mnb6RrECA2Rn91nRv6byVkPZ4xRWUxjUMmFW0Mbpa5SO5z/
# pGQGhcnGOMfZQBCxdwDU17RVetXho3xr2JCmYMG/vjin7lqoY5D+AFl+R7UAm/0p
# C4Pe/uqGRwXlEkL5lwO2RIoMp0bXRDZI9yQeeWVg6uBbWDKBlM2TmHiu6lyRofTl
# 0/Go8nhumecT09xMTFMp5doQL8HSYKv0DnzqFjM8a5Tc5gQSFx4=
# =1KeO
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 19 Oct 2023 11:01:58 JST
# gpg:                using RSA key 3D66AAE474594824C88CE0F81A54AFB8E5646C32
# gpg: Good signature from "Brian Cain <bcain@kernel.org>" [unknown]
# gpg:                 aka "Brian Cain (QuIC) <bcain@quicinc.com>" [unknown]
# gpg:                 aka "Brian Cain (CAF) <bcain@codeaurora.org>" [unknown]
# gpg:                 aka "bcain" [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: 6350 20F9 67A7 7164 79EF  49E0 175C 464E 541B 6D47
#      Subkey fingerprint: 3D66 AAE4 7459 4824 C88C  E0F8 1A54 AFB8 E564 6C32

* tag 'pull-hex-20231018' of https://github.com/quic/qemu:
  target/hexagon: fix some occurrences of -Wshadow=local
  target/hexagon: move GETPC() calls to top level helpers

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

hw/ufs: Modify lu.c to share codes with SCSI subsystem

This patch removes the code that ufs-lu was duplicating from
scsi-hd and allows them to share code.

It makes ufs-lu have a virtual scsi-bus and scsi-hd internally.
This allows scsi related commands to be passed thorugh to the scsi-hd.
The query request and nop command work the same as the existing logic.

Well-known lus do not have a virtual scsi-bus and scsi-hd, and
handle the necessary scsi commands by emulating them directly.

Signed-off-by: Jeuk Kim <jeuk20.kim@samsung.com>

Merge tag 'pull-request-2023-10-27' of https://gitlab.com/thuth/qemu into staging

* Fix global variable shadowing in test code
* Avoid recompiling libfdt in the FreeBSD VM
* Mark old pc machine types as deprecated
* Force IPv4 in the ipmi-bt-test

# -----BEGIN PGP SIGNATURE-----
#
# iQJFBAABCAAvFiEEJ7iIR+7gJQEY8+q5LtnXdP5wLbUFAmU7g+wRHHRodXRoQHJl
# ZGhhdC5jb20ACgkQLtnXdP5wLbWzIQ//WkiEtXAB4zn/b8H0KfRF283FoFMstBPu
# xrF90WcXMidAPRK8iFhwlae0rVyo40V6A3+/0rntMZ6WPJ3pVvFhmM4zIQgX+1+F
# HT8To7ngngeEey5gJNbOs9GdAFSmF8HwB+HWnEGf2fjWsUvIWgm458ikPUWvhXaJ
# 08wH4V+bmEAKYgny8/ys70E1iFeMwfg+LKXHzVAavaQ8GjM9xFXOjRe8XEq2P7/8
# T99O+hn1iGJfrvZTdJ01awRuOUzA1YTKLCYNSzfRc/j0eYwI28CBm12/sZn4dUA8
# 4rV33ISPw55QmxdVLCfGhAJ6Ih0trOgf8eDBkhcgqWt26ymcKyVL/7kJHcuoY9Q2
# onaN3XT39lVz/HmcuTl9a0sePsZd4d7p0vRsccwTSisX8RpLDab8I8oU8Nvs3VQG
# t23K3XLrdSyMbxsbi9b5BdXJVgkz4irEKG1dX8laIGMGD5zVi/LX2BWjP6QaIOZi
# OyopRQN2uTi6cry5NmblpeLDNJwQdQVxJKlqcvDQQl8TjPXleNtIXL4dr+IpNC5T
# Uo8GJ6+b+AzUI4EaNFybVDPAYHFMJKxvN/QUEjExKV1QTP3V2V9jCbGCF8nio8pp
# ISo11Lo/Z4Z39NwDBH13UuSOdD1rlrU2uuq6LaiRTBfCOTYumzUMvlDTNYqGuSLr
# +1q1GTGsHO0=
# =A6CT
# -----END PGP SIGNATURE-----
# gpg: Signature made Fri 27 Oct 2023 18:33:32 JST
# gpg:                using RSA key 27B88847EEE0250118F3EAB92ED9D774FE702DB5
# gpg:                issuer "thuth@redhat.com"
# gpg: Good signature from "Thomas Huth <th.huth@gmx.de>" [full]
# gpg:                 aka "Thomas Huth <thuth@redhat.com>" [full]
# gpg:                 aka "Thomas Huth <huth@tuxfamily.org>" [full]
# gpg:                 aka "Thomas Huth <th.huth@posteo.de>" [unknown]
# Primary key fingerprint: 27B8 8847 EEE0 2501 18F3  EAB9 2ED9 D774 FE70 2DB5

* tag 'pull-request-2023-10-27' of https://gitlab.com/thuth/qemu:
  ipmi-bt-test: force ipv4
  tests/vm/freebsd: Add additional library paths for libfdt
  docs/about: Mark the old pc-i440fx-2.0 - 2.3 machine types as deprecated
  tests/coroutine: Clean up global variable shadowing
  tests/aio: Clean up global variable shadowing
  tests/npcm7xx_adc: Clean up global variable shadowing
  tests/rtl8139: Clean up global variable shadowing
  tests/cdrom-test: Clean up global variable shadowing in prepare_image()
  tests/virtio-scsi: Clean up global variable shadowing
  tests/throttle: Clean up global variable shadowing
  system/qtest: Clean up global variable shadowing in qtest_server_init()

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

hw/net/cadence_gem: enforce 32 bits variable size for CRC

The CRC was stored in an unsigned variable in gem_receive. Change it for
a uint32_t to ensure we have the correct variable size here.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-12-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: perform PHY access on write only

The MDIO access is done only on a write to the PHYMNTNC register. A
subsequent read is used to retrieve the result but does not trigger an
MDIO access by itself.

Refactor the PHY access logic to perform all accesses (MDIO reads and
writes) at PHYMNTNC write time.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-11-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: use FIELD to describe PHYMNTNC register fields

Use the FIELD macro to describe the PHYMNTNC register fields.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-10-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: use FIELD to describe DESCONF6 register fields

Use the FIELD macro to describe the DESCONF6 register fields.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-id: 20231017194422.4124691-9-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: use FIELD to describe IRQ register fields

Use de FIELD macro to describe the IRQ related register fields.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-8-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: use FIELD to describe [TX|RX]STATUS register fields

Use de FIELD macro to describe the TXSTATUS and RXSTATUS register
fields.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-7-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: use FIELD to describe DMACFG register fields

Use de FIELD macro to describe the DMACFG register fields.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-6-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: use FIELD to describe NWCFG register fields

Use de FIELD macro to describe the NWCFG register fields.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-5-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: use FIELD to describe NWCTRL register fields

Use the FIELD macro to describe the NWCTRL register fields.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-4-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: use FIELD for screening registers

Describe screening registers fields using the FIELD macros.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-3-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/net/cadence_gem: use REG32 macro for register definitions

Replace register defines with the REG32 macro from registerfields.h in
the Cadence GEM device.

Signed-off-by: Luc Michel <luc.michel@amd.com>
Reviewed-by: sai.pavan.boddu@amd.com
Message-id: 20231017194422.4124691-2-luc.michel@amd.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

misc/led: LED state is set opposite of what is expected

Testing of the LED state showed that when the LED polarity was
set to GPIO_POLARITY_ACTIVE_LOW and a low logic value was set on
the input GPIO of the LED, the LED was being turn off when it was
expected to be turned on.

Fixes: ddb67f6402 ("hw/misc/led: Allow connecting from GPIO output")
Signed-off-by: Glenn Miles <milesg@linux.vnet.ibm.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Andrew Jeffery <andrew@codeconstruct.com.au>
Message-id: 20231024191945.4135036-1-milesg@linux.vnet.ibm.com
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm: Avoid using 'first_cpu' when first ARM CPU is reachable

Prefer using a well known local first CPU rather than a global one.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20231025065909.57344-1-philmd@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/pxa2xx: Realize PXA2XX_I2C device before accessing it

qbus_new(), called in i2c_init_bus(), should not be called
on unrealized device.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-id: 20231020130331.50048-10-philmd@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/intc/pxa2xx: Factor pxa2xx_pic_realize() out of pxa2xx_pic_init()

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-id: 20231020130331.50048-9-philmd@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/intc/pxa2xx: Pass CPU reference using QOM link property

QOM objects shouldn't access each other internals fields
except using the QOM API.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-id: 20231020130331.50048-8-philmd@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/intc/pxa2xx: Convert to Resettable interface

Factor reset code out of the DeviceRealize() handler.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-id: 20231020130331.50048-7-philmd@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/pcmcia/pxa2xx: Inline pxa2xx_pcmcia_init()

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20231020130331.50048-6-philmd@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/pcmcia/pxa2xx: Do not open-code sysbus_create_simple()

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20231020130331.50048-5-philmd@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/pcmcia/pxa2xx: Realize sysbus device before accessing it

sysbus_mmio_map() should not be called on unrealized device.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20231020130331.50048-4-philmd@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/sd/pxa2xx: Do not open-code sysbus_create_simple()

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20231020130331.50048-3-philmd@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>