Hi Jonathan,
On 4/14/22 5:33 PM, Jonathan Cameron wrote:
On Thu, 14 Apr 2022 15:35:41 +0800
Gavin Shan wrote:
On 4/14/22 10:49 AM, wangyanan (Y) wrote:
On 2022/4/14 10:37, Gavin Shan wrote:
On 4/14/22 10:27 AM, wangyanan (Y) wrote:
On 2022/4/14 8:08, Gavin Shan wrote:
On 4/13/22 8:39 PM, w
Hi Yanan,
On 4/14/22 5:29 PM, wangyanan (Y) wrote:
On 2022/4/14 15:35, Gavin Shan wrote:
On 4/14/22 10:49 AM, wangyanan (Y) wrote:
On 2022/4/14 10:37, Gavin Shan wrote:
On 4/14/22 10:27 AM, wangyanan (Y) wrote:
On 2022/4/14 8:08, Gavin Shan wrote:
On 4/13/22 8:39 PM, wangyanan (Y) wrote:
O
在 2022/4/8 20:28, Maxime Coquelin 写道:
This patch introduces new Vhost backend callbacks to
support RSS, and makes them called in Virtio-net
device.
It will be used by Vhost-user backend implementation to
support RSS feature.
Signed-off-by: Maxime Coquelin
---
hw/net/vhost_net-stub.c
在 2022/4/8 20:28, Maxime Coquelin 写道:
The goal of this series is to add support for Virtio RSS
feature to the Vhost-user backend.
First patches are preliminary reworks to support variable
RSS key and indirection table length. eBPF change only adds
checks on whether the key length is 40B, it do
在 2022/4/8 20:28, Maxime Coquelin 写道:
This patch is a preliminary rework to support RSS with
Vhost-user backends. It enables supporting different types
of hashes, key lengths and indirection table lengths.
This patch does not introduces behavioral changes.
Signed-off-by: Maxime Coquelin
---
On Thu, Apr 14, 2022 at 11:57 PM Niklas Cassel via
wrote:
> The device tree property "mmu-type" is currently exported as either
> "riscv,sv32" or "riscv,sv48".
>
> However, the riscv cpu device tree binding [1] has a specific value
> "riscv,none" for a HART without a MMU.
>
> Set the device tree
On Thu, Apr 14, 2022 at 11:55 PM Niklas Cassel via
wrote:
>
> The device tree property "mmu-type" is currently exported as either
> "riscv,sv32" or "riscv,sv48".
>
> However, the riscv cpu device tree binding [1] has a specific value
> "riscv,none" for a HART without a MMU.
>
> Set the device tree
Eric Auger writes:
> Hi Alex,
>
> On 4/7/22 5:00 PM, Alex Bennée wrote:
>> When trying to work out what the virtio-net-tests where doing it was
>> hard because the g_test_trap_subprocess redirects all output to
>> /dev/null. Lift this restriction by using the appropriate flags so you
>> can see
> -Original Message-
> From: Christian Schoenebeck
> Sent: 2022年4月14日 19:24
> To: qemu-devel@nongnu.org; Shi, Guohuai
> Cc: Bin Meng ; Greg Kurz
> Subject: Re: [RFC PATCH 0/4] 9pfs: Add 9pfs support for Windows host
>
> [Please note: This e-mail is from an EXTERNAL e-mail address]
>
On Thu, Apr 14, 2022 at 07:57:56PM +0200, Paolo Bonzini wrote:
> Signed-off-by: Paolo Bonzini
> ---
> block/nbd.c | 8 +++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
Reviewed-by: Eric Blake
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualiz
On Thu, Apr 14, 2022 at 07:57:55PM +0200, Paolo Bonzini wrote:
> requests[].receiving is set by nbd_receive_replies() under the receive_mutex;
> Read it under the same mutex as well. Waking up receivers on errors happens
> after each reply finishes processing, in nbd_co_receive_one_chunk().
> If t
On Thu, Apr 14, 2022 at 07:57:54PM +0200, Paolo Bonzini wrote:
> Remove the confusing, and most likely wrong, atomics. The only function
> that used to be somewhat in a hot path was nbd_client_connected(),
> but it is not anymore after the previous patches.
>
> The same logic is used both to chec
On Thu, Apr 14, 2022 at 07:57:53PM +0200, Paolo Bonzini wrote:
> Prepare for the next patch, so that the diff is less confusing.
>
> nbd_client_connecting is moved closer to the definition point.
>
> nbd_client_connecting_wait() is kept only for the reconnection
> logic; when it is used to check
On Thu, Apr 14, 2022 at 07:57:52PM +0200, Paolo Bonzini wrote:
> The condition for waiting on the s->free_sema queue depends on
> both s->in_flight and s->state. The latter is currently using
> atomics, but this is quite dubious and probably wrong.
>
> Because s->state is written in the main thre
On Thu, Apr 14, 2022 at 07:57:51PM +0200, Paolo Bonzini wrote:
> Elevate s->in_flight early so that other incoming requests will wait
> on the CoQueue in nbd_co_send_request; restart them after getting back
> from nbd_reconnect_attempt. This could be after the reconnect timer or
> nbd_cancel_in_fl
On Thu, Apr 14, 2022 at 07:57:48PM +0200, Paolo Bonzini wrote:
> The .reply_possible field of s->requests is never set to false. This is
> not a problem as it is only a safeguard to detect protocol errors,
> but it's sloppy. In fact, the field is actually not necessary at all,
> because .coroutin
Remove the confusing, and most likely wrong, atomics. The only function
that used to be somewhat in a hot path was nbd_client_connected(),
but it is not anymore after the previous patches.
The same logic is used both to check if a request had to be reissued
and also in nbd_reconnecting_attempt().
requests[].receiving is set by nbd_receive_replies() under the receive_mutex;
Read it under the same mutex as well. Waking up receivers on errors happens
after each reply finishes processing, in nbd_co_receive_one_chunk().
If there is no currently-active reply, there are two cases:
* either there
Prepare for the next patch, so that the diff is less confusing.
nbd_client_connecting is moved closer to the definition point.
nbd_client_connecting_wait() is kept only for the reconnection
logic; when it is used to check if a request has to be reissued,
use the renamed function nbd_client_will_r
The condition for waiting on the s->free_sema queue depends on
both s->in_flight and s->state. The latter is currently using
atomics, but this is quite dubious and probably wrong.
Because s->state is written in the main thread too, for example by
the yank callback, it cannot be protected by a CoM
Signed-off-by: Paolo Bonzini
---
block/nbd.c | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/block/nbd.c b/block/nbd.c
index 31c684772e..d0d94b40bd 100644
--- a/block/nbd.c
+++ b/block/nbd.c
@@ -81,12 +81,18 @@ typedef struct BDRVNBDState {
NBDClientRequest request
The .reply_possible field of s->requests is never set to false. This is
not a problem as it is only a safeguard to detect protocol errors,
but it's sloppy. In fact, the field is actually not necessary at all,
because .coroutine is set to NULL in NBD_FOREACH_REPLY_CHUNK after
receiving the last ch
It is unnecessary to check nbd_client_connected() because every time
s->state is moved out of NBD_CLIENT_CONNECTED the socket is shut down
and all coroutines are resumed.
The only case where it was actually needed is when the NBD
server disconnects and there is no reconnect-delay. In that
case, n
Elevate s->in_flight early so that other incoming requests will wait
on the CoQueue in nbd_co_send_request; restart them after getting back
from nbd_reconnect_attempt. This could be after the reconnect timer or
nbd_cancel_in_flight have cancelled the attempt, so there is no
need anymore to cancel
The main point of this series is patch 7, which removes the dubious and
probably wrong use of atomics in block/nbd.c. This in turn is enabled
mostly by the cleanups in patches 3-5. Together, they introduce a
QemuMutex that synchronizes the NBD client coroutines, the reconnect_delay
timer and nbd_
Several coroutine functions in block/nbd.c are not marked as such. This
patch adds a few more markers; it is not exhaustive, but it focuses
especially on:
- places that wake other coroutines, because aio_co_wake() has very
different semantics inside a coroutine (queuing after yield vs. entering
i
On 4/14/22 07:56, Alex Bennée wrote:
#ifdef CONFIG_TRACE_LOG
-qemu_loglevel |= LOG_TRACE;
+log_flags |= LOG_TRACE;
#endif
+qemu_loglevel = log_flags;
+
This looked weird - so should we consider a qatomic_set here to avoid an
inconsistent set of flags being read non-atomically el
Hi Alex,
On 4/7/22 5:00 PM, Alex Bennée wrote:
> When trying to work out what the virtio-net-tests where doing it was
> hard because the g_test_trap_subprocess redirects all output to
> /dev/null. Lift this restriction by using the appropriate flags so you
> can see something similar to what the v
On Tue, 12 Apr 2022 at 17:49, Peter Maydell wrote:
>
> The sysbus floppy controllers (devices sysbus-fdc and sun-fdtwo)
> don't support DMA. The core floppy controller code expects this to
> be indicated by setting FDCtrl::dma_chann to -1. This used to be
> done in the device instance_init funct
Make -m syntactic sugar for a compound property "-machine
mem.{size,max-size,slots}". The new property does not have
the magic conversion to megabytes of unsuffixed arguments,
and also does not understand that "0" means the default size
(you have to leave it out to get the default). This means
th
As part of converting -boot to a property with a QAPI type, define
the struct and use it throughout QEMU to access boot configuration.
machine_boot_parse takes care of doing the QemuOpts->QAPI conversion by
hand, for now.
Signed-off-by: Paolo Bonzini
---
hw/arm/nseries.c| 2 +-
hw/core/
Make -boot syntactic sugar for a compound property "-machine
boot.{order,menu,...}".
machine_boot_parse is replaced by the setter for the property.
Signed-off-by: Paolo Bonzini
---
hw/core/machine.c | 99 -
include/hw/boards.h | 1 -
softmmu/vl.c
This allows setting memory properties without going through
vl.c, and have them validated just the same.
Signed-off-by: Paolo Bonzini
---
hw/core/machine.c | 21 +++--
softmmu/vl.c | 17 +++--
2 files changed, 22 insertions(+), 16 deletions(-)
diff --git a/hw/co
Handle HostMemoryBackend creation and setting of ms->ram entirely in
machine_run_board_init.
Signed-off-by: Paolo Bonzini
---
hw/core/machine.c | 70 ++---
hw/core/numa.c | 2 +-
hw/sparc/sun4m.c| 5 ++--
include/hw/boards.h | 4 +--
softmmu/
As the next step in turning command line options into shortcuts, this series
does -boot and -m. It also makes -M memory-backend a link instead of special
casing it in vl.c, and makes the MachineState validate memory configuration
without needing help from vl.c.
Paolo Bonzini (5):
machine: use Q
On 12/4/22 18:49, Peter Maydell wrote:
The sysbus floppy controllers (devices sysbus-fdc and sun-fdtwo)
don't support DMA. The core floppy controller code expects this to
be indicated by setting FDCtrl::dma_chann to -1. This used to be
done in the device instance_init functions sysbus_fdc_initf
On 4/14/22 08:35, Alex Bennée wrote:
+/**
+ * valid_filename_template:
+ *
+ * Validate the filename template. Require %d if per_thread, allow it
+ * otherwise; require no other % within the template.
+ * Return 0 if invalid, 1 if stderr, 2 if strdup, 3 if pid printf.
From a neatness point of
Hi Chenxiang,
On 4/7/22 9:57 AM, chenxiang via wrote:
> From: Xiang Chen
>
> In function memory_region_iommu_replay(), it decides to notify() or not
> according to the perm of returned IOMMUTLBEntry. But for smmuv3, the
> returned perm is always IOMMU_NONE even if the translation success.
I think
On 4/14/22 08:18, Alex Bennée wrote:
Richard Henderson writes:
Rename to emphasize this is the file-scope global variable.
Reviewed-by: Philippe Mathieu-Daudé
Signed-off-by: Richard Henderson
---
util/log.c | 13 ++---
1 file changed, 6 insertions(+), 7 deletions(-)
diff --git
The device tree property "mmu-type" is currently exported as either
"riscv,sv32" or "riscv,sv48".
However, the riscv cpu device tree binding [1] has a specific value
"riscv,none" for a HART without a MMU.
Set the device tree property "mmu-type" to "riscv,none" when the CPU mmu
option is disabled
Richard Henderson writes:
> Add a new log flag, tid, to turn this feature on.
> Require the log filename to be set, and to contain %d.
>
> Do not allow tid to be turned off once it is on, nor let
> the filename be change thereafter. This avoids the need
> for signalling each thread to re-open
On 14/04/2022 15:01, Thomas Huth wrote:
The fuzz tests are currently scheduled for all targets, but their setup
code limits the run to "i386", so that these tests always show "SKIP"
on other targets. Move it to the right x86 list in meson.build, then
we can drop the architecture check during runt
Richard Henderson writes:
> Use FILE* for global_file. We can perform an rcu_read on that
> just as easily as RCUCloseFILE*. This simplifies a couple of
> places, where previously we required taking the rcu_read_lock
> simply to avoid racing to dereference RCUCloseFile->fd.
>
> Only allocate
Richard Henderson writes:
> Merge the close from the changed_name block with the close
> from the !need_to_open_file block.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> s/QemuLogFile/RCUCloseFILE/
> s/qemu_logfile_free/rcu_close_file/
>
> Emphasize that this is only a carrier for passing a pointer
> to call_rcu for closing, and not the real logfile.
>
> Reviewed-by: Philippe Mathieu-Daudé
> Signed-off-by: Richard Henderson
Review
Richard Henderson writes:
> Only call is_daemonized once.
> We require the result on all paths after this point.
>
> Reviewed-by: Philippe Mathieu-Daudé
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> Rename to emphasize this covers the file-scope global variables.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> Rename to emphasize this is the file-scope global variable.
>
> Reviewed-by: Philippe Mathieu-Daudé
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> Rename to emphasize this is the file-scope global variable.
>
> Reviewed-by: Philippe Mathieu-Daudé
> Signed-off-by: Richard Henderson
> ---
> util/log.c | 13 ++---
> 1 file changed, 6 insertions(+), 7 deletions(-)
>
> diff --git a/util/log.c b/util/log.c
Richard Henderson writes:
> Perform all logfile setup at startup in one step.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> The only real use is in cpu_abort, where we have just
> flushed the file via qemu_log_unlock, and are just about
> to force-crash the application via abort. We do not
> really need to close the FILE before the abort.
>
> The two uses in test-logging.c can be handled
Richard Henderson writes:
> Perform all logfile setup in one step.
>
> Signed-off-by: Richard Henderson
> ---
> linux-user/main.c | 11 +--
> 1 file changed, 5 insertions(+), 6 deletions(-)
>
> diff --git a/linux-user/main.c b/linux-user/main.c
> index d263b2a669..0297ae8321 100644
>
Richard Henderson writes:
> Perform all logfile setup in one step.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> Provide a function to set both filename and flags at
> the same time. This is the common case at startup.
>
> Signed-off-by: Richard Henderson
> ---
> v2: Return bool, per recommendations in qapi/error.h (phil).
> ---
> include/qemu/log.h | 1 +
> util/log.c
Like -set and -readconfig, it would not really be too hard to
extend -writeconfig to parsing mechanisms other than QemuOpts.
However, the uses of -writeconfig are substantially more
limited, as it is generally easier to write the configuration
by hand in the first place. In addition, -writeconfig
Richard Henderson writes:
> The bug referenced in os-win32.h was fixed in mingw-w64 v6.
>
> According to repology, version 5 used by ubuntu 18, which is
> not yet out of support, so provide a meson link test for it.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Benn
Richard Henderson writes:
> Move QemuLogFile, qemu_logfile, and all inline functions into qemu/log.c.
> No need to expose these implementation details in the api.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
> ---
> include/qemu/log.h| 38 --
Richard Henderson writes:
> Use qemu_log_lock/unlock instead of the raw rcu_read.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Am 07.04.2022 um 10:39 hat Denis V. Lunev geschrieben:
> Simple grep for the .bdrv_co_check callback presence gives the following
> list of block drivers
> * QED
> * VDI
> * VHDX
> * VMDK
> * Parallels
> which have this callback. The presense of the callback means that
> consistency check is suppor
Richard Henderson writes:
> All uses flush output immediately before or after qemu_log_unlock.
> Instead of a separate call, move the flush into qemu_log_unlock.
>
> Reviewed-by: Philippe Mathieu-Daudé
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> We have extra stuff to log at the same time.
> Hoist the qemu_log_lock/unlock to the caller and use fprintf.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> All uses flush output immediately before or after qemu_log_unlock.
> Instead of a separate call, move the flush into qemu_log_unlock.
>
> Reviewed-by: Philippe Mathieu-Daudé
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> We have extra stuff to log at the same time.
> Hoist the qemu_log_trylock/unlock to the caller and use fprintf.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> Now that the log buffer is flushed after every qemu_log_unlock,
> which includes every call to qemu_log, we do not need to force
> line buffering (or unbuffering for windows). Block buffer the
> entire loggable unit.
>
> Signed-off-by: Richard Henderson
Reviewed-b
Richard Henderson writes:
> Now that all uses have been updated, consider a missing
> test of the result of qemu_log_trylock a bug and Werror.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> The only user of this feature, tcg_dump_ops, has been
> converted to use fprintf directly.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
From: Moteen Shah
The error message in virtio_init_region_cache()
is given a prefix virtio.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/230
Buglink: https://bugs.launchpad.net/qemu/+bug/1919021``
Signed-off-by: Moteen Shah
---
hw/virtio/virtio.c | 6 +++---
1 file changed, 3 inser
Richard Henderson writes:
> This is redundant with the logging done in cpu_common_reset.
>
> Reviewed-by: Philippe Mathieu-Daudé
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
Richard Henderson writes:
> Inside log_cpu_state, we perform qemu_log_trylock/unlock, which need
> not be done if we have already performed the lock beforehand.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
--
Alex Bennée
The fuzz tests are currently scheduled for all targets, but their setup
code limits the run to "i386", so that these tests always show "SKIP"
on other targets. Move it to the right x86 list in meson.build, then
we can drop the architecture check during runtime, too.
Signed-off-by: Thomas Huth
---
On 14/04/2022 14.28, Thomas Huth wrote:
The fuzz-sdcard-test is currently scheduled for all targets,
but the code limits itself to "i386". Move it to the right
list in meson.build and allow it to be run on "x86_64", too.
While we're at it, also clean up the wrong indentation in
fuzz-sdcard-test.
After re-reading what I just wrote, I am leaning towards disabling use of
KVM_CAP_IRQFD_RESAMPLE as it seems last worked on POWER8 and never since :)
Did I miss something in the picture (hey Cedric)?
How about disabling it like this?
=
diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci
Hello Alexey,
Thanks for taking over.
On 4/13/22 06:56, Alexey Kardashevskiy wrote:
On 3/17/22 06:16, Cédric Le Goater wrote:
Timothy,
On 3/16/22 17:29, Cédric Le Goater wrote:
Hello,
I've been struggling for some time with what is looking like a
potential bug in QEMU/KVM on the POWER9
The fuzz-sdcard-test is currently scheduled for all targets,
but the code limits itself to "i386". Move it to the right
list in meson.build and allow it to be run on "x86_64", too.
While we're at it, also clean up the wrong indentation in
fuzz-sdcard-test.c (it was using 3 spaces instead of 4 in s
I will start working on a cleaner patch to allow the correct memory
ordering enforcement to be implemented, as my current PoC is very hacky.
I'll post it to the mailing list as soon as I'm done.
Redha
On 08/04/2022 17:27, Richard Henderson wrote:
> On 4/8/22 05:21, Redha Gouicem wrote:
>> We are
On Wed, Apr 13, 2022 at 12:20:54PM +0800, Liu Yiding wrote:
> Refer to 26ec190964 virtiofsd: Do not use a thread pool by default
>
> Signed-off-by: Liu Yiding
Looks good. Our default used to be --thread-pool-size=64. But we changed
it to using no thread pool because on lower end of workloads it
Allow the same set of tests for all MIPS targets, so that "mipsel"
now gets some additional test coverage, too. While we're at it,
simplify the definitions for qtests_mips64 and qtests_mips64el.
Signed-off-by: Thomas Huth
---
tests/qtest/endianness-test.c | 1 +
tests/qtest/meson.build |
On Mittwoch, 13. April 2022 05:30:57 CEST Shi, Guohuai wrote:
> > We have 3 fs drivers: local, synth, proxy. I don't mind about proxy, it is
> > in bad shape and we will probably deprecate it in near future anyway.
> > But it would be good to have support for the synth driver, because we are
> > u
Add the iommufd backend. The IOMMUFD container class is implemented
based on the new /dev/iommu user API. This backend obviously depends
on CONFIG_IOMMUFD.
So far, the iommufd backend doesn't support live migration and
cache coherency yet due to missing support in the host kernel meaning
that only
From: Eric Auger
Let the vfio-ccw device use vfio_attach_device() and
vfio_detach_device(), hence hiding the details of the used
IOMMU backend.
Also now all the devices have been migrated to use the new
vfio_attach_device/vfio_detach_device API, let's turn the
legacy functions into static functi
From: Eric Auger
Let the vfio-ap device use vfio_attach_device() and
vfio_detach_device(), hence hiding the details of the used
IOMMU backend.
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
---
hw/vfio/ap.c | 62
1 file changed, 9 insertio
From: Eric Auger
This auto/on/off option allows the user to force a the select
the iommu BE (iommufd or legacy).
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
---
hw/vfio/pci.c | 26 ++
1 file changed, 26 insertions(+)
diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
inde
From: Eric Auger
Now we support two types of iommu backends, let's add the capability
to select one of them. This is based on a VFIODevice auto/on/off
iommu_be field. This field is likely to be forced to a given value or
set by a device option.
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
-
Qomify the VFIOContainer object which acts as a base class for a
container. This base class is derived into the legacy VFIO container
and later on, into the new iommufd based container.
The base class implements generic code such as code related to
memory_listener and address space management wher
Compared with legacy vfio container BE, one of the benefits provided by
iommufd is to reduce the redundant page pinning on kernel side through
the usage of IOAS_COPY_DMA. For iommufd containers within the same address
space, IOVA mappings can be copied from a source container to destination
contain
Introduce iommufd utility library which can be compiled out with
CONFIG_IOMMUFD configuration. This code is bound to be called by
several subsystems: vdpa, and vfio.
Co-authored-by: Eric Auger
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
---
MAINTAINERS | 7 ++
hw/Kconfig
From: Eric Auger
Let the vfio-platform device use vfio_attach_device() and
vfio_detach_device(), hence hiding the details of the used
IOMMU backend.
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
---
hw/vfio/platform.c | 42 ++
1 file changed, 2 insert
From: Eric Auger
We want the VFIO devices to be able to use two different
IOMMU callbacks, the legacy VFIO one and the new iommufd one.
Introduce vfio_[attach/detach]_device which aim at hiding the
underlying IOMMU backend (IOCTLs, datatypes, ...).
Once vfio_attach_device completes, the device
From: Eric Auger
Let's turn attach/detach_device as container callbacks. That way,
their implementation can be easily customized for a given backend.
For the time being, only the legacy container is supported.
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
---
hw/vfio/as.c
From: Eric Auger
Imported from https://github.com/luxis1999/iommufd/tree/iommufd-v5.17-rc6
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
---
linux-headers/linux/iommufd.h | 223 ++
linux-headers/linux/vfio.h| 84 +
2 files changed, 307 insert
From: Eric Auger
Using a VFIODevice handle local variable to improve the code readability.
no functional change intended
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
---
hw/vfio/pci.c | 49 +
1 file changed, 25 insertions(+), 24 deletions(-)
Rename VFIOGuestIOMMU iommu field into iommu_mr. Then it becomes clearer
it is an IOMMU memory region.
no functional change intended
Signed-off-by: Yi Liu
---
hw/vfio/common.c | 16
include/hw/vfio/vfio-common.h | 2 +-
2 files changed, 9 insertions(+), 9 deletion
From: Eric Auger
Reset implementation depends on the container backend. Let's
introduce a VFIOContainer class function and register a generic
reset handler that will be able to call the right reset function
depending on the container type. Also, let's move the
registration/unregistration to a pla
With the introduction of iommufd[1], the linux kernel provides a generic
interface for userspace drivers to propagate their DMA mappings to kernel
for assigned devices. This series does the porting of the VFIO devices
onto the /dev/iommu uapi and let it coexist with the legacy implementation.
Other
From: Eric Auger
Properly output the errno string.
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
---
hw/vfio/pci.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
index 67a183f17b..e26e65bb1f 100644
--- a/hw/vfio/pci.c
+++ b/hw/vfio/pci.c
@@
From: Eric Auger
Update the script to import iommufd.h
Signed-off-by: Eric Auger
Signed-off-by: Yi Liu
---
scripts/update-linux-headers.sh | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/scripts/update-linux-headers.sh b/scripts/update-linux-headers.sh
index 839a5ec614..a8
[+cc dgilb...@redhat.com stefa...@redhat.com]
On 4/14/22 1:05 PM, liuyd.f...@fujitsu.com wrote:
> [+cc vgo...@redhat.com]
>
> On 4/13/22 12:20 PM, Liu Yiding wrote:
>> Refer to 26ec190964 virtiofsd: Do not use a thread pool by default
>>
>> Signed-off-by: Liu Yiding
>> ---
>>docs/tools/virtio
Up to now the virt-machine node contains a virtio-mmio node.
However no driver produces any PCI interface node. Hence, PCI
tests cannot be run with aarch64 binary.
Add a GPEX driver node that produces a pci interface node. This latter
then can be consumed by all the pci tests. One of the first mot
ARM does not not support hotplug on pcie.0. Add a flag on the bus
which tells if devices can be hotplugged and skip hotplug tests
if the bus cannot be hotplugged. This is a temporary solution to
enable the other pci tests on aarch64.
Signed-off-by: Eric Auger
Acked-by: Thomas Huth
Reviewed-by: A
Up to now the virt-machine node only contains a virtio-mmio
driver node but no driver that eventually produces any pci-bus
interface.
Hence, PCI libqos tests cannot be run with aarch64 binary.
This series brings the pieces needed to be able to run PCI tests
with the aarch64 binary: a generic-pcih
1 - 100 of 119 matches
Mail list logo
