On 2024/02/21 17:15, Markus Armbruster wrote:
Akihiko Odaki writes:
vfio determines if rombar is explicitly enabled by inspecting QDict.
Inspecting QDict is not nice because QDict is untyped and depends on the
details on the external interface. Add an infrastructure to determine if
rombar is e
On 2024/02/21 16:59, Markus Armbruster wrote:
Akihiko Odaki writes:
Currently there is no way to distinguish the case that rombar is
explicitly specified as 1 and the case that rombar is not specified.
Set rombar -1 by default to distinguish these cases just as it is done
for addr and romsize
For consistency with other commands, and since it already
accepts --image-opts, allow specifying -f fmt too.
Signed-off-by: Michael Tokarev
Reviewed-by: Daniel P. Berrangé
---
docs/tools/qemu-img.rst | 2 +-
qemu-img-cmds.hx| 4 ++--
qemu-img.c | 9 ++---
3 files change
Initializing an unsigned as -1, or using temporary
sval for conversion is awkward. Since we don't allow
other "negative" values anyway, use signed value and
pass it to bdrv_img_create() (where it is properly
converted to unsigned), simplifying code.
Signed-off-by: Michael Tokarev
---
qemu-img.c
From: Thomas Huth
For distros like downstream RHEL, it would be helpful to allow to disable
the CompactFlash device. For making this possible, we need a separate
Kconfig switch for this device, and the code should reside in a separate
file. Let's also introduce a new header ide-dev.h which can be
Signed-off-by: Philippe Mathieu-Daudé
Reviewed-by: BALATON Zoltan
Reviewed-by: Thomas Huth
Message-Id: <20240215105017.57748-2-phi...@linaro.org>
---
hw/ppc/ppc440_pcix.c | 1 -
hw/ppc/ppc4xx_pci.c | 1 -
2 files changed, 2 deletions(-)
diff --git a/hw/ppc/ppc440_pcix.c b/hw/ppc/ppc440_pcix.c
From: Bernhard Beschow
The attribute isn't user-changeable and only true for pc-based machines. Turn it
into a class attribute which allows for inlining pc_guest_info_init() into
pc_machine_initfn().
Signed-off-by: Bernhard Beschow
Reviewed-by: Philippe Mathieu-Daudé
Message-ID: <2024020822034
From: Thomas Huth
These definitions are required outside of the hw/ide/ code, too,
so lets's move them from internal.h to a new header called ide-dma.h.
Signed-off-by: Thomas Huth
Acked-by: Mark Cave-Ayland
Reviewed-by: Philippe Mathieu-Daudé
Message-ID: <20240220085505.30255-4-th...@redhat.c
From: Philippe Mathieu-Daudé
Since the named GPIO lines are a "public" interface to the device,
we can directly call qdev_connect_gpio_out_named(), making it
consistent with how the other A20 input source (port92) is wired.
Suggested-by: Peter Maydell
Signed-off-by: Philippe Mathieu-Daudé
Revi
From: Bernhard Beschow
Fixes: fbd758008f0f "hw/isa: extract FDC37M81X to a separate file"
Signed-off-by: Bernhard Beschow
Reviewed-by: Mark Cave-Ayland
Reviewed-by: Philippe Mathieu-Daudé
Message-ID: <20240218131701.91132-2-shen...@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé
---
hw/isa
ppc440_pcix.c is moved from the target specific ppc_ss[] meson
source set to pci_ss[] which is common to all targets: the
object is built once.
Reviewed-by: Thomas Huth
Signed-off-by: Philippe Mathieu-Daudé
Message-Id: <20240215105017.57748-5-phi...@linaro.org>
---
MAINTAINERS
From: Thomas Huth
Let's consolidate the public IDE bus related functions in a separate
header.
Signed-off-by: Thomas Huth
Acked-by: Mark Cave-Ayland
Reviewed-by: Philippe Mathieu-Daudé
Message-ID: <20240220085505.30255-6-th...@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé
---
include/hw
From: Thomas Huth
include/hw/ide/internal.h is currently included by include/hw/ide/pci.h
and thus exposed to a lot of files that are not part of the IDE subsystem.
Stop including internal.h there and use the appropriate new headers
ide-bus.h and ide-dma.h instead.
Reviewed-by: Philippe Mathieu-
From: Thomas Huth
Untangle internal.h by moving public IDE device related
definitions to ide-dev.h.
Signed-off-by: Thomas Huth
Acked-by: Mark Cave-Ayland
Reviewed-by: Philippe Mathieu-Daudé
Message-ID: <20240220085505.30255-5-th...@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé
---
inclu
'qemu-img resize --help' does not work, since it wants more
arguments. Also it -size is only recognized as a very last
argument, but it is common for tools to handle other options
after positional arguments too.
Tell getopt_long() to return non-options together with options,
and process filename
From: Thomas Huth
There was only one prototype left in this legacy file. Move it to
ide-dev.h to finally get rid of it.
Reviewed-by: Philippe Mathieu-Daudé
Signed-off-by: Thomas Huth
Acked-by: Mark Cave-Ayland
Message-ID: <20240220085505.30255-7-th...@redhat.com>
Signed-off-by: Philippe Mathi
From: Peter Maydell
In the i386 PC machine, we want to run the pc_cmos_init_late()
function only once the IDE and floppy drive devices have been set up.
We currently do this using qemu_register_reset(), and then have the
function call qemu_unregister_reset() on itself, so it runs exactly
once.
T
also add short description to each command and use it in --help
Signed-off-by: Michael Tokarev
---
qemu-img.c | 41 ++---
1 file changed, 34 insertions(+), 7 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index ea284dca2d..299e34e470 100644
--- a/qemu-img.
qemu_opt_set_number() expects signed int64_t.
Use int64_t instead of uint64_t for img_size, use -1 as "unset"
value instead of UINT64_MAX, and do not require temporary sval
for conversion from string.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 19 +++
1 file changed, 7 inse
From: Joe Komlodi
The current logging doesn't tell us which specific smbus device is an
error state.
Signed-off-by: Joe Komlodi
Reviewed-by: Peter Maydell
Reviewed-by: Philippe Mathieu-Daudé
Message-ID: <20240202204847.2062798-3-koml...@google.com>
Signed-off-by: Philippe Mathieu-Daudé
---
When no -l/-a/-c/-d specified, assume -l (list).
Use the same values for SNAPSHOT_LIST/etc constants as the
option chars (lacd), this makes it possible to simplify
option handling a lot, combining cases for 4 options into
one.
Also remove bdrv_oflags handling (only list can use RO mode).
Signed-
Add missing long options and --help output.
Also add -b short option for --backing-chain, and remove
now-unused OPTION_BACKING_CHAIN.
While at it, remove unused option_index variable.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 40 +++-
1 file changed, 27
cvtnum() expects input string to specify some sort of size
(optionally with KMG... suffix). However, there are a lot
of other number conversions in there (using qemu_strtol &Co),
also, not all conversions which use cvtnum, actually expects
size, - like dd count=nn.
Add bool issize argument to cvt
Add missing long options and --help output.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 37 -
1 file changed, 28 insertions(+), 9 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index c8b0b68d67..45fbef5d37 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -4
Quite big patchset trying to implement normal, readable qemu-img --help
(and qemu-img COMMAND --help) output with readable descriptions, and
adding many long options in the process.
In the end I stopped using qemu-img-opts.hx in qemu-img.c, perhaps
this can be avoided, with only list of commands a
From: Bernhard Beschow
pc_system_flash_create() checked for pcmc->pci_enabled which is redundant since
its caller already checked it. The method can be turned into just two lines, so
inline and remove it.
Signed-off-by: Bernhard Beschow
Reviewed-by: Philippe Mathieu-Daudé
Message-ID: <20240208
Add missing long options and --help output.
convert uses -B for --backing, - why not -b?
Signed-off-by: Michael Tokarev
---
qemu-img.c | 90 --
1 file changed, 81 insertions(+), 9 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index fd61b2
Add missing long options and --help output.
While at it, remove unused option_index variable.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 34 --
1 file changed, 24 insertions(+), 10 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index cc51da31cf..3f719bbe
From: Bernhard Beschow
Rather than distributing PC system flash handling across three files, let's
confine it to one. Now, pc_system_firmware_init() creates, configures and cleans
up the system flash which makes the code easier to understand. It also avoids
the extra call to pc_system_flash_clean
From: Bernhard Beschow
Handling most of smbios data generation in the machine_done notifier is similar
to how the ARM virt machine handles it which also calls smbios_set_defaults()
there. The result is that all pc machines are freed from explicitly worrying
about smbios setup.
Signed-off-by: Ber
now once all individual subcommands has --help support, remove
the large unreadable help() thing and replace it with small
global --help, which refers to individual command --help for
more info.
While at it, also line-wrap list of formats after 75 chars.
Since missing_argument() and unrecognized_
Add missing long options and --help output.
Options added:
--format, --cache - for the image in question
--backing, --backing-format, --backing-cache, --backing-unsafe -
for the new backing file
(was eg CACHE vs SRC_CACHE, which is unclear).
Probably should rename local variables.
Signed-of
Add missing long options and --help output.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 39 +--
1 file changed, 29 insertions(+), 10 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index e4027ece20..af7841573c 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@
From: Thomas Huth
leon3.c currently fails to compile with some compilers when the -Wvla
option has been enabled:
../hw/sparc/leon3.c: In function ‘leon3_cpu_reset’:
../hw/sparc/leon3.c:153:5: error: ISO C90 forbids variable length array
‘offset_must_be_zero’ [-Werror=vla]
153 | ResetD
Signed-off-by: Michael Tokarev
---
qemu-img.c | 45 +
1 file changed, 37 insertions(+), 8 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index 1271217272..fd61b25ea7 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -1487,25 +1487,52 @@ static int img_com
From: Thomas Huth
qdev.c is a mixture between IDE bus specific functions and IDE device
functions. Let's split it up to make it more obvious which part is
related to bus handling and which part is related to device handling.
Reviewed-by: Philippe Mathieu-Daudé
Signed-off-by: Thomas Huth
Acked-
Add missing long options and --help output.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 40
1 file changed, 32 insertions(+), 8 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index 8455832d34..e4027ece20 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@
Reviewed-by: BALATON Zoltan
Reviewed-by: Thomas Huth
Signed-off-by: Philippe Mathieu-Daudé
Message-Id: <20240215105017.57748-3-phi...@linaro.org>
---
MAINTAINERS | 1 +
include/hw/pci-host/ppc4xx.h | 17 +
include/hw/ppc/ppc4xx.h | 5 -
hw/ppc/ppc440_
Add missing long options and --help output.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 38 ++
1 file changed, 30 insertions(+), 8 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index 01894c097b..69fa9701e9 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -
Add missing long options and --help output.
Also add -s short option for --size (and remove OPTION_SIZE).
Signed-off-by: Michael Tokarev
---
qemu-img.c | 53 -
1 file changed, 40 insertions(+), 13 deletions(-)
diff --git a/qemu-img.c b/qemu-i
Use common code and simplify error message
Signed-off-by: Michael Tokarev
Reviewed-by: Daniel P. Berrangé
---
qemu-img.c | 63 --
1 file changed, 18 insertions(+), 45 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index 7e4c993b9c..01894c0
Add missing long options and --help output.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 34 ++
1 file changed, 26 insertions(+), 8 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index 45fbef5d37..0d17738fb6 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -4557
From: Bernhard Beschow
Both invocations are the same and either one is always executed. Avoid this
redundancy.
Signed-off-by: Bernhard Beschow
Reviewed-by: Philippe Mathieu-Daudé
Message-ID: <20240208220349.4948-3-shen...@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé
---
hw/i386/pc_piix.c
From: Bernhard Beschow
The interrupt handlers need to be populated before the device is realized since
internal devices such as the RTC are wired during realize(). If the interrupt
handlers aren't populated, devices such as the RTC will be wired with a NULL
interrupt handler, i.e. MC146818RtcStat
From: Bernhard Beschow
Resolves redundant code in the piix and q35 machines.
Signed-off-by: Bernhard Beschow
Reviewed-by: Philippe Mathieu-Daudé
Message-ID: <20240208220349.4948-5-shen...@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé
---
include/hw/i386/pc.h | 2 --
hw/i386/pc.c |
In order to correctly print executable name in various
error messages, pass argv[0] to error_exit() function.
This way, error messages will refer to actual executable
name, which may be different from 'qemu-img'.
For subcommands, pass whole argv[] array, so argv[0] is
the executable name, not subc
From: Peter Maydell
Add the two IDE bus BusState pointers to the set we keep in PCMachineState.
This allows us to avoid passing them to pc_cmos_init(), and also will
allow a refactoring of how we call pc_cmos_init_late().
Signed-off-by: Peter Maydell
Acked-by: Richard Henderson
Reviewed-by: Mi
Add missing long options and --help output.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 45 -
1 file changed, 36 insertions(+), 9 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index ee35768af8..ce939708d4 100644
--- a/qemu-img.c
+++ b/qemu-img
Add missing long options and --help output.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 64 +-
1 file changed, 54 insertions(+), 10 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index 0d17738fb6..8455832d34 100644
--- a/qemu-img.c
+++
sysbus_add_io(...) is a simple wrapper to
memory_region_add_subregion(get_system_io(), ...).
It is used in 3 places; inline it directly.
Rationale: we want to move to an explicit I/O bus,
rather that an implicit one. Besides in heterogeneous
setup we can have more than one I/O bus.
Signed-off-by:
Add missing long options and --help output.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 44
1 file changed, 36 insertions(+), 8 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index eba13724b0..1271217272 100644
--- a/qemu-img.c
+++ b/qemu-img.
ppc4xx_pci.c is moved from the target specific ppc_ss[] meson
source set to pci_ss[] which is common to all targets: the
object is built once.
Declare PPC4XX_PCI selector in pci-host/Kconfig.
Signed-off-by: Philippe Mathieu-Daudé
Reviewed-by: Thomas Huth
Message-Id: <20240215105017.57748-4-phi.
The following changes since commit 760b4dcdddba4a40b9fa0eb78fdfc7eda7cb83d0:
Merge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging
(2024-02-20 10:11:08 +)
are available in the Git repository at:
https://github.com/philmd/qemu.git tags/hw-misc-20240221
This info will be used to generate --help output.
Signed-off-by: Michael Tokarev
---
qemu-img.c | 34 +-
1 file changed, 17 insertions(+), 17 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index 44dbf5be4f..38ac0f1845 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@
Create helper function cmd_help() to display command-specific
help text, and use it to print --help for 'create' subcommand.
Add missing long options (eg --format) in img_create().
Remove usage of missing_argument()/unrecognized_option() in
img_create().
Signed-off-by: Michael Tokarev
---
qemu
Signed-off-by: Michael Tokarev
Reviewed-by: Daniel P. Berrangé
---
qemu-img.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index 69fa9701e9..eba13724b0 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -858,8 +858,9 @@ static int img_check(const img_
Currently we have:
./qemu-img resize none +10
qemu-img: Could not open 'none': Could not open 'none': No such file or
directory
stop printing the message twice, - local_err already has
all the info, no need to prepend additional text there.
There are a few other places like this, but I'm un
On 21/2/24 20:13, Thomas Huth wrote:
On 21/02/2024 19.43, Philippe Mathieu-Daudé wrote:
On 20/2/24 09:55, Thomas Huth wrote:
Unentangle internal.h by moving public IDE device related
Disentangle ... from? Untangle? TIL Unentangle.
You're right, "disentangle" seems to be the more appropriate
On 21/02/2024 19.43, Philippe Mathieu-Daudé wrote:
On 20/2/24 09:55, Thomas Huth wrote:
Unentangle internal.h by moving public IDE device related
Disentangle ... from? Untangle? TIL Unentangle.
You're right, "disentangle" seems to be the more appropriate word.
I'll fix it up.
Anyway I'm no
On 20/2/24 09:55, Thomas Huth wrote:
Let's consolidate the public IDE bus related functions in a separate
header.
Signed-off-by: Thomas Huth
---
include/hw/ide/ide-bus.h | 42 +++
include/hw/ide/internal.h | 40 +
2 fi
On 20/2/24 09:55, Thomas Huth wrote:
Unentangle internal.h by moving public IDE device related
Disentangle ... from? Untangle? TIL Unentangle.
Anyway I'm not a native English speaker, so:
Reviewed-by: Philippe Mathieu-Daudé
definitions to ide-dev.h.
Signed-off-by: Thomas Huth
---
includ
On 20/2/24 09:55, Thomas Huth wrote:
These definitions are required outside of the hw/ide/ code, too,
so lets's move them from internal.h to a new header called ide-dma.h.
Signed-off-by: Thomas Huth
---
include/hw/ide/ide-dma.h | 37 +
include/hw/ide/inte
From: Felix Wu
Signed-off-by: Felix Wu
Signed-off-by: Nabih Estefan
---
hw/smbios/smbios.c | 49 +---
include/hw/firmware/smbios.h | 4 +++
qemu-options.hx | 2 +-
3 files changed, 51 insertions(+), 4 deletions(-)
diff --git a/hw/smbios
From: Roque Arcudia Hernandez
This patch adds a way to specify an NGUID for a given NVMe Namespace using a
string of hexadecimal digits with an optional '-' separator to group bytes. For
instance:
-device nvme-ns,nguid="e9accd3b83904e13167cf0593437f57d"
If provided, the NGUID will be part of th
This patch series modifies the ARM GICv3 ITS to use the already existing
send_msi virtual function when writing the GITS_TRANSLATER in order to be able
to modify the final DeviceID to an implementation specific version that requires
extra information besides the BDF that comes in the requester_id.
From: Roque Arcudia Hernandez
According to the SMMU specification the StreamID construction and size is
IMPLEMENTATION DEFINED, the size being between 0 and 32 bits.
This patch creates virtual functions get_sid and get_iommu_mr to allow different
implementations of how the StreamID is constructe
This patch series modifies the ARM SMMUv3 to be able to work with an
implementation specific StreamID that does not match exactly the PCIe BDF.
The way to achieve this is by converting the smmu_get_sid and smmu_iommu_mr
functions to virtual functions that can be overridden by inheritance, making
su
From: Roque Arcudia Hernandez
Current implementation checks the StreamIDs against STRTAB_BASE_CFG.LOG2SIZE
register field value and a constant SMMU_IDR1_SIDSIZE which is also used as
initial value for field SMMU_IDR1.SIDSIZE.
This limits the possibility of extending the SMMUv3 by inheritance and
From: Roque Arcudia Hernandez
According to the “GICv3 and GICv4 Software Overview” the DeviceID is
IMPLEMENTATION DEFINED. This patch increases the DeviceID in send_msi virtual
function to 32 bits to allow the possibility of a redefinition of it with a
bigger DeviceID.
Signed-off-by: Roque Arcud
From: Roque Arcudia Hernandez
This is trying to achieve 2 things: To be able to redefine the send_msi in a
derived class of arm_gicv3_its and/or to expose a method call interface to
inject interrupts from another device.
Signed-off-by: Roque Arcudia Hernandez
Signed-off-by: Nabih Estefan
---
This patch series implements SMBIOS type 9 descriptor, system slots.
For each system slot, we can assign one descriptor for it if needed.
In versions later than 2.6, a new PCI device field was added to make sure the
descriptor is associated with a certain device, if provided.
For ease of usage, qem
From: Felix Wu
Version 2.1+.
Signed-off-by: Felix Wu
Signed-off-by: Nabih Estefan
---
hw/smbios/smbios.c | 99
include/hw/firmware/smbios.h | 13 +
qemu-options.hx | 3 ++
3 files changed, 115 insertions(+)
diff --git a/hw/smbi
On Wed, Feb 21, 2024 at 07:31:42PM +0300, Michael Tokarev wrote:
> 20.02.2024 21:48, Daniel P. Berrangé:
>
> > This ends up looking a bit muddled together. I don't think we
> > need repeat 'qemu-img ' twice, and could add a little
> > more whitespace
> >
> > eg instead of:
> >
> > $ ./build/qemu
20.02.2024 21:48, Daniel P. Berrangé:
This ends up looking a bit muddled together. I don't think we
need repeat 'qemu-img ' twice, and could add a little
more whitespace
eg instead of:
$ ./build/qemu-img check --help
qemu-img check: Check basic image integrity. Usage:
qemu-img check [-f FMT |
20.02.2024 21:51, Michael Tokarev wrote:
20.02.2024 20:45, Daniel P. Berrangé wrote:
On Sat, Feb 10, 2024 at 12:22:33AM +0300, Michael Tokarev wrote:
also remove bdrv_oflags handling (only list can use RO mode)
---
qemu-img.c | 13 -
1 file changed, 8 insertions(+), 5 deletions(-
Am 21.02.24 um 07:55 schrieb Markus Armbruster:
>> diff --git a/qapi/block-core.json b/qapi/block-core.json
>> index ab5a93a966..ac05483958 100644
>> --- a/qapi/block-core.json
>> +++ b/qapi/block-core.json
>> @@ -2181,6 +2181,15 @@
>> # destination (all the disk, only the sectors allocated in
On 20/02/2024 08:54, Thomas Huth wrote:
While trying to make it possible to compile-out the CompactFlash IDE device
in downstream distributions (first patch), we noticed that there are more
things in the IDE code that could use a proper clean up:
First, hw/ide/qdev.c is quite a mix between IDE
20.02.2024 20:57, Daniel P. Berrangé пишет:
On Sat, Feb 10, 2024 at 12:22:36AM +0300, Michael Tokarev wrote:
'qemu-img resize --help' does not work, since it wants more arguments.
Only eat last option at the beginning if it starts like -N.., and allow
getopt() to do its work, and eat it up at th
Akihiko Odaki writes:
> It is no longer used.
>
> Signed-off-by: Akihiko Odaki
> Reviewed-by: Philippe Mathieu-Daudé
> ---
> include/hw/qdev-core.h | 4
> hw/core/qdev.c | 1 -
> system/qdev-monitor.c | 12 +++-
> 3 files changed, 7 insertions(+), 10 deletions(-)
>
> di
Akihiko Odaki writes:
> Use pci_rom_bar_explicitly_enabled() to determine if rombar is explicitly
> enabled.
>
> Signed-off-by: Akihiko Odaki
> ---
> hw/vfio/pci.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
> index 4fa387f0430d..647f
Akihiko Odaki writes:
> vfio determines if rombar is explicitly enabled by inspecting QDict.
> Inspecting QDict is not nice because QDict is untyped and depends on the
> details on the external interface. Add an infrastructure to determine if
> rombar is explicitly enabled to hw/pci.
>
> Signed-o
Akihiko Odaki writes:
> Currently there is no way to distinguish the case that rombar is
> explicitly specified as 1 and the case that rombar is not specified.
>
> Set rombar -1 by default to distinguish these cases just as it is done
> for addr and romsize. It was confirmed that changing the def
82 matches
Mail list logo
