adding John Snow to CC because he investigated this in 2020.
On Fri, 12 Apr 2024, Eric Blake wrote:
> On Fri, Apr 12, 2024 at 10:06:17AM +0200, Stefan Fritsch wrote:
> > Commit 99868af3d0 changed the hardcoded constant BDRV_SECTOR_SIZE to a
> > dynamic field 'align' but i
in uncommon configurations, e.g. with
icount enabled or when reading from disk directly to device memory.
Fixes: 99868af3d0a75cf6 ("dma-helpers: explicitly pass alignment into DMA
helpers")
Signed-off-by: Stefan Fritsch
---
system/dma-helpers.c | 3 +--
1 file changed, 1 insertion(+), 2
On Thu, 20 Jul 2017, Stefan Fritsch wrote:
> Thanks for the review.
>
> On Thu, 20 Jul 2017, Marc-André Lureau wrote:
>
> > On Thu, Jul 20, 2017 at 11:04 AM Stefan Fritsch <s...@sfritsch.de> wrote:
> >
> > > From: Stefan Fritsch <stefan_frit...@genua.d
On Fri, 21 Jul 2017, Gerd Hoffmann wrote:
> > index bef1f03c42..624dc2c447 100644
> > --- a/hw/usb/dev-smartcard-reader.c
> > +++ b/hw/usb/dev-smartcard-reader.c
> > @@ -54,9 +54,25 @@ do { \
> > #define D_INFO 2
> > #define D_MORE_INFO 3
> > #define D_VERBOSE 4
> > +#define D_TRACE 5
> >
On Thu, 20 Jul 2017, Eric Blake wrote:
> On 07/20/2017 04:31 AM, Stefan Fritsch wrote:
> > From: Stefan Fritsch <stefan_frit...@genua.de>
> >
> > Add an ACPI sleep button and QMP/HMP commands to trigger it. A sleep
> > button is a so called "fixed
Thanks for the review.
On Thu, 20 Jul 2017, Marc-André Lureau wrote:
> On Thu, Jul 20, 2017 at 11:04 AM Stefan Fritsch <s...@sfritsch.de> wrote:
>
> > From: Stefan Fritsch <stefan_frit...@genua.de>
> >
> > Change all DPRINTF()s using (1 == WARN)
On Thu, 20 Jul 2017, Marc-André Lureau wrote:
> On Thu, Jul 20, 2017 at 11:09 AM Stefan Fritsch <s...@sfritsch.de> wrote:
>
> > From: Stefan Fritsch <stefan_frit...@genua.de>
> >
> > - We want to support both T=0 and T=1. Additionally, note that all
> &
From: Stefan Fritsch <stefan_frit...@genua.de>
Add an ACPI sleep button and QMP/HMP commands to trigger it. A sleep
button is a so called "fixed hardware feature", which makes it more
suitable for putting the system to sleep than a laptop lid, for example.
The sleep bu
From: Stefan Fritsch <stefan_frit...@genua.de>
For the ATR interface bytes encoding see ISO/IEC 7816-3. For the
interpretation of the protocol parameter block see the CCID
specification.
- Values that are not present in the ATR default to zero.
- TA_1 is used to fill bmFindexDindex.
- Th
From: Stefan Fritsch <stefan_frit...@genua.de>
- We want to support both T=0 and T=1. Additionally, note that all
fields in USB descriptors are little endian and the supported
protocols are in the lowest byte the of the 32-Bit dwProtocols field.
- We want APDU level exchanges. This sa
From: Stefan Fritsch <stefan_frit...@genua.de>
Incoming packets should never need chaining (bChainParamter => 0).
If they ever do we will have to fix this later. Zero is still better
than an uninitialized value.
Warn if outgoing packets try to use chaining (wLevelParameter != 0).
As we
From: Stefan Fritsch <stefan_frit...@genua.de>
The dump can be activated by the debug command line option for the
device like this "qemu ... -dev usb,ccid,debug=5"
While there move the short read debug message to a higher debug level.
It triggers very often and makes debug o
From: Stefan Fritsch <stefan_frit...@genua.de>
The emulated ccid smartcard reader has a fixed limit for APDUs from the
card. Increase this to 1024. We have seen software that tries to use 768
byte APDUs.
Signed-off-by: Stefan Fritsch <stefan_frit...@genua.de>
---
hw/usb/dev-smartc
From: Stefan Fritsch <stefan_frit...@genua.de>
Change all DPRINTF()s using (1 == WARN) to use symbolic
constants. Most of these DPRINTFs are now only logging at higher log
levels.
This allows to use ccid's debug level 1 == WARN in normal operation.
Signed-off-by: Stefan Fritsch <st
From: Stefan Fritsch <stefan_frit...@genua.de>
Sorry, I have accidentally sent an old patchset that was missing one
fix.
This set of changes contains fixes and extensions to allow to use ccid
card passthru with a physical cardos smartcard using the T=1 protocol.
Unfortunately, our vscard
From: Stefan Fritsch <stefan_frit...@genua.de>
- The number of parameter set TA_i...TD_i is unlimited. The list ends
if TD_i is not present or the high nibble is zero.
- If at least one protocol in any of the TD bytes is non-zero, the
ATR must have a checksum.
- Add a missing length
From: Stefan Fritsch <stefan_frit...@genua.de>
An USB transfer must be terminated by a packet that is shorter than
the maximum USB packet size (usually 64Byte) reported by the device
in its descriptor(s). Thus if the last packet of a transfer is exactly
64-Bytes long, we must send an addi
From: Stefan Fritsch <stefan_frit...@genua.de>
- We want to support both T=0 and T=1. Additionally, note that all
fields in USB descriptors are little endian and the supported
protocols are in the lowest byte the of the 32-Bit dwProtocols field.
- We want APDU level exchanges. This sa
From: Stefan Fritsch <stefan_frit...@genua.de>
Incoming packets should never need chaining (bChainParamter => 0).
If they ever do we will have to fix this later. Zero is still better
than an uninitialized value.
Warn if outgoing packets try to use chaining (wLevelParameter != 0).
As we
From: Stefan Fritsch <stefan_frit...@genua.de>
Change all DPRINTF()s using (1 == WARN) to use symbolic
constants. Most of these DPRINTFs are now only logging at higher log
levels.
This allows to use ccid's debug level 1 == WARN in normal operation.
Signed-off-by: Stefan Fritsch <st
From: Stefan Fritsch <stefan_frit...@genua.de>
For the ATR interface bytes encoding see ISO/IEC 7816-3. For the
interpretation of the protocol parameter block see the CCID
specification.
- Values that are not present in the ATR default to zero.
- TA_1 is used to fill bmFindexDindex.
- Th
From: Stefan Fritsch <stefan_frit...@genua.de>
An USB transfer must be terminated by a packet that is shorter than
the maximum USB packet size (usually 64Byte) reported by the device
in its descriptor(s). Thus if the last packet of a transfer is exactly
64-Bytes long, we must send an addi
From: Stefan Fritsch <stefan_frit...@genua.de>
- The number of parameter set TA_i...TD_i is unlimited. The list ends
if TD_i is not present or the high nibble is zero.
- If at least one protocol in any of the TD bytes is non-zero, the
ATR must have a checksum.
- Add a missing length
From: Stefan Fritsch <stefan_frit...@genua.de>
The dump can be activated by the debug command line option for the
device like this "qemu ... -dev usb,ccid,debug=5"
While there move the short read debug message to a higher debug level.
It triggers very often and makes debug o
From: Stefan Fritsch <stefan_frit...@genua.de>
The emulated ccid smartcard reader has a fixed limit for APDUs from the
card. Increase this to 1024. We have seen software that tries to use 768
byte APDUs.
Signed-off-by: Stefan Fritsch <stefan_frit...@genua.de>
---
hw/usb/dev-smartc
From: Stefan Fritsch <stefan_frit...@genua.de>
This set of changes contains fixes and extensions to allow to use ccid
card passthru with a physical smartcard using the T=1 protocol.
Unfortunately, our vscard client program ist not public, therefore it
is difficult for other people t
From: Stefan Fritsch <stefan_frit...@genua.de>
One string is longer than 40 chars. Set the field width to 50.
Signed-off-by: Stefan Fritsch <stefan_frit...@genua.de>
---
scripts/kvm/vmxcap | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/scripts/kvm/vmxcap b
On Wed, 19 Jul 2017, Thomas Huth wrote:
> On 11.06.2017 09:48, Stefan Fritsch wrote:
> > Sending a break on a serial console can be useful for debugging the
> > guest. But not all chardev backends support sending breaks (only telnet
> > and mux do). The chardev-send-break c
On Thursday, 29 June 2017 16:00:54 CEST Dr. David Alan Gilbert wrote:
> * Eric Blake (ebl...@redhat.com) wrote:
> > On 06/11/2017 02:48 AM, Stefan Fritsch wrote:
> > > Sending a break on a serial console can be useful for debugging the
> > > guest. But not all chard
Sending a break on a serial console can be useful for debugging the
guest. But not all chardev backends support sending breaks (only telnet
and mux do). The chardev-send-break command allows to send a break even
if using other backends.
Signed-off-by: Stefan Fritsch <s...@sfritsch.de>
On Tue, 6 Jun 2017, Markus Armbruster wrote:
> >> Is there an obvious test that we can enhance to add coverage of the new
> >> QMP command?
> >
> > You could have a new test covering hw/char/serial.c, but I wouldn't let
> > that hold the patch.
>
> Holding patches is pretty much the only leverage
Sending a break on a serial console can be useful for debugging the
guest. But not all chardev backends support sending breaks (only telnet
and mux do). The chardev-send-break command allows to send a break even
if using other backends.
Signed-off-by: Stefan Fritsch <s...@sfritsch
Hi,
Am Dienstag, 25. Februar 2014, 11:06:33 schrieb Amos Kong:
On Wed, Feb 12, 2014 at 10:46:28PM +0100, Stefan Fritsch wrote:
If VIRTIO_NET_F_CTRL_VLAN is not negotiated, do not filter out
all
VLAN-tagged packets but send them to the guest.
AFAICS, no fix has been committed, yet
On Fri, 21 Feb 2014, Amos Kong wrote:
On Wed, Feb 12, 2014 at 10:46:28PM +0100, Stefan Fritsch wrote:
If VIRTIO_NET_F_CTRL_VLAN is not negotiated, do not filter out all
VLAN-tagged packets but send them to the guest.
Can we just update receive_filter() to filter out VLAN-tagged packets
On Sun, 16 Feb 2014, Michael S. Tsirkin wrote:
On Wed, Feb 12, 2014 at 10:46:28PM +0100, Stefan Fritsch wrote:
If VIRTIO_NET_F_CTRL_VLAN is not negotiated, do not filter out all
VLAN-tagged packets but send them to the guest.
Signed-off-by: Stefan Fritsch s...@sfritsch.de
Thanks
If VIRTIO_NET_F_CTRL_VLAN is not negotiated, do not filter out all
VLAN-tagged packets but send them to the guest.
Signed-off-by: Stefan Fritsch s...@sfritsch.de
---
This time CCing the maintainers.
This fixes VLANs with OpenBSD guests (and probably NetBSD, too, because
the OpenBSD driver
Hi,
if the feature VIRTIO_NET_F_CTRL_VLAN is not negotiated, virtio-net should
not filter VLANs. That means it should send all packets to the guest
instead of dropping all packets that have any VLAN id. The following patch
fixes the issue.
Cheers,
Stefan
--- a/hw/net/virtio-net.c
+++
37 matches
Mail list logo