Hello QEMU-DEVEL! It's been a while…
I work on a solution for "smart" IO caching on the host side.
The configuration is virtio-block device backed by SPDK with OCF/OpenCAS
on top of remote storage.
To improve decision making for caching, I would like to have an additional
per-request contextual i
e host. This patch replaces the
> affected assert() with a conditional statement, returning false if the current
> data fragment exceeds max_raw_frags.
>
Reviewed-by: Dmitry Fleytman
> Reported-by: Alexander Bulekov
> Reported-by: Ziming Zhang
> Signed-off-by: Mauro Matteo Casce
Reviewed-by: Dmitry Fleytman
The idea looks good to me. I believe it makes sense to do the check in
net_tx_pkt_add_raw_fragment() as suggested by Jason.
> On 27 Jul 2020, at 20:29, Alexander Bulekov wrote:
>
> I sent a reproducer for the to the list some time ago, but never cr
te with 1 NetClientState(index 0) without peers, set
> max_queue_num to 0 - It prevents undefined behavior and possible crashes,
> especially during pcie hotplug.
>
> Signed-off-by: Andrew Melnychenko
Reviewed-by: Dmitry Fleytman
> ---
> hw/net/e1000e.c | 2 +-
> 1 file cha
> On 5 Mar 2020, at 3:04, Philippe Mathieu-Daudé wrote:
>
> Express the macreg[] arrays using typedefs.
> No logical changes introduced here.
>
> Signed-off-by: Philippe Mathieu-Daudé
Reviewed-by: Dmitry Fleytman
> ---
> hw/net/e1000.c | 6 --
&g
-Daudé
Reviewed-by: Dmitry Fleytman
> ---
> hw/net/e1000.c | 4 ++--
> hw/net/e1000e_core.c | 4 ++--
> 2 files changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/hw/net/e1000.c b/hw/net/e1000.c
> index 972d9b5083..9233248c9a 100644
> --- a/hw/net/e1000.
1 NetClientState(index 0) without peers, set
> max_queue_num to 0 - It prevents undefined behavior and possible crashes,
> especially during pcie hotplug.
>
> Signed-off-by: Andrew Melnychenko
Reviewed-by: Dmitry Fleytman
> ---
> hw/net/e1000e.c | 2 +-
> 1 file changed,
sted.
> Use this type in e1000e hash calculation for IPv6 TCP, which
> should take in account overrides of the addresses.
>
> Signed-off-by: Yuri Benditovich
Acked-by: Dmitry Fleytman
> ---
> hw/net/e1000e_core.c | 2 +-
> hw/net/net_rx_pkt.c | 2 +-
> 2 files changed, 2 insert
> On 27 Jan 2020, at 13:54, Yuri Benditovich
> wrote:
>
> Add support for following hash types:
> IPV6 TCP with extension headers
> IPV4 UDP
> IPV6 UDP
> IPV6 UDP with extension headers
>
> Signed-off-by: Yuri Benditovich
Acked-by: Dmitry Fleytman
&g
of RCTL register define legacy format of
> transfer descriptors. Current commit discards check
> for BSIZE0 and BSIZE1 when legacy mode used.
>
> Signed-off-by: Yuri Benditovich
Acked-by: Dmitry Fleytman
> ---
> hw/net/e1000e_core.c | 13 -
> 1 file changed, 8
> On 22 Aug 2019, at 14:12, Dr. David Alan Gilbert (git)
> wrote:
>
> From: "Dr. David Alan Gilbert"
>
> Commit 78dd48df3 reworked vmxnet3's live migration but left a straggling
> unregister_savevm call. Remove it, although it doesn't seem to have
Reviewed-by: Dmitry Fleytman
> On 8 Aug 2019, at 17:34, Philippe Mathieu-Daudé wrote:
>
> This is a preparatory cleanup series.
>
> Commit 75020a70215 introduced 4 very equivalent structures:
> - tcp_header and tcp_hdr,
> - udp_header and udp_hdr.
>
> Choose
> On 8 Jul 2019, at 19:03, Dr. David Alan Gilbert wrote:
>
> * Marcel Apfelbaum (marcel.apfelb...@gmail.com) wrote:
>>
>>
>>> On 7/5/19 2:14 PM, Sukrit Bhatnagar wrote:
>>>> On Fri, 5 Jul 2019 at 16:29, Dmitry Fleytman
>>>> wrote
,
> then the above macro is enough.
>
> Use the opportunity to move to the standard VMSTATE_MSIX
> instead of the deprecated SaveVMHandlers.
>
> Signed-off-by: Marcel Apfelbaum
Reviewed-by: Dmitry Fleytman
> ---
> hw/net/vmxnet3.c | 52 ++---
> On 25 Jun 2019, at 11:49, Marcel Apfelbaum wrote:
>
> Hi Dmitry,
>
> On 6/25/19 11:39 AM, Dmitry Fleytman wrote:
>>
>>> On 25 Jun 2019, at 11:14, Marcel Apfelbaum
>>> wrote:
>>>
>>> Hi Sukrit
>>>
>>> On 6/21/19
> On 25 Jun 2019, at 11:14, Marcel Apfelbaum wrote:
>
> Hi Sukrit
>
> On 6/21/19 5:45 PM, Sukrit Bhatnagar wrote:
>> Hi,
> [...]
>> This RFC is meant to request suggestions on the things which are
>> working and for help on the things which are not.
>>
> [...]
>> What is not working:
>>
> [
>
> hw/net/e1000e_core.h
> hw/net/e1000x_common.h
> hw/net/vmxnet3_defs.h
Unintentional.
Reviewed-by: Dmitry Fleytman
> On 28 May 2019, at 19:40, Philippe Mathieu-Daudé wrote:
>
> Since the BusState is accesible from the SCSIBus object,
> it is pointless to use qbus_reset_all_fn.
> Use qbus_reset_all() directly.
>
> Signed-off-by: Philippe Mathieu-Daudé
Nice catch!
Reviewed-by: Dmitry Fleytman mailto:dmitry.fleyt...@gmail.com>>
> On 21 Mar 2019, at 3:35, yuchenlin wrote:
>
> Ping?
>
> On 2019-03-13 14:56, yuchen...@synology.com wrote:
>> From: yuchenlin
>> Due to too early RCT0 interrput, win10x32 may han
ter assigned
> to the slot (as these events do not have TRB), so current
> commit adds the field of assigned interrupter to the
> XHCISlot structure. Guest software assigns interrupter to the
> slot on 'Address Device' and 'Evaluate Context' commands.
>
> Sig
ed-off-by: Yuri Benditovich
Reviewed-by: Dmitry Fleytman
> ---
> hw/usb/hcd-xhci.c | 5 +
> 1 file changed, 5 insertions(+)
>
> diff --git a/hw/usb/hcd-xhci.c b/hw/usb/hcd-xhci.c
> index 8f1a01a..1a8fd96 100644
> --- a/hw/usb/hcd-xhci.c
> +++ b/hw/usb/hcd-xhci.c
tor vmxnet3 device state.
> 3. To configure node_guid accoring to vmxnet3 device's MAC address.
>
> To be able to access vmxnet3 device the definition of VMXNET3State is
> moved to a new header file.
>
> Signed-off-by: Yuval Shaia
Reviewed-by: Dmitry Fleytm
Hi Dave,
I’m not sure e1000e was ever tested with Slirp. It might be a bug existing from
the beginning.
Dmitry
> On 6 Aug 2018, at 22:14, Dr. David Alan Gilbert wrote:
>
> Hi Sam, Jan, Dmitry,
> Any idea where this is coming from; my backtrace shows it's a free in
> slirp, but only happenin
> On 16 Jan 2018, at 8:28, Jason Wang wrote:
>
>
>
> On 2018年01月16日 10:48, Michael S. Tsirkin wrote:
>> On Tue, Jan 09, 2018 at 12:10:10PM +1100, David Gibson wrote:
>>> On Mon, Jan 08, 2018 at 08:10:23PM +0200, Michael S. Tsirkin wrote:
On Mon, Jan 08, 2018 at 09:14:41AM +, Peter Ma
; F: tests/megasas-test.c
>
> Network packet abstractions
> -M: Dmitry Fleytman
> +M: Dmitry Fleytman
> S: Maintained
> F: include/net/eth.h
> F: net/eth.c
> @@ -1165,7 +1165,7 @@ F: hw/net/net_rx_pkt*
> F: hw/net/net_tx_pkt*
>
> Vmware
> -M: Dmitry Fleytman
&
Reviewed-by: Dmitry Fleytman
> On 13 Nov 2017, at 18:57 PM, Thomas Huth wrote:
>
> Since commit ab06ec43577177a442e8 we test the vmxnet3 device in the
> pxe-tester, too (when running "make check SPEED=slow"). This now
> revealed that the vmxnet3 code is not working r
or -95
> [vmxnet3][WR][vmxnet3_pci_realize]: Failed to initialize MSI-X, configuration
> is inconsistent.
>
> Since vmxnet3 is a para-virtualized device that is only useful on x86,
> it should also only be enabled on the x86 targets.
>
Reviewed-by: Dmitry Fleytman
> Signed-of
Reviewed-by: Dmitry Fleytman
> On 4 Sep 2017, at 10:23 AM, Fam Zheng wrote:
>
> On Sun, 09/03 18:37, Kamil Rytarowski wrote:
>> SunOS defines SEC in as 1 (commonly used time symbols).
>>
>> This fixes build on SmartOS (Joyent).
>>
>> Patch cherr
Hi Gerd, qemu-devel
We are trying to use USB host device assignment to propagate
Logitech H340 headset to QEMU guest with virtual XHCI controller
running Windows 10.
For some reason headphone function of the device does not work -
guest OS claims that the device does not complete requests.
We ha
> On 2 Aug 2017, at 24:24 AM, Alex Williamson
> wrote:
>
> On Fri, 28 Jul 2017 08:22:45 +0300
> Dmitry Fleytman mailto:dmi...@daynix.com>> wrote:
>
>>> On 28 Jul 2017, at 07:51 AM, Zhang, Xiong Y wrote:
>>>
>>>>>
couple of such entries already
> in the MAINTAINERS file.
>
>> MAINTAINERS | 1 +
>> 1 file changed, 1 insertion(+)
>>
>> diff --git a/MAINTAINERS b/MAINTAINERS
>> index d8aef164ca..9670a3901a 100644
>> --- a/MAINTAINERS
>> +++ b/MAINTAINERS
>>
> On 28 Jul 2017, at 07:51 AM, Zhang, Xiong Y wrote:
>
>>> On 26 Jul 2017, at 08:22 AM, Zhang, Xiong Y
>> wrote:
>>>
>>> Sorry, we indeed found Intel windows guest graphic driver couldn't be bind
>> when GMS memory size is zero. And we have fixed it and the next intel
>> windows driver release
f(opthdr),
> src_addr, sizeof(*src_addr));
>
> -return bytes_read == sizeof(src_addr);
> +return bytes_read == sizeof(*src_addr);
> }
>
Reviewed-by: Dmitry Fleytman
> opt_offset += optlen;
> --
> 2.13.3
>
> On 26 Jul 2017, at 11:41 AM, Mao Zhongyi wrote:
>
> qemu_bh_new() is a wrapper around aio_bh_new(), which returns
> null only when g_new() does. It doesn't. So remove the dead
> error handling.
>
Acked-by: Dmitry Fleytman mailto:dmi...@daynix.com>>
>
> On 26 Jul 2017, at 11:41 AM, Mao Zhongyi wrote:
>
> Convert a device model where initialization obviously
> can't fail, make it implement realize() rather than init().
Acked-by: Dmitry Fleytman mailto:dmi...@daynix.com>>
>
> Cc: Dmitry Fleytman
> Cc: Paolo
gt; -Original Message-
>> From: Alex Williamson [mailto:alex.william...@redhat.com]
>> Sent: Tuesday, July 25, 2017 11:32 PM
>> To: Dmitry Fleytman
>> Cc: qemu-devel@nongnu.org; Zhang, Xiong Y
>> Subject: Re: [PATCH] vfio/pci-quirks: Set non-zero GMS memo
-by: Dmitry Fleytman
---
hw/vfio/pci-quirks.c | 13 +
hw/vfio/pci.c| 2 +-
2 files changed, 10 insertions(+), 5 deletions(-)
diff --git a/hw/vfio/pci-quirks.c b/hw/vfio/pci-quirks.c
index 349085e..197cb90 100644
--- a/hw/vfio/pci-quirks.c
+++ b/hw/vfio/pci-quirks.c
@@ -1527,10
On 17 Jul 2017, at 19:59 PM, Michael S. Tsirkin wrote:
>
> On Mon, Jul 17, 2017 at 03:48:03PM +0300, Dmitry Fleytman wrote:
>> Am I understand correctly that there are no special cases for
>> IDE controllers, i.e. bus master bit must be set by SW same
>> way as for oth
On 16 Jul 2017, at 19:56 PM, Marcel Apfelbaum wrote:
>
> On 16/07/2017 11:29, Dmitry Fleytman wrote:
>> According to PCI spec. bit 1 of command
>> register (PCI_COMMAND_MEMORY) controls
>> a device's response to memory space accesses.
>> A value of 0 disables t
0aa00, Incorrect checksum
Arg4: 0418, Previous disk read's checksum
According to our tests this problem happens at least on
Windows 8/8.1/2012/2012R2/10/2016.
This patch solves https://bugzilla.redhat.com/show_bug.cgi?id=988351
Signed-off-by: Dmitry Fleytman
---
to free it then.
>
> Signed-off-by: Laurent Vivier
> ---
> hw/net/vmxnet3.c | 8 ++—
Acked-by: Dmitry Fleytman mailto:dmi...@daynix.com>>
> hw/s390x/s390-skeys.c | 9 +++--
> hw/s390x/s390-virtio-ccw.c | 8 ++--
> include/migration/vms
, IMS: 0xa4)
> 27563@1494850819.412441:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa4)
> 27563@1494850819.412998:e1000e_irq_pending_interrupts ICR PENDING: 0x100
> (ICR: 0x815000c2, IMS: 0x1a4)
>
> This commit solves:
> https://bugzi
Hello Dave,
It looks like we identified the problem.
We are working on fix and will send it as soon as it is ready.
~Dmitry.
Sent from my iPhone
> On 15 May 2017, at 12:22, Dr. David Alan Gilbert wrote:
>
> * Dmitry Fleytman (dmi...@daynix.com) wrote:
>> Hello Dave,
Hello Dave,
We are trying to reproduce this issue on our systems but with no luck so far…
From what you describe it looks like some bit in ICR is not being cleared by
the driver.
This usually means that this bit should never be set in that specific interrupt
mode.
Could you please check which
loop. Add check to avoid it.
Hello P J P,
Except the commit message that does not exactly correspond to what this patch
does,
Reviewed-by: Dmitry Fleytman
Thanks for the patch,
Dmitry.
>
> Reported-by: YY Z
> Signed-off-by: P J P
> ---
> hw/scsi/vmw_pvscsi.c | 2 +-
> 1 file
Reviewed-by: Dmitry Fleytman
> On 13 Mar 2017, at 21:55 PM, Eric Blake wrote:
>
> An upcoming patch will let the compiler warn us when we are silently
> losing precision in traces; update the trace definitions to pass
> through the full value at the callsite. In one case, t
> On 3 Mar 2017, at 18:52 PM, Philippe Mathieu-Daudé wrote:
>
> On 02/16/2017 09:29 AM, Dmitry Fleytman wrote:
>> Make VLAN stripping functions return number of bytes
>> copied to given Ethernet header buffer.
>>
>> This information should be used to re
> On 3 Mar 2017, at 18:39 PM, Philippe Mathieu-Daudé wrote:
>
> Hi Dmitry,
>
> On 02/16/2017 09:29 AM, Dmitry Fleytman wrote:
>> In case of VLAN stripping ETH header is stored in a
>> separate chunk and length of IOV should take this into
>> account.
>>
> On 17 Feb 2017, at 5:05, Jason Wang wrote:
>
>
>
>> On 2017年02月16日 20:39, Peter Maydell wrote:
>>> On 16 February 2017 at 12:29, Dmitry Fleytman wrote:
>>> This series fix a few issues related
>>> to processing of RX packets with VLAN head
This is a refactoring commit that does not change behavior.
Signed-off-by: Dmitry Fleytman
---
hw/net/net_rx_pkt.c | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/hw/net/net_rx_pkt.c b/hw/net/net_rx_pkt.c
index c7ae33d..3899211 100644
--- a/hw/net/net_rx_pkt.c
+++ b
der) + sizeof(struct vlan_header)
bytes into it.
Devices affected by this problem: vmxnet3.
Reported-by: Peter Maydell
Signed-off-by: Dmitry Fleytman
---
hw/net/net_rx_pkt.c | 34 +-
1 file changed, 17 insertions(+), 17 deletions(-)
diff --git a/hw/net/net_rx_pkt.c b
In case of VLAN stripping ETH header is stored in a
separate chunk and length of IOV should take this into
account.
This patch fixes checksum validation for RX packets
with VLAN header.
Devices affected by this problem: e1000e and vmxnet3.
Signed-off-by: Dmitry Fleytman
---
hw/net
In case of VLAN stripping, ETH header put into a
separate buffer, therefore amont of data copied
from original IOV should be smaller.
Signed-off-by: Dmitry Fleytman
---
hw/net/net_rx_pkt.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/hw/net/net_rx_pkt.c b/hw/net
This series fix a few issues related
to processing of RX packets with VLAN headers.
See commit messages of specific patches
for information regarding affected devices.
Dmitry Fleytman (5):
eth: Extend vlan stripping functions
NetRxPkt: Fix memory corruption on VLAN header stripping
Make VLAN stripping functions return number of bytes
copied to given Ethernet header buffer.
This information should be used to re-compose
packet IOV after VLAN stripping.
Signed-off-by: Dmitry Fleytman
---
include/net/eth.h | 4 ++--
net/eth.c | 25 ++---
2 files
Reviewed-by: Dmitry Fleytman
This patch also fixes compilation with NET_RX_PKT_DEBUG defined.
> On 15 Feb 2017, at 10:31 AM, Fam Zheng wrote:
>
> This has been pointless since commit 605d52e62, which was a
> search-and-replace, overlooked the redundancy.
>
> Signed
Reviewed-by: Dmitry Fleytman
> On 10 Feb 2017, at 04:19 AM, Li Qiang wrote:
>
> This issue is like the issue in e1000 network card addressed in
> this commit:
> e1000: eliminate infinite loops on out-of-bounds transfer start.
>
> Signed-off-by: Li Qiang
> ---
>
> On 9 Feb 2017, at 04:11 AM, Li Qiang wrote:
>
> From: Li Qiang
>
> This issue is like the issue in e1000 network card addressed in
> this commit:
> e1000: eliminate infinite loops on out-of-bounds transfer start.
Hi,
Please see my comments below.
>
> Signed-off-by: Li Qiang
> ---
>
> C
> On 8 Feb 2017, at 12:17 PM, Li Qiang wrote:
>
> Hello Dmitry,
>
>
> 2017-02-08 18:01 GMT+08:00 Dmitry Fleytman :
>
>> On 8 Feb 2017, at 11:30 AM, Li Qiang wrote:
>>
>> Hello,
>>
>> 2017-02-08 16:38 GMT+08:00 Dmitry Fleytman :
>>
> On 8 Feb 2017, at 11:30 AM, Li Qiang wrote:
>
> Hello,
>
> 2017-02-08 16:38 GMT+08:00 Dmitry Fleytman <mailto:dmi...@daynix.com>>:
> Hello,
>
> Thanks for the patch!
>
> The problem of infinite loop indeed exists in e1000e, however it is different
Hello,
Thanks for the patch!
The problem of infinite loop indeed exists in e1000e, however it is different
from the one fixed in e1000.
Please find my comments inline.
~Dmitry
> On 7 Feb 2017, at 11:43 AM, Li Qiang wrote:
>
> From: Li Qiang
>
> This issue is the same as e1000 network card
Good catch!
Reviewed-by: Dmitry Fleytman
> On 26 Jan 2017, at 12:10 PM, Paolo Bonzini wrote:
>
> Because is_first is declared inside a loop, it is always true. The store
> is dead, and so is the "else" branch of "if (is_first)". is_last is
> oka
ned multibyte reads, LE/BE considerations,
> architecture-specific optimizations, etc. This patch still keeps
> things simple and readable.
Reviewed-by: Dmitry Fleytman
>
> Signed-off-by: Ladi Prosek
> ---
> net/checksum.c | 21 +
> 1 file changed, 13
> On 3 Jan 2017, at 21:40 PM, Dr. David Alan Gilbert
> wrote:
>
> * Dmitry Fleytman (dmi...@daynix.com <mailto:dmi...@daynix.com>) wrote:
>>>
>>> On 16 Dec 2016, at 14:19 PM, Dr. David Alan Gilbert
>>> wrote:
>>>
>>> * Dr.
Acked-by: Dmitry Fleytman
> On 21 Dec 2016, at 10:21 AM, Cao jin wrote:
>
> When user specify invalid value for property aer_log_max, device should
> fail to create, and report appropriate message.
>
> Signed-off-by: Cao jin
> Reviewed-by: Marcel Apfelbaum
> A
Acked-by: Dmitry Fleytman
> On 15 Dec 2016, at 22:05 PM, Dr. David Alan Gilbert (git)
> wrote:
>
> From: "Dr. David Alan Gilbert"
>
> The index's in the Vmxnet3Ring were migrated as 32bit ints
> yet are declared as size_t's. They appear to be deri
>> The problem is vmxnet3-ring size/cell_size/next are declared as size_t
>> but written as 32bit.
>
> Oops, I should have removed that warning in the commit message when
> I added the 1st patch in.
Acked-by: Dmitry Fleytman
>
>
Acked-by: Dmitry Fleytman
> On 3 Nov 2016, at 14:57 PM, Cao jin wrote:
>
> When user specify invalid property aer_log_max, device should fail to
> create, and report appropriate message.
>
> Signed-off-by: Cao jin
> ---
> v3 changelog:
> 1. get rid of PCIE_AER
t; 1) we avoid confusing double-negatives like
> "disable_vnet_hdr=false";
> 2) we avoid underscores in property names.
>
> This breaks command-line compatibility, but I am assuming the
> property is not being used in production by anybody.
Reviewed-by: Dmitry Fleytman
> On 26 Oct 2016, at 22:21 PM, Eduardo Habkost wrote:
>
> Follow the usual QOM property naming style, and make the property
> name clearer.
>
> This breaks command-line compatibility, but I am assuming the
> property is not being used in production by anybody.
I’m basically Ok with this patch
> On 26 Oct 2016, at 22:21 PM, Eduardo Habkost wrote:
>
> The user (or management software) is responsible for keeping the
> same configuration on both sides while migrating. Remove the
> configuration validation code at e1000e_post_load, and the
> unnecessary subsys_used/subsys_ven_used fields.
> On 27 Oct 2016, at 06:12 AM, Jason Wang wrote:
>
>
>
> On 2016年10月27日 03:21, Eduardo Habkost wrote:
>> Instead of hacking custom PropertyInfo structs, use the regular
>> DEFINE_PROP_ macros for the e1000e properties.
>>
>> This also fixes a bug in the disable_vnet_hdr property
>> definition
> On 19 Oct 2016, at 13:07 PM, Kevin Wolf wrote:
>
> Am 19.10.2016 um 09:57 hat Dmitry Fleytman geschrieben:
>>
>>On 19 Oct 2016, at 10:25 AM, Kevin Wolf wrote:
>>
>>Am 19.10.2016 um 08:48 hat Dmitry Fleytman geschrieben:
>>
>>
> On 19 Oct 2016, at 10:25 AM, Kevin Wolf wrote:
>
> Am 19.10.2016 um 08:48 hat Dmitry Fleytman geschrieben:
>>Another related thing that I noticed while debugging this and turning on
>>tracing is that the interrupt throttling timers kept firing even if
>>
Reviewed-by: Dmitry Fleytman
> On 19 Oct 2016, at 09:47 AM, Cao jin wrote:
>
> On migration target, msix_vector_use() will be called in vmxnet3_post_load()
> in second time, without a matching second call to msi_vector_unuse(),
> which results in vector reference leak.
> On 18 Oct 2016, at 17:27 PM, Kevin Wolf wrote:
>
> Am 18.10.2016 um 16:10 hat Dmitry Fleytman geschrieben:
>>On 17 Oct 2016, at 01:35 AM, Kevin Wolf wrote:
>>
>>The e1000e emulation zeroes out any used rx descriptor and then writes a
>>complet
Reviewed-by: Dmitry Fleytman
> On 19 Oct 2016, at 09:47 AM, Cao jin wrote:
>
> Internal flag msix_used is unnecessary, it has the same effect as
> msix_enabled().
>
> The corresponding msi flag is already dropped in commit 1070048e.
>
> CC: Dmitry Fleytman
> C
ptors unchanged even after the descriptor has been used.
Hi Kevin,
Reviewed-by: Dmitry Fleytman
Thanks for catching this!
~Dmitry
>
> Signed-off-by: Kevin Wolf
> ---
> hw/net/e1000e_core.c | 8
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a
> On 11 Oct 2016, at 6:35, Cao jin wrote:
>
>
>
>> On 10/06/2016 05:39 PM, Dmitry Fleytman wrote:
>>
>
>>
>> Hello,
>>
>> Yes, from what I see, this call is wrong and leads to
>> reference leaks on device unload at migration target
> On 6 Oct 2016, at 17:43, Dr. David Alan Gilbert wrote:
>
> * Dmitry Fleytman (dmi...@daynix.com) wrote:
>>
>>> On 30 Sep 2016, at 15:08 PM, Markus Armbruster wrote:
>>>
>>> Cao jin writes:
>>>
>>>>> On 0
> On 30 Sep 2016, at 15:08 PM, Markus Armbruster wrote:
>
> Cao jin writes:
>
>> On 09/29/2016 10:42 PM, Markus Armbruster wrote:
>>> Cao jin writes:
>>>
>>
static int vmxnet3_post_load(void *opaque, int version_id)
{
VMXNET3State *s = opaque;
-PCIDevice *d =
> On 22 Sep 2016, at 09:40 AM, Jason Wang wrote:
>
>
>
> On 2016年09月15日 14:14, Dmitry Fleytman wrote:
>> This patch fixes incorrect check for
>> interrypt type being used.
>>
>> PBSCLR register is valid for MSI-X only.
>>
>> See spec. 10.
Hi qemu-devel,
Ping.
~Dmitry
> On 15 Sep 2016, at 09:14 AM, Dmitry Fleytman wrote:
>
> This series contains a few RX logic fixes for e1000e.
>
> These fixes improve compliance to HW device specification
> and fix occasional interruptions of RX traffic with
> th
Interrupt mask for legacy OTHER causes should
not apply to MSI-X OTHER cause.
Signed-off-by: Dmitry Fleytman
---
hw/net/e1000e_core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/net/e1000e_core.c b/hw/net/e1000e_core.c
index c38ed10..e70e669 100644
--- a/hw/net
This patch fixes 2 issues:
1. Bits set in EIAC register should be cleared
from IMS when EIAM is not used.
2. Only bit that corresonds to the interrupt being
raised should be cleared.
See spec. 10.2.4.7 Interrupt Auto Clear
Signed-off-by: Dmitry Fleytman
---
hw/net/e1000e_core.c | 14
Do not raise ACK interrupts when
RFCTL.ACKDIS bit is set (see spec. 10.2.5.16).
Signed-off-by: Dmitry Fleytman
---
hw/net/e1000e_core.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/hw/net/e1000e_core.c b/hw/net/e1000e_core.c
index e70e669..f4ec79f 100644
--- a/hw/net
Before this patch first netdev queue only was flushed.
Signed-off-by: Dmitry Fleytman
---
hw/net/e1000e.c | 2 +-
hw/net/e1000e_core.c | 2 +-
hw/net/e1000e_core.h | 3 +++
3 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/hw/net/e1000e.c b/hw/net/e1000e.c
index bad43f4
This patch fixes incorrect check for
interrypt type being used.
PBSCLR register is valid for MSI-X only.
See spec. 10.2.3.13 MSI—X PBA Clear
Signed-off-by: Dmitry Fleytman
---
hw/net/e1000e_core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/net/e1000e_core.c b/hw
Signed-off-by: Dmitry Fleytman
---
hw/net/e1000e_core.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/hw/net/e1000e_core.c b/hw/net/e1000e_core.c
index ade4d25..a9603e0 100644
--- a/hw/net/e1000e_core.c
+++ b/hw/net/e1000e_core.c
@@ -1807,6 +1807,7 @@ e1000e_core_set_link_status
This series contains a few RX logic fixes for e1000e.
These fixes improve compliance to HW device specification
and fix occasional interruptions of RX traffic with
the latest linux driver.
Dmitry Fleytman (7):
e1000e: Flush all receive queues on receive enable
e1000e: Flush receive queue on
CTRL_EXT.EIAME bit controls clearing of IAM bits,
but current code clears IMS bits instead.
See spec. 10.2.2.5 Extended Device Control Register.
Signed-off-by: Dmitry Fleytman
---
hw/net/e1000e_core.c | 4 ++--
hw/net/trace-events | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff
> On 13 Sep 2016, at 15:53 PM, P J P wrote:
>
> Hello Dmitry,
>
> +-- On Tue, 13 Sep 2016, Dmitry Fleytman wrote --+
> | It should be in the same patch because otherwise a broken logic will be
> | introduced.
>
> Not even same patch-set?
Same patc
> On 13 Sep 2016, at 13:48 PM, P J P wrote:
>
> Hello Dmitry,
>
> +-- On Tue, 13 Sep 2016, Dmitry Fleytman wrote --+
> | > On 13 Sep 2016, at 10:00 AM, P J P wrote:
> | >
> | > +-- On Wed, 7 Sep 2016, P J P wrote --+
> | > | From: Prasad J Pandit
&
> On 13 Sep 2016, at 14:20 PM, P J P wrote:
>
> Hello Dmitry,
>
> +-- On Tue, 13 Sep 2016, Dmitry Fleytman wrote --+
> | Better put here:
> |
> | if(elemcnt++ >= PVSCSI_MAX_SG_ELEM) {
> | return;
> | }
> |
> | And ditch additional conditions from
Hello Prasad,
See my comments inline.
> On 13 Sep 2016, at 10:01 AM, P J P wrote:
>
> +-- On Tue, 6 Sep 2016, P J P wrote --+
> | From: Prasad J Pandit
> |
> | In PVSCSI paravirtual SCSI bus, pvscsi_convert_sglist can take a very
> | long time or go into an infinite loop due to two different
Hello Prasad,
Please see my questions inline.
> On 13 Sep 2016, at 10:00 AM, P J P wrote:
>
> +-- On Wed, 7 Sep 2016, P J P wrote --+
> | From: Prasad J Pandit
> |
> | Vmware Paravirtual SCSI emulator while processing IO requests
> | could run into an infinite loop if 'pvscsi_ring_pop_req_des
> On 9 Sep 2016, at 10:15 AM, Gonglei (Arei) wrote:
>
> Who can pick up this patch? Dmitry or Jason? Thanks!
Jason, would you please?
>
>
> Regards,
> -Gonglei
>
>
>> -----Original Message-
>> From: Dmitry Fleytman [mailto:dmi...@daynix.com]
>
Reviewed-by: Dmitry Fleytman
> On 1 Sep 2016, at 14:00 PM, P J P wrote:
>
> From: Prasad J Pandit
>
> Vmware Paravirtual SCSI emulation uses command descriptors to
> process SCSI commands. These descriptors come with their ring
> buffers. A guest could set the page cou
> On 1 Sep 2016, at 07:50 AM, P J P wrote:
>
> Hello Dmitry,
>
> +-- On Wed, 31 Aug 2016, Dmitry Fleytman wrote --+
> | > -if ((ri->reqRingNumPages > PVSCSI_SETUP_RINGS_MAX_NUM_PAGES)
> | > -|| (ri->cmpRingNumPages > PVSCSI_SETUP_RINGS_MAX
> On 31 Aug 2016, at 09:49 AM, P J P wrote:
>
> From: Prasad J Pandit
>
> Vmware Paravirtual SCSI emulation uses command descriptors to
> process SCSI commands. These descriptors come with their ring
> buffers. A guest could set the page count for these rings to
> an arbitrary value, leading t
1 - 100 of 421 matches
Mail list logo
