On 12/22/2015 5:32 AM, Thomas Monjalon wrote:
> 2015-12-21 17:20, Wiles, Keith:
>> On 12/21/15, 9:21 AM, "Xie, Huawei" wrote:
>>> On 12/19/2015 3:27 AM, Wiles, Keith wrote:
On 12/18/15, 11:32 AM, "dev on behalf of Stephen Hemminger" >>> at dpdk.org on behalf of stephen at networkplumber.org>
On Mon, Dec 21, 2015 at 03:32:53PM +, Xie, Huawei wrote:
> > +
> > + /*
> > +* mmap from 0 to workaround a hugepage mmap bug: mmap will be
> > +* failed when offset is not page size aligned.
> > +*/
> s /will be failed/will fail/
> mmap will fail when offset is not zero.
> Also
On Mon, Dec 21, 2015 at 03:06:43PM +, Xie, Huawei wrote:
> On 12/17/2015 11:11 AM, Yuanhan Liu wrote:
> > Introduce vhost_log_write() helper function to log the dirty pages we
> > touched. Page size is harded code to 4096 (VHOST_LOG_PAGE), and each
> > log is presented by 1 bit.
> >
> >
> -Original Message-
> From: Yuanhan Liu [mailto:yuanhan.liu at linux.intel.com]
> Sent: Tuesday, December 22, 2015 10:26 AM
> To: Xie, Huawei
> Cc: dev at dpdk.org; Michael S. Tsirkin; Victor Kaplansky; Iremonger,
> Bernard; Pavel Fedin; Peter Xu
> Subject: Re: [PATCH v2 1/6] vhost:
On 12/22/2015 10:40 AM, Yuanhan Liu wrote:
> On Mon, Dec 21, 2015 at 03:06:43PM +, Xie, Huawei wrote:
>> On 12/17/2015 11:11 AM, Yuanhan Liu wrote:
>>> Introduce vhost_log_write() helper function to log the dirty pages we
>>> touched. Page size is harded code to 4096 (VHOST_LOG_PAGE), and each
On 12/22/2015 2:04 AM, Martinx - ? wrote:
> On 10 December 2015 at 02:45, Xie, Huawei wrote:
>> On 12/10/2015 6:49 AM, Martinx - ? wrote:
>>> On 9 December 2015 at 18:05, Thomas Monjalon
>>> wrote:
2015-12-09 15:54, Martinx - ?:
> Sorry to insist on this subject but, the
On Tue, Dec 22, 2015 at 02:45:52AM +, Xie, Huawei wrote:
> >>> +static inline void __attribute__((always_inline))
> >>> +vhost_log_write(struct virtio_net *dev, uint64_t addr, uint64_t len)
> >>> +{
> >>> + uint64_t page;
> >>> +
> >> Before we log, we need memory barrier to make sure updates
On 12/18/2015 3:03 PM, Yuanhan Liu wrote:
> While we use a single linked list to maintain all devices, we could
> use a static array to achieve the same goal, just like what we did
> to maintain the eth devices with rte_eth_devices array. This could
> simplifies the code a bit.
>
> Signed-off-by:
On Fri, Dec 18, 2015 at 10:01:25AM -0800, Rich Lane wrote:
> I'm using the vhost callbacks and struct virtio_net with the vhost PMD in a
> few
> ways:
Rich, thanks for the info!
>
> 1. new_device/destroy_device: Link state change (will be covered by the link
> status interrupt).
> 2.
Hi,
I got a question related to how virtio pmd driver work without
UIO layer.
I see that in virtio PMD driver, DPDK will first try to init the
device using UIO interfaces. If it fails, it will try to init by
manipulating IO ports directly (see virtio_resource_init()).
For the ioport case, is it
On Mon, Dec 21, 2015 at 11:10:10AM +0900, Tetsuya Mukawa wrote:
> nes: 168
>
> On 2015/12/19 3:01, Rich Lane wrote:
> > I'm using the vhost callbacks and struct virtio_net with the vhost PMD in a
> > few ways:
> >
> > 1. new_device/destroy_device: Link state change (will be covered by the
> >
On Thu, Dec 17, 2015 at 11:11:57AM +0800, Yuanhan Liu wrote:
> +static inline void __attribute__((always_inline))
> +vhost_log_write(struct virtio_net *dev, uint64_t addr, uint64_t len)
> +{
> + uint64_t page;
> +
> + if (likely(((dev->features & (1ULL << VHOST_F_LOG_ALL)) == 0) ||
> +
On Mon, Dec 21, 2015 at 08:47:28PM -0800, Rich Lane wrote:
> On Mon, Dec 21, 2015 at 7:41 PM, Yuanhan Liu
> wrote:
>
> On Fri, Dec 18, 2015 at 10:01:25AM -0800, Rich Lane wrote:
> > I'm using the vhost callbacks and struct virtio_net with the vhost PMD
> in
> a few
> > ways:
>
On Tue, Dec 22, 2015 at 01:11:02PM +0800, Peter Xu wrote:
> On Thu, Dec 17, 2015 at 11:11:57AM +0800, Yuanhan Liu wrote:
> > +static inline void __attribute__((always_inline))
> > +vhost_log_write(struct virtio_net *dev, uint64_t addr, uint64_t len)
> > +{
> > + uint64_t page;
> > +
> > + if
On 12/18/2015 3:03 PM, Yuanhan Liu wrote:
> While we use a single linked list to maintain all devices, we could
> use a static array to achieve the same goal, just like what we did
> to maintain the eth devices with rte_eth_devices array. This could
> simplifies the code a bit.
>
> Signed-off-by:
On Fri, Dec 18, 2015 at 06:16:36PM +0530, Santosh Shukla wrote:
> On Fri, Dec 18, 2015 at 4:54 AM, Stephen Hemminger
> wrote:
> > On Thu, 17 Dec 2015 17:32:38 +0530
> > Santosh Shukla wrote:
> >
> >> On Mon, Dec 14, 2015 at 6:30 PM, Santosh Shukla
> >> wrote:
> >> > virtio_recv_pkts_vec and
On 12/18/2015 3:03 PM, Yuanhan Liu wrote:
> We could first check if we need realloc vq or not, if so,
> reallocate it. We then do similar to vhost dev realloc.
>
> This could get rid of the tons of repeated "if (realloc_dev)"
> and "if (realloc_vq)" statements, therefore, makes code
> a bit more
On 12/21/2015 6:20 PM, Shaopeng He wrote:
> rx_descriptor_done is used by interrupt mode example application
> (l3fwd-power) to check rxd DD bit to decide the RX trend,
> then l3fwd-power will adjust the cpu frequency according to
> the result.
>
> Signed-off-by: Shaopeng He
> Acked-by: Jing Chen
On Tue, Dec 22, 2015 at 06:46:32AM +, Xie, Huawei wrote:
> On 12/18/2015 3:03 PM, Yuanhan Liu wrote:
> > We could first check if we need realloc vq or not, if so,
> > reallocate it. We then do similar to vhost dev realloc.
> >
> > This could get rid of the tons of repeated "if (realloc_dev)"
>
On 12/21/2015 6:20 PM, Shaopeng He wrote:
> Previous dev_stop function stops the rx/tx queues. This patch adds logic
> to disable rx queue interrupt, clean the datapath event and queue/vec map.
>
> Signed-off-by: Shaopeng He
> Acked-by: Jing Chen
> ---
> drivers/net/fm10k/fm10k_ethdev.c | 22
On Thu, Dec 17, 2015 at 11:11:58AM +0800, Yuanhan Liu wrote:
> +static inline void __attribute__((always_inline))
> +vhost_log_used_vring(struct virtio_net *dev, struct vhost_virtqueue *vq,
> + uint64_t offset, uint64_t len)
> +{
One thing optional: I feel it a little bit
On 12/18/2015 3:03 PM, Yuanhan Liu wrote:
> vq is allocated on pairs, hence we should do pair reallocation
> at numa_realloc() as well, otherwise an error like following
> occurs while do numa reallocation:
>
> VHOST_CONFIG: reallocate vq from 0 to 1 node
> PANIC in rte_free():
> Fatal
On Tue, Dec 22, 2015 at 11:50:41AM +0800, Peter Xu wrote:
> Hi,
>
> I got a question related to how virtio pmd driver work without
> UIO layer.
>
> I see that in virtio PMD driver, DPDK will first try to init the
> device using UIO interfaces. If it fails, it will try to init by
> manipulating
On 10/26/2015 11:48 AM, He, Shaopeng wrote:
> Signed-off-by: Shaopeng He
> ---
> doc/guides/rel_notes/release_2_2.rst | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/doc/guides/rel_notes/release_2_2.rst
> b/doc/guides/rel_notes/release_2_2.rst
> index 73dba47..44b3aea 100644
> ---
On 12/22/2015 11:03 AM, Yuanhan Liu wrote:
> On Tue, Dec 22, 2015 at 02:45:52AM +, Xie, Huawei wrote:
> +static inline void __attribute__((always_inline))
> +vhost_log_write(struct virtio_net *dev, uint64_t addr, uint64_t len)
> +{
> + uint64_t page;
> +
Before we log,
On 12/22/2015 2:56 PM, Peter Xu wrote:
> On Thu, Dec 17, 2015 at 11:11:58AM +0800, Yuanhan Liu wrote:
>> +static inline void __attribute__((always_inline))
>> +vhost_log_used_vring(struct virtio_net *dev, struct vhost_virtqueue *vq,
>> + uint64_t offset, uint64_t len)
>> +{
[...]
>
On Tue, Dec 22, 2015 at 02:55:52PM +0800, Peter Xu wrote:
> On Thu, Dec 17, 2015 at 11:11:58AM +0800, Yuanhan Liu wrote:
> > +static inline void __attribute__((always_inline))
> > +vhost_log_used_vring(struct virtio_net *dev, struct vhost_virtqueue *vq,
> > +uint64_t offset,
While we use a single linked list to maintain all devices, we could
use a static array to achieve the same goal, just like what we did
to maintain the eth devices with rte_eth_devices array. This could
simplifies the code a bit.
Signed-off-by: Yuanhan Liu
Acked-by: Huawei Xie
---
Note that
We could first check if we need realloc vq or not, if so,
reallocate it. We then do similar to vhost dev realloc.
This could get rid of the tons of repeated "if (realloc_dev)"
and "if (realloc_vq)" statements, therefore, makes code
a bit more readable.
Signed-off-by: Yuanhan Liu
---
v2: fix
vq is allocated on pairs, hence we should do pair reallocation
at numa_realloc() as well, otherwise an error like following
occurs while do numa reallocation:
VHOST_CONFIG: reallocate vq from 0 to 1 node
PANIC in rte_free():
Fatal error: Invalid memory
The reason we don't catch it is
On 12/21/2015 6:20 PM, Shaopeng He wrote:
> In interrupt mode, each rx queue can have one interrupt to notify the up
> layer application when packets are available in that queue. Some queues
> also can share one interrupt.
> Currently, fm10k needs one separate interrupt for mailbox. So, only those
On Tue, Dec 22, 2015 at 07:07:25AM +, Xie, Huawei wrote:
> On 12/22/2015 2:56 PM, Peter Xu wrote:
> > Got a question here:
> >
> > I see that we are logging down changes when we are marking
> > used_vring. Do we need to log down buffer copy in rte_memcpy() too?
> > I am not sure whether I
On Tue, Dec 22, 2015 at 03:13:49PM +0800, Yuanhan Liu wrote:
> On Tue, Dec 22, 2015 at 02:55:52PM +0800, Peter Xu wrote:
> > On Thu, Dec 17, 2015 at 11:11:58AM +0800, Yuanhan Liu wrote:
> > > +static inline void __attribute__((always_inline))
> > > +vhost_log_used_vring(struct virtio_net *dev,
On Thu, Dec 17, 2015 at 11:12:00AM +0800, Yuanhan Liu wrote:
> It's actually a feature already enabled in Linux kernel. What we need to
> do is simply to claim that we support such feature, and nothing else.
>
> With that, the guest will send GARP messages after live migration.
>
>
On Tue, Dec 22, 2015 at 04:11:08PM +0800, Peter Xu wrote:
> On Thu, Dec 17, 2015 at 11:12:00AM +0800, Yuanhan Liu wrote:
> > It's actually a feature already enabled in Linux kernel. What we need to
> > do is simply to claim that we support such feature, and nothing else.
> >
> > With that, the
On Tue, Dec 22, 2015 at 03:00:29PM +0800, Yuanhan Liu wrote:
> On Tue, Dec 22, 2015 at 11:50:41AM +0800, Peter Xu wrote:
> > Hi,
> >
> > I got a question related to how virtio pmd driver work without
> > UIO layer.
> >
> > I see that in virtio PMD driver, DPDK will first try to init the
> >
Hello!
> > diff --git a/lib/librte_vhost/virtio-net.c b/lib/librte_vhost/virtio-net.c
> > index 03044f6..0ba5045 100644
> > --- a/lib/librte_vhost/virtio-net.c
> > +++ b/lib/librte_vhost/virtio-net.c
> > @@ -74,6 +74,7 @@ static struct virtio_net_config_ll *ll_root;
> > #define
On Tue, Dec 22, 2015 at 04:23:38PM +0800, Peter Xu wrote:
> On Tue, Dec 22, 2015 at 03:00:29PM +0800, Yuanhan Liu wrote:
> > On Tue, Dec 22, 2015 at 11:50:41AM +0800, Peter Xu wrote:
> > > Hi,
> > >
> > > I got a question related to how virtio pmd driver work without
> > > UIO layer.
> > >
> > >
As demonstrated by the following code, CRC32c computation is not valid
when buffer length is not a multiple of 4 bytes:
(Output obtained by code below)
CRC of 1 NULL bytes expected: 0x527d5351
soft: 527d5351
rte accelerated: 48674bc7
rte soft: 48674bc7
CRC of 2 NULL bytes expected:
On Mon, Dec 21, 2015 at 9:47 PM, Yuanhan Liu
wrote:
> On Mon, Dec 21, 2015 at 08:47:28PM -0800, Rich Lane wrote:
> > The queue state change callback is the one new API that needs to be
> > added because
> > normal NICs don't have this behavior.
>
> Again I'd ask, will vring_state_changed() be
On Tue, Dec 22, 2015 at 04:32:46PM +0800, Yuanhan Liu wrote:
> Actually, you are right. I mentioned in the last email that this is
> for configuration part. To answer your question in this email, you
> will not be able to go that further (say initiating virtio pmd) if
> you don't unbind the origin
On 12/22/2015 5:57 PM, Peter Xu wrote:
> On Tue, Dec 22, 2015 at 04:32:46PM +0800, Yuanhan Liu wrote:
>> Actually, you are right. I mentioned in the last email that this is
>> for configuration part. To answer your question in this email, you
>> will not be able to go that further (say initiating
On 12/22/2015 6:48 PM, Xie, Huawei wrote:
> On 12/22/2015 5:57 PM, Peter Xu wrote:
>> On Tue, Dec 22, 2015 at 04:32:46PM +0800, Yuanhan Liu wrote:
>>> Actually, you are right. I mentioned in the last email that this is
>>> for configuration part. To answer your question in this email, you
>>> will
v2:
* Removed the bug fix unrelated code change to make this patch a pure
bug fix patch.
* Fixed the "PATCH 1/2" mistake in the patch title, rewrote the subject.
v1:
* Initial version for tx_vlan set command support bug fix.
Wang Xiao W (1):
testpmd: fix a bug in tx_vlan set command support
Now in cmd_tx_vlan_set_parsed function, we check the vlan_offload
capability first, if it's an invalid port_id we'll get a strange
prompt saying "Error, as QinQ has been enabled.". We should always
make sure that we get a valid port_id first before we check other
information. It's the same problem
On Tue, Dec 22, 2015 at 10:47:21AM +, Xie, Huawei wrote:
> On 12/22/2015 5:57 PM, Peter Xu wrote:
> > On Tue, Dec 22, 2015 at 04:32:46PM +0800, Yuanhan Liu wrote:
> >> Actually, you are right. I mentioned in the last email that this is
> >> for configuration part. To answer your question in
Hi,
just tried to forward a lot of tiny packets with testpmd (dpdk-2.2.0)
running on Broadcom Corporation NetXtreme II BCM57810S 10 Gigabit Ethernet
(rev 10) adapter. I see forwarding performance of 2.6Mpps instead of
expected 14.8Mpps. What should be done to achieve better results?
Thank you,
> -Original Message-
> From: Qiu, Michael
> Sent: Tuesday, December 22, 2015 2:51 PM
> To: He, Shaopeng; dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v4 1/6] fm10k: implement
> rx_descriptor_done function
>
> On 12/21/2015 6:20 PM, Shaopeng He wrote:
> > rx_descriptor_done is used by
> -Original Message-
> From: Qiu, Michael
> Sent: Tuesday, December 22, 2015 2:55 PM
> To: He, Shaopeng; dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v4 3/6] fm10k: remove rx queue interrupts
> when dev stops
>
> On 12/21/2015 6:20 PM, Shaopeng He wrote:
> > Previous dev_stop function
> -Original Message-
> From: Qiu, Michael
> Sent: Tuesday, December 22, 2015 3:00 PM
> To: He, Shaopeng; dev at dpdk.org
> Cc: Chen, Jing D
> Subject: Re: [PATCH v2 7/7] doc: release note update for fm10k intr mode
>
> On 10/26/2015 11:48 AM, He, Shaopeng wrote:
> > Signed-off-by:
Depending on non-doc targets being built before and the setting of DESTDIR
the examples dir could in some cases not end up in the right target.
Reason is just a typo variable reference in the copy target.
Signed-off-by: Christian Ehrhardt
---
[diffstat]
rte.sdkinstall.mk |2 +-
1 file
On 12/22/2015 7:39 PM, Peter Xu wrote:
> On Tue, Dec 22, 2015 at 10:47:21AM +, Xie, Huawei wrote:
>> On 12/22/2015 5:57 PM, Peter Xu wrote:
>>> On Tue, Dec 22, 2015 at 04:32:46PM +0800, Yuanhan Liu wrote:
Actually, you are right. I mentioned in the last email that this is
for
On 12/22/2015 7:39 PM, Peter Xu wrote:
> On Tue, Dec 22, 2015 at 10:47:21AM +, Xie, Huawei wrote:
>> On 12/22/2015 5:57 PM, Peter Xu wrote:
>>> On Tue, Dec 22, 2015 at 04:32:46PM +0800, Yuanhan Liu wrote:
Actually, you are right. I mentioned in the last email that this is
for
There is really no way to safely give a user full access to a DMA
capable device without an IOMMU to protect the host system. There is
also no way to provide DMA translation, for use cases such as device
assignment to virtual machines. However, there are still those users
that want userspace
Hey guys,
I'm using dpdk 2.1.0, and get SIGILL|SIGSEGV when calling
ixgbevf_vlan_filter_set from secondary process.
Is it possible to do that in the first place?
Thanks,
Shaham
There is really no way to safely give a user full access to a DMA
capable device without an IOMMU to protect the host system. There is
also no way to provide DMA translation, for use cases such as device
assignment to virtual machines. However, there are still those users
that want userspace
1. Implement rte_vect.h and enable LPM lookup;
2. It also has some code cleanup and fixes for potential crash when quiting pmd
under high traffic;
3. Changes have dependency on serie (http://dpdk.org/dev/patchwork/patch/9571)
which is pending in the patchwork queue.
Liming Sun (2):
rte_vect.h was missing earlier thus LPM was disabled and l3fwd is
not able to compile. This commit implements the vector api and
enable LPM in the tilegx configuration by default. It also includes
a minor optimization to use __insn_fetchadd4() instead of
rte_atomic32_xxx() in
1. Fixed the compiling issue of the ethtool example on tilegx
platform.
2. Fixed the hung/crash issue when quitting testpmd under high
traffic rate. The buffer error bit needs to be checked before
processing the idesc and releasing the buffer. Code logic is
also simplified.
On 12/3/15, 5:05 PM, "Stephen Hemminger" wrote:
>From: Stephen Hemminger
>
>The vmxnet3 interface specification supports having multiple
>receive rings. The first ring has buffers of BTYPE_HEAD which
>are used for the start of the packet, the second ring has buffers
>of type BTYPE_BODY which
On Mon, 2015-12-21 at 12:22 -0700, Alex Williamson wrote:
> On Mon, 2015-12-21 at 11:46 +, Yigit, Ferruh wrote:
> > On Fri, Dec 18, 2015 at 02:50:17PM -0700, Alex Williamson wrote:
> > > On Fri, 2015-12-18 at 07:38 -0700, Alex Williamson wrote:
> > > > On Fri, 2015-12-18 at 10:43 +, Yigit,
I checked with our engineers internally and it turns out that the changes to
lower the max ring2 size never got checked-in in 6.0. This should explain why
it works on ESXi6.0. And for 5.5, the plan is to revert the changes in 55p08
patch to bring it back to 4096.
On 12/13/15, 11:06 PM,
On 7/23/15, 5:53 PM, "dev on behalf of Marco Lee" wrote:
>The RX of VMXNET3 PMD will have deadlock when a lot of traffic coming in.
>The root cause is due to mbuf allocation fail in vmxnet3_post_rx_bufs()
>and there is no error handling when it is called from vmxnet3_recv_pkts().
>The RXD will
63 matches
Mail list logo
