> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Wednesday, December 10, 2014 11:26 PM
> To: Zhang, Helin
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v4] mbuf: fix of enabling all newly added RX
> error flags
>
> 2014-12-10 13:50,
This patch set is based on latest vhost.
It fixes vhost-user memory map/unmap alignment issue.
It uses VHOST_USER_GET_VRING_BASE as the message for vhost device stop
in vhost-user.
It uses VHOST_SET_VRING_KICK as the message that tells us vhost device
is ready in
And add xen_dom0/compat.h for be compatible with older kernel.
Signed-off-by: Jincheng Miao
---
lib/librte_eal/linuxapp/xen_dom0/compat.h | 16
lib/librte_eal/linuxapp/xen_dom0/dom0_mm_misc.c | 3 ++-
2 files changed, 18 insertions(+), 1 deletion(-)
create mode 100644
Search the right segment to increase its data length, rather than
wrongly early return and exit the tx function, which leads to drop all jumbo
frame packets
when vm2vm is in hard forward mode.
Signed-off-by: Changchun Ouyang
---
examples/vhost/main.c | 25 -
1 file
When vfio module is not loaded when kernel support vfio feature,
the routine still try to open the container to get file
description.
This action is not safe, and of cause got error messages:
EAL: Detected 40 lcore(s)
EAL: unsupported IOMMU type!
EAL: VFIO support could not be initialized
EAL:
na ez :)
On Wed, 10 Dec 2014, Bruce Richardson wrote:
> On Wed, Dec 10, 2014 at 09:29:26AM -0500, Neil Horman wrote:
>> On Wed, Dec 10, 2014 at 10:32:25AM +, Bruce Richardson wrote:
>>> On Tue, Dec 09, 2014 at 02:10:32PM -0800, Stephen Hemminger wrote:
On Tue, 9 Dec 2014 11:45:07 -0800
Hi Neil,
Moving __rte_cache_aligned right after struct keyword will help. On the
other hand, enforcing this rule for existing (100+) and future definitions
will be difficult. It?s clearer and a good practice to include header file
explicitly.
Thanks,
Jia
On 12/9/14, 7:22 AM, "Neil Horman"
And add xen_dom0/compat.h for be compatible with older kernel.
Signed-off-by: Jincheng Miao
---
lib/librte_eal/linuxapp/xen_dom0/compat.h | 16
lib/librte_eal/linuxapp/xen_dom0/dom0_mm_misc.c | 2 +-
2 files changed, 17 insertions(+), 1 deletion(-)
create mode 100644
Add kni/compat.h for be compatible with older kernel.
Signed-off-by: Jincheng Miao
---
lib/librte_eal/linuxapp/kni/compat.h| 16
lib/librte_eal/linuxapp/kni/kni_vhost.c | 3 ++-
2 files changed, 18 insertions(+), 1 deletion(-)
create mode 100644
Function pci_num_vf() is introduced from upstream linux-2.6.34. So
this patch make compatible with longterm kernel linux-2.6.32.63.
For RHEL, function pci_num_vf() begins from RHEL5 update9. And
it is stub-defined when CONFIG_PCI_IOV is not enabled.
So dropped the CONFIG_PCI_IOV checking of
The related kernel function is:
- pci_num_vf, it is introduced from upstream linux-2.6.34. For RHEL-based
kernel, it is defined from RHEL5.9.
- kstrtoul, this function is united kernel API to replace strict_strtoul in
the furture. It is introduced from linux-2.6.39. For RHEL6, it is defined
from
lib/librte_eal/linuxapp/eal/eal_memory.c:324:4: error: comparison
is always false due to limited range of data type [-Werror=type-limits]
|| (hugepage_sz == RTE_PGSIZE_16G)) {
^
cc1: all warnings being treated as errors
This was introuduced by commit b77b5639:
mem: add huge page
> From: Jay Rolette [mailto:rolette at infiniteio.com]
> Sent: Wednesday, December 10, 2014 5:58 PM
> To: Ananyev, Konstantin
> Cc: Qiu, Michael; Richardson, Bruce; dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] Avoid possible memory cpoy when sort hugepages
>
>
> On Wed, Dec 10, 2014 at
2014-12-10 13:50, Zhang, Helin:
>
> > -Original Message-
> > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > Sent: Wednesday, December 10, 2014 5:35 PM
> > To: Zhang, Helin
> > Cc: dev at dpdk.org
> > Subject: Re: [dpdk-dev] [PATCH v4] mbuf: fix of enabling all newly added
On Thu, 11 Dec 2014 00:23:24 +0100
Thomas Monjalon wrote:
> 2014-12-09 08:37, Stephen Hemminger:
> > Add missing setup for X540 MAC type when setting up VF.
> > Additional check exists in Linux driver but not in DPDK.
> >
> > Signed-off-yb: Bill Hong
> > Signed-off-by: Stephen Hemminger
> >
2014-12-09 10:11, Pablo de Lara:
> Since commit b91c67e5a693211862aa7dc3b78630b4e856c2af,
> maximum number of cores is 128, which has increase
> the total memory necessary for a rte_mempool structure,
> as the per-lcore local cache has been doubled in size.
> Therefore, eal_flags unit test was
On Wed, 10 Dec 2014 11:16:46 -0500
Neil Horman wrote:
> This really seems like a false savings to me. If an application intends to
> use
> multiple processes (which by all rights it seems like the use case that the
> dpdk
> is mostly designed for) then you need locking one way or another, and
On Wed, Dec 10, 2014 at 12:39 PM, Ananyev, Konstantin <
konstantin.ananyev at intel.com> wrote:
> > I just got through replacing that entire function in my repo with a call
> to qsort() from the standard library last night myself. Faster
> > (although probably not material to most deployments)
Hi Bruce,
>>>The standard sample applications with DPDK use a simple buffering
scheme, where
>>>we buffer the packets until a full burst of 32 are ready for sending.
Once we
>>>have a full burst of packets - or a timeout occurs - we then send that
burst
>>>of packets using tx_burst function. Would
Subject: [PATCH] When "make clean" is executed following error msg is
thrown == Clean lib/librte_eal/linuxapp/kni tr: missing operand after '.-'
Two strings must be given when translating.
due to missing "comma" definition in Makefile.
Signed-off-by: Ravi Kerur
---
When 'make clean' is executed, following message is thrown on console due
to missing 'comma' definition.
tr: missing operand after '.-'
Two strings must be given when translating.
Ravi Kerur (1):
When "make clean" is executed following msg is thrown
==Clean lib/librte_eal/linuxapp/kni tr:
2014-12-05 13:46, Bruce Richardson:
> When using multiple processes, the TX function used in all processes
> should be the same, otherwise the secondary processes cannot transmit
> more than tx-ring-size - 1 packets.
> To achieve this, we extract out the code to select the ixgbe TX function
> to
On Wed, Dec 10, 2014 at 09:47:45AM -0500, Neil Horman wrote:
> On Wed, Dec 10, 2014 at 08:18:36AM +, Wodkowski, PawelX wrote:
> > > Though, that said, doesn't it seem to anyone else like serialization of
> > > enqueue
> > > to a port should be the responsibility of the library, not the
> > >
2014-12-09 13:05, r k:
> Subject: [PATCH] Fix power_of_2 macro. Avoid branching when
> calculating RTE_MIN and RTE_MAX.
Please could you add more explanations about the problem you are solving?
You should also add a Signed-off like explained in this page:
http://dpdk.org/dev#send
Thanks
On Wed, Dec 10, 2014 at 01:09:32PM +0100, Sachin Sharma wrote:
> Hi Bruce,
>
> >>>I'm not entirely clear on what you mean by filling one queue and
> emptying another. Is this just a form of buffering you are trying to
> implement?
>
> Yes, you are right! I am implementing a buffering mechanism
Hi Bruce,
>>>I'm not entirely clear on what you mean by filling one queue and
emptying another. Is this just a form of buffering you are trying to
implement?
Yes, you are right! I am implementing a buffering mechanism in which a node
will have three queues and it fills one queue with packets
Thanks for your reply.
What do you mean "device assignment" ? You mean pass-through?
> Yes, the same thing. Pass through of PCI device to VM via qemu
(pci-assign, vfio-pci, ivshmem etc)
As I knows, the answer is *NO*, because the iommu is not support in guest.
> NO was essentially what I was
Hi Bruce,
In my use case, I want to have three NIC TX queues per port, and want to
fill one NIC TX queue and want to empty the other queue. Is it possible
this through tx_burst or do I need to implement these queues in
applications as you suggested before. However, in this case, I would have
then
Thanks for your reply. Sorry, that was a typo, It was meant to be this
only.
00:04.0 Ethernet controller: Intel Corporation 82599 Ethernet Controller
Virtual Function (rev 01)
On Wed, Dec 10, 2014 at 7:45 AM, Kiran KN wrote:
> Looks like virtual functions are not created.
> For 82599, use
Dear all,
In my algorithm, I am interested to perform two activities - (1)
transmitting packets to a tx_queue and (2) transmitting packets from
tx_queue to a wire - separately. I have gone through the code by putting
logs in the dpdk code and found that there is a function rte_eth_tx_burst
On Wed, Dec 10, 2014 at 5:08 AM, Ananyev, Konstantin <
konstantin.ananyev at intel.com> wrote:
> I wonder why we do need to write our own bubble sort procedure?
> Why we can't use standard qsort() here?
>
Sadly, even bubble sort would be better than the selection sort being used
here. It's
On Wed, Dec 10, 2014 at 12:31:27PM +0100, Sachin Sharma wrote:
> Hi Bruce,
>
> In my use case, I want to have three NIC TX queues per port, and want to
> fill one NIC TX queue and want to empty the other queue. Is it possible
> this through tx_burst or do I need to implement these queues in
>
>From upstream kernel commit 3db2e9cd, strict_strto* serial functions
are removed. So that we should directly used kstrtoul instead.
And add kni/compat.h for be compatible with older kernel.
Signed-off-by: Jincheng Miao
---
lib/librte_eal/linuxapp/kni/compat.h| 16
>From upstream kernel commit 3db2e9cd, strict_strto* serial functions
are removed. So that we should directly used kstrtoul instead.
Kstrtoul exists from RHEL6.4, so for compatible with old kernel and RHEL,
add some logic to igb_uio/compat.h, same as what we do for pci_num_vf().
Signed-off-by:
On 12/10/2014 5:21 PM, Burakov, Anatoly wrote:
>> When vfio module is not loaded when kernel support vfio feature, the
>> routine still try to open the container to get file description.
>>
>> This action is not safe, and of cause got error messages:
>>
>> EAL: Detected 40 lcore(s)
>> EAL:
On Wed, Dec 10, 2014 at 12:03:41PM +0100, Sachin Sharma wrote:
> Dear all,
>
> In my algorithm, I am interested to perform two activities - (1)
> transmitting packets to a tx_queue and (2) transmitting packets from
> tx_queue to a wire - separately. I have gone through the code by putting
>
On Wed, Dec 10, 2014 at 02:54:56PM +, Bruce Richardson wrote:
> On Wed, Dec 10, 2014 at 09:47:45AM -0500, Neil Horman wrote:
> > On Wed, Dec 10, 2014 at 08:18:36AM +, Wodkowski, PawelX wrote:
> > > > Though, that said, doesn't it seem to anyone else like serialization of
> > > > enqueue
>
There is a bug in querying reta, of storing the data to the correct entry.
Code changes is straightforward for this bug.
Signed-off-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_ethdev.c| 2 +-
lib/librte_pmd_i40e/i40e_ethdev_vf.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
v2
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Qiu, Michael
> Sent: Wednesday, December 10, 2014 10:59 AM
> To: Richardson, Bruce
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] Avoid possible memory cpoy when sort hugepages
>
> On 12/10/2014 6:41
On 12/10/2014 6:41 PM, Richardson, Bruce wrote:
> On Wed, Dec 10, 2014 at 10:25:41AM +0800, Michael Qiu wrote:
>> When the first address is the compared address in the loop,
>> it will also do memory copy, which is meaningless,
>> worse more, when hugepg_tbl is mostly in order. This should
>> be a
Here is my patch for it, and it also resolves issue of pci_num_vf()
definition.
And I will send V3 for a while.
On 12/10/2014 10:38 AM, Helin Zhang wrote:
> strict_strtoul() was just a redefinition of kstrtoul() for a long
> time. From kernel version of 3.18, strict_strtoul() will not be
>
On Wed, Dec 10, 2014 at 10:25:41AM +0800, Michael Qiu wrote:
> When the first address is the compared address in the loop,
> it will also do memory copy, which is meaningless,
> worse more, when hugepg_tbl is mostly in order. This should
> be a big deal in large hugepage memory systerm(like
strict_strtoul() was just a redefinition of kstrtoul() for a long
time. From kernel version of 3.18, strict_strtoul() will not be
defined at all. A compile time kernel version check is needed to
decide which function or macro can be used for a specific version
of kernel.
Signed-off-by: Helin
2014-12-10 16:55, Helin Zhang:
> Before redefining mbuf structure, there was lack of free bits in 'ol_flags'
> (32 bits in total) for new RX or TX flags. So it tried to reuse existant
> bits as most as possible, or even assigning 0 to some of bit flags. After
> new mbuf structure defined, there
> -Original Message-
> From: Doherty, Declan
> Sent: Monday, December 8, 2014 11:22 AM
> To: dev at dpdk.org
> Cc: Iremonger, Bernard; Doherty, Declan
> Subject: [PATCH] doc: add license header to link bonding diagrams
>
>
> Signed-off-by: Declan Doherty
Acked-by: Bernard Iremonger
When the first address is the compared address in the loop,
it will also do memory copy, which is meaningless,
worse more, when hugepg_tbl is mostly in order. This should
be a big deal in large hugepage memory systerm(like hunderd
or thousand GB).
Meanwhile smallest_idx never be a value of -1,so
When vfio module is not loaded when kernel support vfio feature,
the routine still try to open the container to get file
description.
This action is not safe, and of cause got error messages:
EAL: Detected 40 lcore(s)
EAL: unsupported IOMMU type!
EAL: VFIO support could not be initialized
EAL:
On Wed, Dec 10, 2014 at 08:18:36AM +, Wodkowski, PawelX wrote:
> > Though, that said, doesn't it seem to anyone else like serialization of
> > enqueue
> > to a port should be the responsibility of the library, not the application?
> >
> > Neil
>
> From my knowledge it is an application
2014-12-10 11:33, Jincheng Miao:
> From upstream kernel commit 3db2e9cd, strict_strto* serial functions
> are removed. So that we should directly used kstrtoul instead.
>
> And add xen_dom0/compat.h for be compatible with older kernel.
>
> Signed-off-by: Jincheng Miao
Same comments as patch
Hi Jincheng,
2014-12-10 11:33, Jincheng Miao:
> From upstream kernel commit 3db2e9cd, strict_strto* serial functions
> are removed. So that we should directly used kstrtoul instead.
>
> And add kni/compat.h for be compatible with older kernel.
>
> Signed-off-by: Jincheng Miao
[...]
> new file
On Wed, Dec 10, 2014 at 10:32:25AM +, Bruce Richardson wrote:
> On Tue, Dec 09, 2014 at 02:10:32PM -0800, Stephen Hemminger wrote:
> > On Tue, 9 Dec 2014 11:45:07 -0800
> > wrote:
> >
> > > > Hey Folks,
> > > >
> > > > Our DPDK application deals with very large in memory data structures,
>
Hi all,
Any idea?
Leave this issue in DPDK?
Thanks,
Michael
On 12/8/2014 5:07 PM, Qiu, Michael wrote:
> Hi all,
> My platform is:
>
> uname -a
> Linux suse-11-sp3 3.0.77-0.11-xen #1 SMP Tue Mar 11 16:48:56 CST 2014
> x86_64 x86_64 x86_64 GNU/Linux
>
> gcc -v
> Using built-in specs.
>
>
> When vfio module is not loaded when kernel support vfio feature, the
> routine still try to open the container to get file description.
>
> This action is not safe, and of cause got error messages:
>
> EAL: Detected 40 lcore(s)
> EAL: unsupported IOMMU type!
> EAL: VFIO support could not
The patch enhances the tx_checksum command and reworks csum forwarding engine
due to the change of tx_checksum command.
The main changes of the tx_checksum command are listed below,
1. add "tx_checksum set tunnel (hw|sw|none) (port-id)" command
2. add "tx_checksum set outer-ip (hw|sw)
The DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM flag is added in i40e capability set, which
means the i40e supports TX checksum offload of tunneling packet.
Signed-off-by: Jijiang Liu
---
lib/librte_pmd_i40e/i40e_ethdev.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git
If the flag is set in a PMD, which means the NIC(s) support TX checksum offload
of tunneling packet.
Signed-off-by: Jijiang Liu
---
lib/librte_ether/rte_ethdev.h |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/lib/librte_ether/rte_ethdev.h
In the current codes, the "tx_checksum set (ip|udp|tcp|sctp|vxlan) (hw|sw)
(port-id)" command is not easy to understand and extend, so the patch set
enhances the tx_checksum command and reworks csum forwarding engine due to the
change of tx_checksum command.
The main changes of the tx_checksum
Just sent another patch explaining the fix and signed-off.
Thanks,
Ravi
On Wed, Dec 10, 2014 at 4:58 AM, Thomas Monjalon
wrote:
> 2014-12-09 13:05, r k:
>> Subject: [PATCH] Fix power_of_2 macro. Avoid branching when
>> calculating RTE_MIN and RTE_MAX.
>
> Please could you add more explanations
Subject: [PATCH] Fix rte_is_power_of_2() function since it returns
true for zero and zero is not a power of 2. Avoid branching in RTE_MIN
and RTE_MAX when calculating minimum and maximum of two numbers.
Signed-off-by: Ravi Kerur
---
lib/librte_eal/common/include/rte_common.h | 6 +++---
Subject: [PATCH] Simple fix in rte_is_power_of_2 and RTE_MIN/RTE_MAX
Number 0 is not a power_of_2, fix it in rte_is_power_of_2 function.
Avoid code which generates branching when calculating min and max.
Ravi Kerur (1):
Fix rte_is_power_of_2() function since it returns true for zero and
> Though, that said, doesn't it seem to anyone else like serialization of
> enqueue
> to a port should be the responsibility of the library, not the application?
>
> Neil
>From my knowledge it is an application responsibility to serialize access to
queue on particular port.
Pawel
- Original Message -
> 2014-12-10 19:04, Jincheng Miao:
> > And add xen_dom0/compat.h for be compatible with older kernel.
> >
> > Signed-off-by: Jincheng Miao
> [...]
> > new file mode 100644
> > index 000..89dab27
> > --- /dev/null
> > +++
2014-12-10 19:04, Jincheng Miao:
> And add xen_dom0/compat.h for be compatible with older kernel.
>
> Signed-off-by: Jincheng Miao
[...]
> new file mode 100644
> index 000..89dab27
> --- /dev/null
> +++ b/lib/librte_eal/linuxapp/xen_dom0/compat.h
[...]
> ---
Hi Jincheng
I have seen your v3 patch, and ack-ed it. I just dropped my patch for the same
issue. Thank you so much!
Regards,
Helin
> -Original Message-
> From: Jincheng Miao [mailto:jmiao at redhat.com]
> Sent: Wednesday, December 10, 2014 10:55 AM
> To: Zhang, Helin; dev at dpdk.org
Acked-by: Helin Zhang
Note: Don't forget to add the version number to your patch name next time. E.g.
[PATCH v3 0/4]
> -Original Message-
> From: Jincheng Miao [mailto:jmiao at redhat.com]
> Sent: Wednesday, December 10, 2014 11:33 AM
> To: dev at dpdk.org
> Cc: thomas.monjalon at
Hi Thomas,
What's going on with this patch?
I really do not have other better solution.
Thanks,
Michael
On 12/8/2014 9:30 AM, Qiu, Michael wrote:
> On 12/4/2014 9:35 PM, Michael Qiu wrote:
>> app/test-pmd/csumonly.c: In function ?get_psd_sum?:
>> build/include/rte_ip.h:161: error: dereferencing
On 12/10/2014 11:02 AM, Zhang, Helin wrote:
> Hi Jincheng
>
> Did you attach anything? I can see the text only.
> Could you forward your patch mail to me directly? Thanks a lot!
Hi Helin,
I indeed ever saw one patch that was similar this one :) but not sure if
it is post by Jincheng.
You can go
Hi Jincheng
Did you attach anything? I can see the text only.
Could you forward your patch mail to me directly? Thanks a lot!
Regards,
Helin
> -Original Message-
> From: Jincheng Miao [mailto:jmiao at redhat.com]
> Sent: Wednesday, December 10, 2014 10:55 AM
> To: Zhang, Helin; dev at
increase MAX_QUEUES from 256 to 512
In vhost example, MAX_QUEUES macro should be the maximum possible queue number
of the port.
Theoretically we should only set up the queues that are used, i.e., first rx
queue of each pool, or
at most queues from 0 to MAX_QUEUES. Before we revise the
On 12/10/2014 8:17 AM, Xie, Huawei wrote:
> Hi Michael:
>
>> -Original Message-
>> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Qiu, Michael
>> Sent: Monday, December 08, 2014 8:28 AM
>> To: Burakov, Anatoly; dev at dpdk.org
>> Subject: Re: [dpdk-dev] [PATCH v4] VFIO: Avoid to
Hi Michael,
> -Original Message-
> From: Qiu, Michael
> Sent: Wednesday, December 10, 2014 12:03 AM
> To: Ouyang, Changchun; Richardson, Bruce
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [RFC PATCH 00/17] Single virtio implementation
>
> On 2014/12/9 22:19, Ouyang, Changchun wrote:
>
Hi Michael:
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Qiu, Michael
> Sent: Monday, December 08, 2014 8:28 AM
> To: Burakov, Anatoly; dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v4] VFIO: Avoid to enable vfio while the module
> not loaded
>
> On
72 matches
Mail list logo