Linux kernel provides UIO as well as VFIO mechanism to support writing user
space device driver. Comparing to UIO which is available since 2.6.32 kernel,
the VFIO is introduced into kernel since version 3.6.0 with better interrupt
and memory protection (build on top of Intel VT-d technology) suppor
Hi,
My setup:
Qemu+KVM
Host OS: CentOS, Linux kernel version 2.6.32
Guest OS: Linux kernel version 3.10.0, with a DPDK application running.
Physical NIC: Intel I350 T4 using igb driver
On guest VM: 2 (out of 4) ports of the I350 are virtualized using
PCI-passthrough (no VFs, SR-IOV is not enabled
Hi Thomas
Sorry, I should have answer your comments together with my reworking the latest
patch set.
Please see my answers for your comments! I really appreciate your comments!
> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Wednesday, October 2
Hi Thomas
See my answers inlined.
> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Monday, October 27, 2014 10:24 PM
> To: Zhang, Helin
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v3 7/8] ethdev: support of multiple sizes of
> redirecti
Acked-by: Helin Zhang
I have reviewed the differences between v3 and v4. It seems OK for me.
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Jingjing Wu
> Sent: Wednesday, October 22, 2014 9:01 AM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH v4 00/21]
Hi, Thomas
> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Tuesday, October 28, 2014 12:58 AM
> To: Wu, Jingjing
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v4 04/21] ethdev: define structures for
> adding/deleting flow director
>
> 20
Tested-by: Michael Qiu
? 10/28/2014 1:49 AM, Danny Zhou ??:
> Linux kernel provides UIO as well as VFIO mechanism to support writing user
> space device driver. Comparing to UIO which is available since 2.6.32 kernel,
> the VFIO is introduced into kernel since version 3.6.0 with better interrupt
On Mon, Oct 27, 2014 at 07:49:03PM +0200, Kevin Wilson wrote:
> Poke!
> Can anybody advice about this question ?
> Kevin
>
> On Fri, Oct 24, 2014 at 12:54 PM, Kevin Wilson wrote:
> > Hi,
> >
> > I am looking in the file hierarchy of dpdk, and I see that under
> > /dpdk-1.7.1/lib/librte_eal/linuxa
2014-10-28 00:37, Zhang, Helin:
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > 2014-10-22 19:53, Helin Zhang:
> > > +#define RTE_BIT_WIDTH_64 (CHAR_BIT * sizeof(uint64_t))
> >
> > How can it be different of 64?
> > Using 64 would be simpler to understand than RTE_BIT_WIDTH_64.
>
2014-10-28 00:33, Zhang, Helin:
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > 2014-09-25 16:40, Helin Zhang:
> > > /* Definitions used for redirection table entry size */
> > > -#define ETH_RSS_RETA_NUM_ENTRIES 128
> > > -#define ETH_RSS_RETA_MAX_QUEUE 16
> > > +#define ETH_R
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Thomas Monjalon
> Sent: Tuesday, October 28, 2014 10:10 AM
> To: Zhang, Helin
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2 04/13] ethdev: support of multiple sizes of
> redirection table
>
> 2014-10
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Sergio Gonzalez
> Monroy
> Sent: Thursday, October 23, 2014 4:36 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] mk: link combined lib using CC
>
> Building combined shared libs fails if we set EXTRA_CF
Hi Erik,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Erik Ziegenbalg
> Sent: Saturday, October 25, 2014 4:19 AM
> To: dev at dpdk.org
> Subject: [dpdk-dev] rte_acl test-acl app
>
> Hi everyone,
>
> I am having trouble to successfully perform a packet cl
On Wed, Oct 22, 2014 at 12:17:01PM +0100, Pablo de Lara wrote:
> rte_ipv4_fragment_packet() and rte_ipv6_fragment packet()
> call rte_pktmbuf_attach() to attach the segment of the original
> packet to the segment of the new fragmented one. Such function
> is not declared if RTE_MBUF_REFCNT is disab
Hi Thomas
> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Tuesday, October 28, 2014 6:10 PM
> To: Zhang, Helin
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2 04/13] ethdev: support of multiple sizes of
> redirection table
>
> 2014-10-2
2014-10-28 12:00, Zhang, Helin:
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > 2014-10-28 00:33, Zhang, Helin:
> > > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > > > 2014-09-25 16:40, Helin Zhang:
> > > > > /* Definitions used for redirection table entry size
> -Original Message-
> From: Mario Gianni [mailto:m.gianni at engineer.com]
> Sent: Friday, October 24, 2014 4:03 PM
> To: Richardson, Bruce
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] Cannot mmap device resource in DPDK 1.7.0 multi-
> process/multi-thread
>
> So you are telling me tha
On Fri, Oct 10, 2014 at 08:29:57PM +0800, Cunming Liang wrote:
> v2 update:
> Rebase code to the latest master branch.
>
> --
>
> Tested-by: Zhaochen Zhan
> This patch has been verified on ixgbe and it is ready to be integrated to
> dpdk.org.
> For e1000 and i40e, lacking loopback s
Hi Thomas
> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Tuesday, October 28, 2014 8:13 PM
> To: Zhang, Helin
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2 04/13] ethdev: support of multiple sizes of
> redirection table
>
> 2014-10-2
The set of patches split x86 architecture specific operations from DPDK and put
them to x86 arch directory.
This will make the adoption of DPDK much easier on other computer architecture.
For a new architecture, just add an architecture specific directory and
necessary building configuration files,
This breaks compilation but is a preliminary step for headers rework.
Signed-off-by: David Marchand
---
lib/librte_eal/common/Makefile |4 +-
.../common/include/arch/i686/rte_atomic.h | 373
.../common/include/arch/x86_64/rte_atomic.h
From: Chao Zhu
This patch first adds architecture specific directories to eal.
Then split the atomic operations to architecture specific and generic files.
Architecture specific files are put into the corresponding architecture
directory and common header are put into generic directory.
Update d
From: Chao Zhu
This patch splits the byte order operations from DPDK and push them to
architecture specific arch directories, so that other processor
architecture to support DPDK can be easily adopted.
Signed-off-by: Chao Zhu
Signed-off-by: David Marchand
---
lib/librte_eal/common/Makefile
From: Chao Zhu
This patch splits the CPU TSC read operations from DPDK and push them to
architecture specific arch directories, so that other processors that
don't have tsc register can implement its own functions.
Signed-off-by: Chao Zhu
Signed-off-by: David Marchand
---
lib/librte_eal/commo
From: Chao Zhu
This patch splits the prefetch operations from DPDK and push them to
architecture specific arch directories, so that other processor
architecture to support DPDK can implement their own functions.
Signed-off-by: Chao Zhu
Signed-off-by: David Marchand
---
lib/librte_eal/common/M
From: Chao Zhu
This patch splits the spinlock operations from DPDK and push them to
architecture specific arch directories, so that other processor
architecture to support DPDK can be easily adopted.
Signed-off-by: Chao Zhu
Signed-off-by: David Marchand
---
lib/librte_eal/common/Makefile
From: Chao Zhu
This patch splits CPU flags related operations from DPDK and push them
to architecture specific arch directories, so that other processor
architecture can implement its own CPU flag functions to support DPDK.
Signed-off-by: Chao Zhu
Signed-off-by: David Marchand
---
lib/librte_
From: Chao Zhu
This patch splits the SSE based memory copy function from DPDK and push
them to architecture specific arch directories. Other processor
architecture can implement its own vector based memory copy functions.
Signed-off-by: Chao Zhu
Signed-off-by: David Marchand
---
lib/librte_ea
Architecture can have their own specific headers, just install all headers from
arch directory.
Signed-off-by: David Marchand
---
lib/librte_eal/common/Makefile |6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/lib/librte_eal/common/Makefile b/lib/librte_eal/common/Mak
No need to keep the same code duplicated for 32 and 64bits x86.
Signed-off-by: David Marchand
---
.../common/include/arch/i686/rte_atomic.h | 393
.../common/include/arch/i686/rte_byteorder.h | 129 ---
.../common/include/arch/i686/rte_cpuflags.h
2014-10-28 01:18, Wu, Jingjing:
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > 2014-10-22 09:01, Jingjing Wu:
> > > +#define RTE_ETH_FDIR_MAX_FLEXWORD_LEN 8
> > > +/**
> > > + * A structure used to contain extend input of flow */ struct
> > > +rte_eth_fdir_flow_ext {
> > > + ui
Hi Bruce
> -Original Message-
> From: Richardson, Bruce
> Sent: Tuesday, October 28, 2014 6:18 PM
> To: Thomas Monjalon; Zhang, Helin
> Cc: dev at dpdk.org
> Subject: RE: [dpdk-dev] [PATCH v2 04/13] ethdev: support of multiple sizes of
> redirection table
>
> > -Original Message-
2014-10-22 09:01, Jingjing Wu:
> --- a/app/test-pmd/testpmd.h
> +++ b/app/test-pmd/testpmd.h
> @@ -73,6 +73,9 @@ int main(int argc, char **argv);
> #define NUMA_NO_CONFIG 0xFF
> #define UMA_NO_CONFIG 0xFF
>
> +#define BYTES_PER_WORD 2
This is an example of a constant which have a meaning.
Bu
2014-10-22 09:01, Jingjing Wu:
> extend fdir field to support flex bytes reported when fdir match
The commit log should explain why it is required (i40e?).
It will help to understand when digging into git history of mbuf file.
> --- a/lib/librte_mbuf/rte_mbuf.h
> +++ b/lib/librte_mbuf/rte_mbuf.h
2014-10-22 09:01, Jingjing Wu:
> +/**
> + * A structure used to report the status of the flow director filters in use.
> + */
> +struct rte_eth_fdir {
> + /** Number of filters with collision indication. */
> + uint16_t collision;
> + /** Number of free (non programmed) filters. */
> +
2014-10-22 09:01, Jingjing Wu:
> + "flush_flow_diretor (port_id)\n"
Typo: diretor
I know it's hard to write so much code in a small timeframe without any typo.
Especially when you are not english native.
But I'd like such simple errors would have been already reviewed
after t
2014-10-22 09:01, Jingjing Wu:
> +/**
> + * A structure defined a field vector to specify each field.
> + */
> +struct rte_eth_field_vector {
> + uint8_t offset; /**< Source word offset */
> + uint8_t size; /**< Field Size defined in word units */
> +};
I'm sorry but I don't understa
2014-10-22 09:01, Jingjing Wu:
> add test command to configure flexible masks for each flow type
[...]
> --- a/app/test-pmd/cmdline.c
> +++ b/app/test-pmd/cmdline.c
> + "flow_director_flex_mask (port_id)"
> + " flow (ether|ip4|tcp4|udp4|sctp4|ip6|tcp6|udp6|sc
2014-10-28 13:20, Zhang, Helin:
> From: Richardson, Bruce
> > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Thomas Monjalon
> > > 2014-10-28 00:33, Zhang, Helin:
> > > > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > > > > 2014-09-25 16:40, Helin Zhang:
> > > > > > /* De
Currently I have an issue in a pass-thorugh mode. I am running in multi process
(2 processes) via ixgbe_rxtx. When running with a single process I have a
perfect traffic while with 2 processes I have no traffic at all.
Perhaps there is a special configuration for that, can you please assist ?
T
Hi Yong,
Is there any progress with this patchset?
Thanks
--
Thomas
2014-10-22 07:07, Cao, Waterman:
> Hi Yong,
>
> We verified your patch with VMWare ESXi 5.5 and found VMware L2fwd and
> L3fwd cmd can't run.
> But We use DPDK1.7_rc1 package to validate VMware regression, It works
2014-10-28 14:38, Yan Freedland:
> Currently I have an issue in a pass-thorugh mode.
> I am running in multi process (2 processes) via ixgbe_rxtx.
> When running with a single process I have a perfect traffic
> while with 2 processes I have no traffic at all.
> Perhaps there is a special configura
On Tue, Oct 28, 2014 at 11:45:28AM +, De Lara Guarch, Pablo wrote:
>
>
> > -Original Message-
> > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Sergio Gonzalez
> > Monroy
> > Sent: Thursday, October 23, 2014 4:36 PM
> > To: dev at dpdk.org
> > Subject: [dpdk-dev] [PATCH] mk:
Hi Alan,
Did you make any progress in Qemu/KVM community?
We need to be sync'ed up with them to be sure we share the same goal.
I want also to avoid using a solution which doesn't fit with their plan.
Remember that we already had this problem with ivshmem which was planned
to be dropped.
Thanks
-
2014-10-28 14:51, Sergio Gonzalez Monroy:
> On Tue, Oct 28, 2014 at 11:45:28AM +, De Lara Guarch, Pablo wrote:
> > > Building combined shared libs fails if we set EXTRA_CFLAGS=-O0.
> > >
> > > /usr/bin/ld: test: hidden symbol `mknod' in
> > > /usr/lib64/libc_nonshared.a(mknod.oS) is referenced
On Tue, Oct 28, 2014 at 04:33:18PM +0100, Thomas Monjalon wrote:
> 2014-10-28 14:51, Sergio Gonzalez Monroy:
> > On Tue, Oct 28, 2014 at 11:45:28AM +, De Lara Guarch, Pablo wrote:
> > > > Building combined shared libs fails if we set EXTRA_CFLAGS=-O0.
> > > >
> > > > /usr/bin/ld: test: hidden
If we set EXTRA_CFLAGS=-O0, build fails with following error:
/usr/bin/ld: test: hidden symbol `mknod' in
/usr/lib64/libc_nonshared.a(mknod.oS) is referenced by DSO
Fix: link combined shared lib using CC if LINK_USING_CC is enabled.
Signed-off-by: Sergio Gonzalez Monroy
---
mk/rte.lib.mk
> -Original Message-
> From: Gonzalez Monroy, Sergio
> Sent: Tuesday, October 28, 2014 3:49 PM
> To: dev at dpdk.org
> Cc: thomas.monjalon at 6wind.com; De Lara Guarch, Pablo
> Subject: [PATCH v2] mk: link combined shared lib using CC
>
> If we set EXTRA_CFLAGS=-O0, build fails with foll
Hi,
I mistakenly pressed on reply all instead of reply and sent my question to the
whole dev. Sorry about that.
In fact I am not lazy and looks like I solved the problem by myself. I will
gladly share my knowledge with you or anyone in dev and of course open a new
mail thread with an appropria
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Sergio Gonzalez
> Monroy
> Sent: Wednesday, October 22, 2014 5:36 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] mk: fix build 32bits shared libs on 64bits system
>
> Incompatible libraries error when
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Sergio Gonzalez
> Monroy
> Sent: Thursday, October 23, 2014 4:37 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] mk: fix app linking for combined libs
>
> Building combined shared libraries results in a
Hi
In my multi process system I need to support 2 modes of work: pass through and
VIRTIO.
I saw that in order to work in a VIRTIO mode I need to update the txq_flags
(part of the rte_eth_txconf structure) as follows:
.txq_flags = ETH_TXQ_FLAGS_NOMULTSEGS | ETH_TXQ_FLAGS_NOOFFLOADS
In pass - thr
Each device can tell you the driver name its bound to . This is enough to
have separate configuration paths for vmxnet/virtio/ixgbe/etc...
On Tue, Oct 28, 2014 at 7:21 PM, Yan Freedland wrote:
> Hi
>
> In my multi process system I need to support 2 modes of work: pass through
> and VIRTIO.
> I s
Thomas/Waterman,
I couldn't reproduce the reported issue on v1.8.0-rc1 and both l2fwd and l3fwd
works fine using the same command posted.
# dpdk_nic_bind.py --status
Network devices using DPDK-compatible driver
:0b:00.0 'VMXNET3 Ethernet Controll
On 10/22/14, 6:39 AM, "Stephen Hemminger"
wrote:
>On Mon, 13 Oct 2014 18:42:18 +
>Yong Wang wrote:
>
>> Are you referring to the patch as a whole or your comment is about the
>>reset of vlan_tci on the "else" (no vlan tags stripped) path? I am not
>>sure I get your comments here. This pat
Hi, im trying the kni example. When i hit ctrl-c in the terminal stopping
the example, the os signals with a segmentation fault.
The issue is in this block of code:
static int kni_free_kni(uint8_t port_id) {
uint8_t i;
struct kni_port_params **p = kni_port_params_array;
if (port_id >= RTE_MAX_
On Wed, Oct 29, 2014 at 03:27:58AM +, Qiu, Michael wrote:
> I just saw one return path with value '0', and no any other place
> return a negative value, so it is better to be designed as one
> non-return function,
>
> +void
> +rte_eal_hugepage_free(void)
> +{
> + struct hugepage_file *h
57 matches
Mail list logo
