Hi Bruce,
> From: Rasesh Mody [mailto:rasesh.mody at qlogic.com]
> Sent: Friday, May 06, 2016 9:30 PM
>
> This patch set adds support for enabling 100G mode for QEDE PMD.
> It also adds support for APIs like
> - mtu_set
> - reta_update
> - reta_query
> - rss_hash_update
> - rss_hash_conf_get
> -Original Message-
> From: Xing, Beilei
> Sent: Friday, May 20, 2016 11:17 PM
> To: Wu, Jingjing
> Cc: dev at dpdk.org; Xing, Beilei
> Subject: [PATCH v4] i40e: configure MTU
>
> This patch enables configuring MTU for i40e.
> Since changing MTU needs to reconfigure queue, stop port fir
On Fri, May 20, 2016 at 02:49:31PM +, Mcnamara, John wrote:
> > -Original Message-
> > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Christian Ehrhardt
> > Sent: Friday, May 20, 2016 9:07 AM
> > To: dev ; Stephen Hemminger
> > Subject: Re: [dpdk-dev] Suggestions for the dpdk
Hi,
sorry on my late reply I was on sick leave. Sure I can do that. This fix
was the fastest possible without interfering with DPDK API. I will add
the callback then.
Regards,
Zyta Szpak
On 20.05.2016 10:25, Remy Horton wrote:
> Morning,
>
> On 11/05/2016 11:48, zr at semihalf.com wrote:
>> Fro
--exclude became a positional option in tar 1.29, breaking the
test app filtering in "make install", causing .map files and all test
apps to get installed in bindir. Adjust the tar arguments accordingly,
this is compatible with older versions too since they do not care about
the order.
Ref: https:
Hi Panu, Thomas,
On 05/20/2016 11:09 AM, Thomas Monjalon wrote:
> 2016-05-20 11:42, Panu Matilainen:
>> Just noticed this series "breaks" --no-huge as a regular user, commit
>> 593a084afc2b to be exact:
>>
>> mmap(NULL, 4194304, PROT_READ|PROT_WRITE,
>> MAP_PRIVATE|MAP_ANONYMOUS|MAP_LOCKED, 0, 0
Following the discussions from:
http://dpdk.org/ml/archives/dev/2015-July/021721.html
http://dpdk.org/ml/archives/dev/2016-April/038143.html
The value of these flags is 0, making them useless. Today, no example
application checks them on Rx, and only few drivers sets them and
silently give wrong p
Hi John,
On 05/12/2016 10:36 PM, John Daley (johndale) wrote:
>> ... This is a draft patch that implements what was previously
>> discussed, except the packet_type, which does not exist for vlan
>> today (and I think it is not mandatory for now, let's do it in
>> another patch).
>>
>> After doing
Setup
ThinkServer RD650, Intel Xeon E5-2680 v3
2*x710 ports pass-through (i40e driver)
Dpdk 2.1
Dpdk app doing forwarding from one port to another
Results
When more than 7.5(50% of wire speed) MPPS sent with 64 bytes, Tx
fails(rte_eth_tx_burst return with l
The behavior of PKT_RX_VLAN_PKT was not very well defined, resulting in
PMDs not advertising the same flags in similar conditions.
Following discussion in [1], introduce 2 new flags PKT_RX_VLAN_STRIPPED
and PKT_RX_QINQ_STRIPPED that are better defined:
PKT_RX_VLAN_STRIPPED: a vlan has been stri
Hi Olivier,
> -Original Message-
> From: Olivier Matz [mailto:olivier.matz at 6wind.com]
> Sent: Monday, May 23, 2016 9:47 AM
> To: dev at dpdk.org
> Cc: johndale at cisco.com; Ananyev, Konstantin; Zhang, Helin;
> adrien.mazarguil at 6wind.com; rahul.lakkireddy at chelsio.com;
> alejandro
Hi Konstantin,
On 05/23/2016 10:59 AM, Ananyev, Konstantin wrote:
> Hi Olivier,
>
>> -Original Message-
>> From: Olivier Matz [mailto:olivier.matz at 6wind.com]
>> Sent: Monday, May 23, 2016 9:47 AM
>> To: dev at dpdk.org
>> Cc: johndale at cisco.com; Ananyev, Konstantin; Zhang, Helin;
>
> -Original Message-
> From: Olivier Matz [mailto:olivier.matz at 6wind.com]
> Sent: Monday, May 23, 2016 9:47 AM
> To: dev at dpdk.org
> Cc: johndale at cisco.com; Ananyev, Konstantin; Zhang, Helin;
> adrien.mazarguil at 6wind.com; rahul.lakkireddy at chelsio.com;
> alejandro.lucero at
> -Original Message-
> From: Olivier Matz [mailto:olivier.matz at 6wind.com]
> Sent: Monday, May 23, 2016 10:13 AM
> To: Ananyev, Konstantin; dev at dpdk.org
> Cc: johndale at cisco.com; Zhang, Helin; adrien.mazarguil at 6wind.com;
> rahul.lakkireddy at chelsio.com;
> alejandro.lucero at
Hi,
>>> I don't see ixgbe/i4oe_rxtx_vec.c updated.
>>> Would it be another patch for them?
>>
>> The ixgbe vector and i40e vector do not support vlan stripping,
>
> As I remember, they do.
You are right, I misinterpreted this code in condition_check():
#ifndef RTE_IXGBE_RX_OLFLAGS_ENABLE
>> static inline uint64_t
>> -rx_desc_status_to_pkt_flags(uint32_t rx_status)
>> +rx_desc_status_to_pkt_flags(uint32_t rx_status, uint8_t vlan_strip)
>> {
>> uint64_t pkt_flags;
>> +uint64_t vlan_flags;
>> +
>> +/* if vlan is stripped, set the proper flag */
>> +if (vlan_strip)
On Fri, May 20, 2016 at 03:50:04PM +0300, Ilya Maximets wrote:
> In current implementation guest application can reinitialize vrings
> by executing start after stop. In the same time host application
> can still poll virtqueue while device stopped in guest and it will
> crash with segmentation faul
On 23.05.2016 13:57, Yuanhan Liu wrote:
> On Fri, May 20, 2016 at 03:50:04PM +0300, Ilya Maximets wrote:
>> In current implementation guest application can reinitialize vrings
>> by executing start after stop. In the same time host application
>> can still poll virtqueue while device stopped in gue
Hi,
On 05/19/2016 05:50 PM, Thomas Monjalon wrote:
> 2016-05-19 19:05, Jerin Jacob:
>> On Thu, May 19, 2016 at 12:18:57PM +, Ananyev, Konstantin wrote:
On Thu, May 19, 2016 at 12:20:16AM +0530, Jerin Jacob wrote:
> On Wed, May 18, 2016 at 05:43:00PM +0100, Bruce Richardson wrote:
On 05/19/2016 02:36 PM, Slawomir Mrozowicz wrote:
> Fix issue reported by Coverity.
>
> Coverity ID 13243: Division or modulo by zero
> In function call rte_mempool_xmem_size, division by expression total_size
> which may be zero has undefined behavior.
>
> Fixes: 148f963fb532 ("xen: core libra
On 05/20/2016 05:06 PM, Neil Horman wrote:
[...]
> Look, I think we're simply not going to agree on this issue at all. What
> about
> this in the way of compromise. I simply am not comfortable with automatically
> trying to guess what hardware support will exist in an application based on
> the
qw app at its init stage reserve 2*sizeof(int) memory space for quota
and low_watermark shared variables, but both apps (qw and qwctl) assign
wrong address for low_watermark pointer (out of reserved memzone space)
due to wrong pointer arithmetic.
CID 30709 : Extra sizeof expression (SIZEOF_MISMATC
Calling i40e_switch_tx_queue without checking return value.
Fixed by add warning log information if return failed.
Fixes: 71d35259ff67 ("i40e: tear down flow director")
Coverity ID 13208
Signed-off-by: Slawomir Mrozowicz
---
drivers/net/i40e/i40e_fdir.c | 5 -
1 file changed, 4 insertions(+
Hello David,
please, see my comments inline.
I didn't see the previous versions of the mempool (well, only very roughly) so
I am
probably missing some points... My point of view is as a user of the handler
API.
I need to understand the API to implement a custom handler for my purposes.
On Thu,
On Thu, 19 May 2016 14:45:01 +0100
David Hunt wrote:
> By default, the mempool handler used for mbuf allocations is a multi
> producer and multi consumer ring. We could imagine a target (maybe some
> network processors?) that provides an hardware-assisted pool
> mechanism. In this case, the defau
On Thu, 19 May 2016 14:45:00 +0100
David Hunt wrote:
> Use a minimal custom mempool external handler and check that it also
> passes basic mempool autotests.
>
> Signed-off-by: Olivier Matz
> Signed-off-by: David Hunt
>
> ---
> app/test/test_mempool.c | 113 +++
Hi David,
Please find some comments below.
On 05/19/2016 04:48 PM, David Hunt wrote:
> This is a mempool handler that is useful for pipelining apps, where
> the mempool cache doesn't really work - example, where we have one
> core doing rx (and alloc), and another core doing Tx (and return). In
>
Hi David,
On 05/19/2016 04:48 PM, David Hunt wrote:
> For security, any data structure with function pointers should be const
>
> Signed-off-by: David Hunt
I think this should be merged in the patchset adding the mempool
handler feature.
On Tue, May 17, 2016 at 05:54:01PM +0200, David Marchand wrote:
> > +pci_uio_ioport_map(struct rte_pci_device *dev, int bar,
> > + struct rte_pci_ioport *p)
> > +{
> > + FILE *f;
> > + char buf[BUFSIZ];
> > + char filename[PATH_MAX];
> > + uint64_t phys_addr
On Fri, May 20, 2016 at 11:37:47AM +0100, Bruce Richardson wrote:
> On Thu, May 19, 2016 at 06:44:44PM +0200, Thomas Monjalon wrote:
> > 2016-05-19 17:28, Ferruh Yigit:
> > > On 5/19/2016 9:33 AM, Thomas Monjalon wrote:
> > > > 2016-05-18 18:10, Ferruh Yigit:
> > > >> Add rte_eth_from_vhost() API t
On Thu, May 19, 2016 at 04:20:40PM +, Yoni Gilad wrote:
> Hi,
>
> We have encountered a crash in virtio_xmit_pkts (specifically, in the call to
> virtqueue_notify) when running DPDK in a multi-process setup. This is a
> regression in DPDK 16.04.
>
> The culprit seems to be the field vtpci_o
Hello dpdk dev,
Do you know if the vfio_pci can be bind to network interface from within RedHat
virtual machine ? I read the doc that igb_uio should not be used ; it is not
stable. (http://people.redhat.com/~pmatilai/dpdk-guide/index.html) however I
cannot use vfio_pci driver from inside VM.
Hi Yuanhan,
On 05/23/2016 03:07 PM, Yuanhan Liu wrote:
> On Tue, May 17, 2016 at 05:54:01PM +0200, David Marchand wrote:
>>> +pci_uio_ioport_map(struct rte_pci_device *dev, int bar,
>>> + struct rte_pci_ioport *p)
>>> +{
>>> + FILE *f;
>>> + char buf[BUFSIZ];
>>> +
On Mon, May 09, 2016 at 06:19:35PM +0200, Olivier Matz wrote:
> The commit dd856dfcb9e74 introduced an optimization that prepends virtio
> header to mbuf data. It can be used when the tx mbuf is writeable, so we
> need to check that the mbuf is direct (i.e. it embeds its own data).
>
> Fixes: dd85
On Mon, May 23, 2016 at 02:56:18PM +0300, Panu Matilainen wrote:
> On 05/20/2016 05:06 PM, Neil Horman wrote:
> [...]
> > Look, I think we're simply not going to agree on this issue at all. What
> > about
> > this in the way of compromise. I simply am not comfortable with
> > automatically
> >
This patch docs the issue on EAL argument that the last EAL
argument is replaced by program name in argv[].
Reported-by: Ziye Yang
Signed-off-by: Jingjing Wu
---
doc/guides/rel_notes/known_issues.rst | 21 +
1 file changed, 21 insertions(+)
diff --git a/doc/guides/rel_notes
From: "Jain, Deepak K"
Fix the control issues for return value
Fixes: 924e84f87306 ("aesni_mb: add driver for multi buffer based crypto")
Coverity ID 126585
Signed-off-by: Deepak Kumar Jain
---
drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
d
From: "Jain, Deepak K"
Fix null pointer dereferencing by reporing if null and
exiting the function.
Fixes: c0f87eb5252b ("cryptodev: change burst API to be crypto op oriented")
Coverity issue: 126584
Signed-off-by: Deepak Kumar Jain
---
drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c | 4 +++-
1 f
Sorry for the delayed reply, just sent V2~
On 18 May 2016 at 03:33, Ferruh Yigit wrote:
> On 5/14/2016 7:22 PM, Alex Wang wrote:
> > From: Alex Wang
> >
> > The 'mbufs' alloc/free descriptions for 'rte_kni_tx_burst()'
> > and 'rte_kni_rx_burst()' should be inverted.
> >
> > Signed-off-by: Alex
On 5/21/2016 8:25 AM, Alex Wang wrote:
> From: Alex Wang
>
> The 'mbufs' alloc/free descriptions for
> 'rte_kni_tx_burst()' and 'rte_kni_rx_burst()'
> should be inverted.
>
> Fixes: 3fc5ca2 (kni: initial import)
>
> Signed-off-by: Alex Wang
Acked-by: Ferruh Yigit
On 5/21/2016 8:25 AM, Alex Wang wrote:
> From: Alex Wang
>
> Function like 'rte_kni_rx_burst()' keeps
> allocating 'MAX_MBUF_BURST_NUM' mbufs to
> kni fifo queue unless the queue's capacity
> ('KNI_FIFO_COUNT_MAX') is reached. So, if
> the mempool is under-provisioned, user may
> run into "Out o
On 5/23/2016 2:24 PM, Yuanhan Liu wrote:
> On Fri, May 20, 2016 at 11:37:47AM +0100, Bruce Richardson wrote:
>> On Thu, May 19, 2016 at 06:44:44PM +0200, Thomas Monjalon wrote:
>>> 2016-05-19 17:28, Ferruh Yigit:
On 5/19/2016 9:33 AM, Thomas Monjalon wrote:
> 2016-05-18 18:10, Ferruh Yigit
On 5/23/2016 6:00 PM, Ferruh Yigit wrote:
> On 5/21/2016 8:25 AM, Alex Wang wrote:
>> From: Alex Wang
>>
>> Function like 'rte_kni_rx_burst()' keeps
>> allocating 'MAX_MBUF_BURST_NUM' mbufs to
>> kni fifo queue unless the queue's capacity
>> ('KNI_FIFO_COUNT_MAX') is reached. So, if
>> the mempoo
Converted rte_eth_dev_get_port_by_name to public API.
Signed-off-by: Reshma Pattan
---
lib/librte_ether/rte_ethdev.c | 2 +-
lib/librte_ether/rte_ethdev.h | 15 +++
lib/librte_ether/rte_ether_version.map | 1 +
3 files changed, 17 insertions(+), 1 deletion(-)
dif
This patchset include below changes
1)Changes to librte_ether.
2)New library librte_pdump added for packet capture framework.
3)New app/pdump tool added for packet capturing.
4)Test pmd changes done to initialize packet capture framework.
5)Documentation update.
1)librte_pdump
==
To s
Added spinlocks around add/remove logic of rxtx callbacks to
avoid corruption of callback lists in multithreaded context.
Signed-off-by: Reshma Pattan
---
lib/librte_ether/rte_ethdev.c | 82 +--
1 file changed, 40 insertions(+), 42 deletions(-)
diff --git
New fields nb_rx_queues and nb_tx_queues will be added to
rte_eth_dev_info structure.
Changes to API rte_eth_dev_info_get() will be done to update
these new fields to rte_eth_dev_info object.
Signed-off-by: Reshma Pattan
---
doc/guides/rel_notes/deprecation.rst | 6 ++
1 file changed, 6 inse
New tool added for packet capturing on dpdk.
This tool supports command line options.
This tool runs as secondary process by default.
Command line supports various parameters to capture
the packets.
User should pass on a)port and queue (or) b)pci address
and queue (or) c)device name and queue to
Add new fields to rte_eth_dev_info struct
New fields nb_rx_queues and nb_tx_queues are added to
rte_eth_dev_info structure.
Changes to API rte_eth_dev_info_get() are done to update
these new fields to rte_eth_dev_info object.
Signed-off-by: Reshma Pattan
---
lib/librte_ether/rte_ethdev.c
Added programmers guide for librte_pdump.
Added sample application guide for app/pdump application.
Updated release note for packet capture framework changes.
Signed-off-by: Reshma Pattan
---
MAINTAINERS | 3 +
doc/guides/prog_guide/index.rst | 1 +
doc/gu
Added new public api rte_eth_add_first_rx_callback to add given
callback as head of list.
Signed-off-by: Reshma Pattan
---
lib/librte_ether/rte_ethdev.c | 35 ++
lib/librte_ether/rte_ethdev.h | 27 ++
lib/librte_ether/rte_
Added new library for packet capturing support.
Added public api rte_pdump_init, applications should call
this as part of their application setup to have packet
capturing framework ready.
Added public api rte_pdump_uninit to un initialize the packet
capturing framework.
Added public apis rte_pdu
Call rte_pdump_init and rte_pdump_uninit for packet
capturing initialization and uninitialization.
Signed-off-by: Reshma Pattan
---
app/test-pmd/testpmd.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index 9d11830..645bf50 100644
--- a
On Mon, 23 May 2016 22:38:26 +0100
Reshma Pattan wrote:
> Add new fields to rte_eth_dev_info struct
> New fields nb_rx_queues and nb_tx_queues are added to
> rte_eth_dev_info structure.
> Changes to API rte_eth_dev_info_get() are done to update
> these new fields to rte_eth_dev_info object.
>
>
Sht, sorry for missing that, sending V3,
On Mon, May 23, 2016 at 10:10 AM, Ferruh Yigit
wrote:
> On 5/23/2016 6:00 PM, Ferruh Yigit wrote:
> > On 5/21/2016 8:25 AM, Alex Wang wrote:
> >> From: Alex Wang
> >>
> >> Function like 'rte_kni_rx_burst()' keeps
> >> allocating 'MAX_MBUF_BURST_NUM'
The first 3 patches are related to drop counters. The remaining
patches make up a refactoring, cleanup and optimization of the Tx path.
Changes since v1 are:
- subject line fixups after running check-git-log.sh.
- Errors reported from patchworks fixed. Note: ./scripts/checkpatches.pl was
run
rx_no_bufs is a hardware counter of packets dropped on the
interface due to no host buffers and should be used to update
r_stats->imissed counter instead of rx_nombuf.
Include rx_drop in ierrors. rx_drop is incremented if packets
arrive when the receive queue is disabled.
Add a structure and func
Truncated packets occur on enic if an mbuf is not big enough to
receive it or there aren't enough mbufs if rx scatter is in use.
They show up as error packets but unlike other error packets (like
packets bad FCS) there are no nic drop counts incremented for them.
Truncated packets are calculated by
Following the discussions from:
http://dpdk.org/ml/archives/dev/2015-July/021721.html
http://dpdk.org/ml/archives/dev/2016-April/038143.html
Remove the unused flag from enic driver. Also, the enic driver is
modified to drop bad packets.
Signed-off-by: Olivier Matz
Signed-off-by: John Daley
---
Add an ASSERT macro for the enic driver which is enabled when the log
level is >= RTE_LOG_DEBUG. Assert that number of mbufs to return to
the pool in the Tx function is never greater than the max allowed.
Signed-off-by: John Daley
---
drivers/net/enic/enic.h | 12
drivers/net/e
Functions existed which were never called. Removed them. Also
rename the 'pmd' from the name of the Tx function to improve clarity.
Signed-off-by: John Daley
---
drivers/net/enic/base/vnic_wq.h | 45
drivers/net/enic/enic.h | 5 ++-
drivers/net/enic/enic_ethdev.
Reduce host CPU overhead of Tx packet processing:
* Use local variables inside per packet loop instead of fields in structs.
* Factor book keeping and conditionals out of the per packet loop where
possible.
* Post buffers to the nic at a maximum of every 64 packets
Signed-off-by: Nelson Escobar
Signed-off-by: John Daley
---
drivers/net/enic/Makefile | 2 +-
drivers/net/enic/enic.h| 3 +
drivers/net/enic/enic_ethdev.c | 65 --
drivers/net/enic/enic_main.c | 82 +--
drivers/net/enic/enic_rx.c | 343 -
drivers/net/enic/enic_rxtx.
The list of mbufs held by the driver on Tx was allocated in chunks
(a hold-over from the enic kernel mode driver). The structure used
next pointers across chunks which led to cache misses.
Allocate the array used to hold mbufs in flight on Tx with
rte_zmalloc_socket(). Remove unnecessary fields fr
The NIC can either DMA a separate completion message for each
completed send or periodically just DMA an index of the last
completed send. Switch to the second method which improves
cache locality and performance.
Signed-off-by: John Daley
---
drivers/net/enic/base/vnic_wq.c | 1 +
drivers/net/
Mbufs were returned to the pool one at a time. Use rte_mempool_put_bulk
instead. There were muiltiple function calls for each buffer returned.
Refactor this code into just 2 functions.
Signed-off-by: John Daley
---
drivers/net/enic/base/vnic_wq.h | 27 -
drivers/net/enic/enic
Remove some files, functions and variables left unused after
Tx performance improvements.
Signed-off-by: John Daley
---
drivers/net/enic/base/enic_vnic_wq.h | 63
drivers/net/enic/base/vnic_cq.h | 44 -
drivers/net/enic/base/vnic_
67 matches
Mail list logo
