Fixes: ac5e1d838dc1 ("virtio: skip error when probing kernel managed device")
Signed-off-by: Huawei Xie
---
drivers/net/virtio/virtio_pci.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/net/virtio/virtio_pci.c b/drivers/net/virtio/virtio_pci.c
index
We keep a common vq structure, containing only vq related fields,
and then split others into RX, TX and control queue respectively.
Signed-off-by: Huawei Xie
---
drivers/net/virtio/virtio_ethdev.c | 352 ++--
drivers/net/virtio/virtio_ethdev.h | 2 +-
This patch fixes rte_memcpy performance in Haswell and Broadwell for
vhost when copy size larger than 256 bytes.
It is observed that for large copies like 1024/1518 ones, rte_memcpy
suffers high ratio of store buffer full issue which causes pipeline
to stall in scenarios like vhost enqueue. This
On Thu, May 19, 2016 at 02:44:59PM +0100, David Hunt wrote:
> Until now, the objects stored in mempool mempool were internally stored a
> ring. This patch introduce the possibility to register external handlers
> replacing the ring.
>
> The default behavior remains unchanged, but calling the new
Problem:
The flag for RSS and flow director is not set correctly in the
vector RX function, so the upper layer APP which base on the related
flags will not work correctly.
Fix this problem by change the shuffle table. the original shuffle
table is not correct.
Fixes: 9ed94 (i40e: add vector Rx)
This patch is used to add the class_id (class_code,
subclass_code, programming_interface) support for
pci_device probe. With this patch, it will be
flexible for users to probe a class of devices
by class_id.
Signed-off-by: Ziye Yang
---
Changes in v4: adjust title name and change RTE_PCI_DEVICE
On Tue, May 24, 2016 at 04:59:47PM +0200, Olivier Matz wrote:
> Hi Jerin,
>
>
> On 05/24/2016 04:50 PM, Jerin Jacob wrote:
> > Signed-off-by: Jerin Jacob
> > ---
> > lib/librte_mempool/rte_mempool.h | 5 ++---
> > 1 file changed, 2 insertions(+), 3 deletions(-)
> >
> > diff --git
Signed-off-by: Jerin Jacob
---
lib/librte_mempool/rte_mempool.h | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/lib/librte_mempool/rte_mempool.h b/lib/librte_mempool/rte_mempool.h
index ed2c110..ebe399a 100644
--- a/lib/librte_mempool/rte_mempool.h
+++
On 5/24/2016 1:52 PM, Burakov, Anatoly wrote:
>> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Anatoly Burakov
>> Sent: Tuesday, May 24, 2016 1:50 PM
>> To: dev at dpdk.org
>> Subject: [dpdk-dev] [PATCH] ivshmem: fix overlap detection code
>>
>> Partial revert of an earlier ill-conceived
Any news Tomasz, Cristian?
2016-05-11 10:15, Dumitrescu, Cristian:
> CC-ing Tomasz, who is the original author of RED implementation and its
> autotest. Tomasz, what do you think?
>
> > -Original Message-
> > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > Sent:
> -Original Message-
> From: Singh, Jasvinder
> Sent: Friday, May 6, 2016 8:09 PM
> To: dev at dpdk.org
> Cc: Dumitrescu, Cristian ; Chokkalingam,
> SankarX
> Subject: [PATCH] ip_pipeline: add command for multiple execution of run
>
>
> From: Sankar Chokkalingam
>
> The new command
2016-05-11 16:26, Thomas Monjalon:
> In order to make autotests easy to run often,
> the time and memory constraints are reduced.
Applied
There is some remaining work:
> Red autotest: Success [01m 36s] [02m 13s]
> The RED autotest needs some work.
2016-05-11 14:16, Thomas Monjalon:
> The LPM performance tests should not be mixed with
> functional tests: they are long and show some statistics.
>
> LPM6 functional tests are added to parallel tests.
>
> Signed-off-by: Thomas Monjalon
> ---
> v2: move LPM6 autotest in parallel tests
>
>
2016-05-17 09:05, Thomas Monjalon:
> The log unit test was checking display of low priority messages.
> It was not working if RTE_LOG_LEVEL is not RTE_LOG_DEBUG.
> It is even easier to see since the default level is INFO (9b9d7ca).
>
> Now the test use ERR and CRIT levels which should be always
On Tue, May 24, 2016 at 05:10:01PM +0100, Bruce Richardson wrote:
> On Fri, May 06, 2016 at 11:55:44AM +0530, Jianbo Liu wrote:
> > Implement ixgbe vPMD on ARM with NEON intrinsic.
> >
> > v3:
> > - rebase to rel_16_07 branch on dpdk-next-net.
> >
> > v2:
> > - move the common code to new
On Fri, May 06, 2016 at 11:55:44AM +0530, Jianbo Liu wrote:
> Implement ixgbe vPMD on ARM with NEON intrinsic.
>
> v3:
> - rebase to rel_16_07 branch on dpdk-next-net.
>
> v2:
> - move the common code to new header file.
>
> Jianbo Liu (4):
> ixgbe: rearrange vector PMD code for x86
>
2016-05-11 11:46, Thomas Monjalon:
> We can define any custom whitelist of tests to run.
> The predefined lists for ring and mempool tests are not quite useful.
>
> Signed-off-by: Thomas Monjalon
> ---
> v2: completely remove targets
Series applied
Hi Thomas,
>From my side, I am OK to remove RED from the fast autotest, as long as it is
>kept available as part of the normal/full autotest of DPDK.
Some of the RED autotests need a long time to run in order to train the history
for the average queue size stochastic variable, therefore it is
Hi Jerin,
On 05/24/2016 04:50 PM, Jerin Jacob wrote:
> Signed-off-by: Jerin Jacob
> ---
> lib/librte_mempool/rte_mempool.h | 5 ++---
> 1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/lib/librte_mempool/rte_mempool.h
> b/lib/librte_mempool/rte_mempool.h
> index
2016-05-10 21:08, Yuanhan Liu:
> Patch 1 fix a false positive warning; patch 2 helps to spot bad content
> clearly.
>
> ---
> Yuanhan Liu (2):
> scripts: fix false positive warning
> scripts: high light bad patterns
The justification of these patches is the word "ctx".
It should be written
2016-04-19 15:16, Thomas Monjalon:
> It was suggesting to use a Linux macro:
> CHECK:BIT_MACRO: Prefer using the BIT macro
>
> It is ignored because it is a false positive in DPDK case.
>
> Signed-off-by: Thomas Monjalon
Applied
By the way, maybe that someone would like to implement RTE_BIT?
2016-05-20 08:24, Panu Matilainen:
> On 05/19/2016 09:05 PM, Thomas Monjalon wrote:
> > This script was forgotten when dropping the combined library.
> >
> > Fixes: 948fd64befc3 ("mk: replace the combined library with a linker
> > script")
> >
> > Signed-off-by: Thomas Monjalon
> > ---
>
> I
2016-05-24 13:28, Panu Matilainen:
> The qede driver depends on libz but the LDLIBS entry in makefile
> was missing. Also because of the external dependency, make it
> disabled in default config as per common DPDK policy on external deps.
>
> Fixes: ec94dbc57362 ("qede: add base driver")
>
>
2016-05-19 18:18, Harish Patil:
> >
> >Some 64-bit variables are printed for debug.
> >%PRIx64 qualifier must be used because %lx is not long enough
> >on 32-bit systems
> >
> >Signed-off-by: Thomas Monjalon
>
> Looks good, thanks.
>
> Acked-by: Harish Patil
Applied
PS: please remove useless
This tool searches for the primer sting PMD_DRIVER_INFO= in any ELF binary,
and, if found parses the remainder of the string as a json encoded string,
outputting the results in either a human readable or raw, script parseable
format
Note that, in the case of dynamically linked applications,
Modify the compilation makefile to identify C files that export PMD information,
and use that to trigger execution of the pmdinfo binary. If the execution of
pmdinfo is successful, compile the output C file to an object, and use the
linker to do relocatable linking on the resultant object file
Export a symbol containing the string:
DPDK_PLUGIN_PATH="$(CONFIG_RTE_EAL_PMD_PATH)"
Where the latter half of the string is set at build time to a location from
which autoloaded DSO's will be found. This string is used by pmdinfo in
'plugin' mode, whereby a user can specify a dpdk installation
Modify the PMD_REGISTER_DRIVER macro, adding a name argument to it. The
addition of a name argument creates a token that can be used for subsequent
macros in the creation of unique symbol names to export additional bits of
information for use by the pmdinfogen tool. For example:
pmdinfogen is a tool used to parse object files and build json strings for use
in
later determining hardware support in a dso or application binary. pmdinfo
looks for the non-exported symbol names this_pmd_name and this_pmd_tbl
(where n is a integer counter). It records the name of each of
Hey all-
So heres attempt number 2 at a method for exporting PMD hardware support
information. As we discussed previously, the consensus seems to be that pmd
information should be:
1) Able to be interrogated on any ELF binary (application binary or individual
DSO)
2) Equally functional
2016-05-18 21:15, Mauricio Vasquez B:
> The RTE_ETH_VALID_PORTID_OR_ERR_RET macro is used in some places
> to check if a port id is valid or not. This commit makes use of it in
> some new parts of the code.
>
> Signed-off-by: Mauricio Vasquez B studenti.polito.it>
> ---
> v3:
> - use it also in
On 5/23/2016 1:35 PM, Jan Viktorin wrote:
> 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
On 5/23/2016 1:55 PM, Olivier Matz wrote:
> Hi David,
>
> On 05/19/2016 04:48 PM, David Hunt wrote:
>> For security, any data structure with function pointers should be const
> I think this should be merged in the patchset adding the mempool
> handler feature.
Agreed, I'm working on that now,
Olivier,
The patch set looks good. I think little endian is good enough. We
internally decide to only support little endian previously. You can keep the
endian conversion such kind of things in the code.
-Original Message-
From: Olivier Matz [mailto:olivier.m...@6wind.com]
Sent:
It adds a flag, which can be used to tell the firmware to
disable the link on all ports.
Signed-off-by: Helin Zhang
---
doc/guides/rel_notes/release_16_07.rst | 4
drivers/net/i40e/Makefile | 2 +-
drivers/net/i40e/base/i40e_adminq_cmd.h | 3 +++
3 files changed, 8
It add opcodes and structures to support RSS configuration
by PF driver on behalf of the VF drivers.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_virtchnl.h | 45 ---
1 file changed, 42 insertions(+), 3 deletions(-)
diff --git
It adds input set mask definitions for RSS, flow director
and flex bytes.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_type.h | 33 +
1 file changed, 33 insertions(+)
diff --git a/drivers/net/i40e/base/i40e_type.h
It increases the supported AQ API version to 1.5
for X722.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_type.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/i40e/base/i40e_type.h
b/drivers/net/i40e/base/i40e_type.h
index 7ed3048..bc68b47 100644
It adds more device capabilities for NVM management.
- if update is available
- if security check is needed
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_adminq_cmd.h | 1 +
drivers/net/i40e/base/i40e_common.c | 6 ++
drivers/net/i40e/base/i40e_type.h | 5 +
3
It fixes the debug output messages.
Fixes: f388b435bc33 ("i40e/base: clean adminq debug")
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_common.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/i40e/base/i40e_common.c
It corrects the number of MSIX vector in a debug info.
Fixes: 889bc9f0cd3a ("i40e/base: unify the capability function")
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_common.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/i40e/base/i40e_common.c
It adds new device IDs and PHY types.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_adminq_cmd.h | 4
drivers/net/i40e/base/i40e_common.c | 4
drivers/net/i40e/base/i40e_devids.h | 4
lib/librte_eal/common/include/rte_pci_dev_ids.h | 8
It removes the problematic mirror rule ID check. It
returns an error if the mirror rule ID is 0, which is
a valid value.
Fixes: 0bf2dbbe077c ("i40e/base: support mirroring rules")
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_common.c | 5 +
1 file changed, 1 insertion(+), 4
It exposes the configuration of mirroring or not egress
traffic to VSIs in promiscuous mode, as latest firmware
supports that from API version 1.5.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_common.c| 9 ++---
drivers/net/i40e/base/i40e_prototype.h | 4 ++--
The hardware doesn't layout the Geneve VNI (Virtual Network
Identifier) quite the same as the VxLAN VNI, so it needs to
adjust it before sending through the AQ commands as the
workaround.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_common.c | 35 ++-
It trim the source code, with limiting pieces of code for
PF or VF driver only, code style fixes, and annotation
rewording.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_adminq.c | 52 +++--
drivers/net/i40e/base/i40e_adminq.h | 4 +--
It refactors the NVM update command processing, with adding
a new element of nvm_wait_opcode in struct i40e_hw to indicate
the opcode it waits on, and putting the wait event check into
a function. In addition, that element needs to be initialized
or updated properly.
Signed-off-by: Helin Zhang
It centralizes all NVM update status info into a single
structure, by moving nvm_release_on_done from struct
i40e_adminq_info to struct i40e_hw, for better management.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_adminq.c | 6 +++---
drivers/net/i40e/base/i40e_adminq.h | 1 -
HMC AQ APIs were removed from the latest datasheet, and
hence remove its implementations and relevant.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/base/i40e_adminq.c | 4 ---
drivers/net/i40e/base/i40e_adminq_cmd.h | 25 --
drivers/net/i40e/base/i40e_common.c | 61
This is the i40e base driver update, which includes bug fixes,
enhancements, refactoring, and new device enabling. Below are
the details.
v2:
- reworded commit titles and logs.
Helin Zhang (15):
i40e/base: remove HMC AQ APIs
i40e/base: move field of NVM update status info
i40e/base:
In this scenario, kernel PF + DPDK VF, when PF finds the link
state is changed, up -> down or down -> up, it will send a
mailbox message to VF.
This patch enables the support of the interruption of mailbox,
so VF can receive the message for link up/down.
Signed-off-by: Wenzhuo Lu
---
In this scenario, kernel PF + DPDK VF, when PF finds the link
state is changed, up -> down or down -> up, it will send a
mailbox message to VF.
This patch enables the support of the interruption of mailbox,
so VF can receive the message for link up/down.
Signed-off-by: Wenzhuo Lu
---
This patch set addes the support of the mailbox interruption on VF.
So, VF can receice the messges for physical link down/up.
PS: This patch set is splitted from a previous patch set, *automatic
link recovery on ixgbe/igb VF*.
Wenzhuo Lu (2):
ixgbe: VF supports mailbox interruption for PF link
Partial revert of an earlier ill-conceived "fix".
Adjacent segments can never be considered overlapping because we
are not comparing ends to starts, but rather starts to starts.
Therefore the earlier fix was wrong (plus it also had a typo).
Signed-off-by: Anatoly Burakov
---
The qede driver depends on libz but the LDLIBS entry in makefile
was missing. Also because of the external dependency, make it
disabled in default config as per common DPDK policy on external deps.
Fixes: ec94dbc57362 ("qede: add base driver")
Signed-off-by: Panu Matilainen
---
On Mon, May 23, 2016 at 03:40:58PM +0200, Olivier Matz wrote:
> For reference, here is the report of the ABI checker for EAL:
>
> [?] struct rte_pci_ioport (2)
>
> 1 Field len has been added to this type.
>1) This field will not be initialized by old clients.
>2) Size of the inclusive
On Mon, May 23, 2016 at 06:06:21PM +0100, Ferruh Yigit wrote:
> 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
On 05/20/2016 08:08 PM, Thomas Monjalon wrote:
> 2016-05-20 18:50, Christian Ehrhardt:
>> The individual libraries have various cross dependencies.
>> This is already refelcted in the DEPDIR dependency, but not yet in
>> proper DT_NEEDED flags in the .so's.
>> This adds the -l flags so that is
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Anatoly Burakov
> Sent: Tuesday, May 24, 2016 1:50 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] ivshmem: fix overlap detection code
>
> Partial revert of an earlier ill-conceived "fix".
> Adjacent segments can never be considered
On Wed, 25 May 2016 01:28:03 +0800
Zhe Tao wrote:
> This patch-set add the support for floating VEB in i40e.
> All the VFs VSIs can decide whether to connect to the legacy VEB/VEPA or
> the floating VEB. When connect to the floating VEB a new floating VEB is
> created. Now all the VFs need to
Hi Panu,
I already agreed to Thomas on IRC but won't have time until next week.
Thanks for making a patch that does that already - I'll give it a look and
some test on my end next week then.
Christian Ehrhardt
Software Engineer, Ubuntu Server
Canonical Ltd
On Tue, May 24, 2016 at 11:56 AM,
Hi All,
After successful deployment of OVS (2.4) -DPDK (2.0) on HP GEN 7 blade ,
I tried to activate it also on HP GEN9 (BL460C), got DMA error ("PTE Read
access is not set" , see attached file).
I found several discussions describing similar scenarios (for HP GEN8),
On 05/20/2016 08:24 PM, Neil Horman wrote:
> This tool searches for the primer sting PMD_DRIVER_INFO= in any ELF binary,
> and, if found parses the remainder of the string as a json encoded string,
> outputting the results in either a human readable or raw, script parseable
> format
>
> Note that,
Hi Tao,
On 23/05/2016 14:39, DING, TAO wrote:
> 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.
Hi Ziye,
Please check the title.
It could be "pci: add class id".
Please try to add a changelog below the 3 dashes when making a new revision.
Other comments below:
2016-05-19 21:17, Ziye Yang:
> --- a/lib/librte_eal/common/include/rte_pci.h
> +++ b/lib/librte_eal/common/include/rte_pci.h
> @@
2016-05-19 12:16, Jerin Jacob:
> On Wed, May 18, 2016 at 06:02:08PM +0200, Olivier Matz wrote:
> > Some architectures (ex: Power8) have a cache line size of 128 bytes,
> > so the drivers should not expect that prefetching the second part of
> > the mbuf with rte_prefetch0(>cacheline1) is valid.
>
bytes
Desc: dpdk_dma_error.PNG
URL:
<http://dpdk.org/ml/archives/dev/attachments/20160524/b7550454/attachment-0002.png>
-- next part --
A non-text attachment was scrubbed...
Name: dpdk_hp_bios.PNG
Type: image/png
Size: 60513 bytes
Desc: dpdk_hp_bios.PNG
URL:
<http://d
On Tue, May 24, 2016 at 10:41:28AM +0300, Panu Matilainen wrote:
> On 05/20/2016 08:24 PM, Neil Horman wrote:
> > This tool searches for the primer sting PMD_DRIVER_INFO= in any ELF binary,
> > and, if found parses the remainder of the string as a json encoded string,
> > outputting the results in
Please take care of the title.
Here you know it is not a modulo but you keep copying the coverity message.
It is fixed to "mempool: avoid division by zero"
2016-05-23 13:28, Olivier Matz:
> On 05/19/2016 02:36 PM, Slawomir Mrozowicz wrote:
> > Fix issue reported by Coverity.
> >
> > Coverity ID
Hi,
I have rte_pmd_init_all used in my code. While upgrading from DPDK 1.6 to
DPDK 1.7, I see this function is missing in DPDK 1.7. I
removed rte_pmd_init_all from my code, but I am not sure whether
initialization is done for all poll mode drivers
What is the right way to do this?. Can anyone
On Tue, May 24, 2016 at 09:15:41AM +0300, Panu Matilainen wrote:
> On 05/23/2016 04:55 PM, Neil Horman wrote:
> > 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
On Tue, May 24, 2016 at 01:11:26PM +0800, Yuanhan Liu wrote:
> On Mon, May 23, 2016 at 06:06:21PM +0100, Ferruh Yigit wrote:
> > 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
On 05/20/2016 08:24 PM, Neil Horman wrote:
> This tool searches for the primer sting PMD_DRIVER_INFO= in any ELF binary,
> and, if found parses the remainder of the string as a json encoded string,
> outputting the results in either a human readable or raw, script parseable
> format
>
> Note that,
On 05/20/2016 08:24 PM, Neil Horman wrote:
> Modify the PMD_REGISTER_DRIVER macro, adding a name argument to it. The
> addition of a name argument creates a token that can be used for subsequent
> macros in the creation of unique symbol names to export additional bits of
> information for use by
Calling i40e_switch_rx_queue without checking return value.
Fixed by add warning log information if return failed.
Fixes: 71d35259ff67 ("i40e: tear down flow director")
Coverity ID 13198
Signed-off-by: Daniel Mrzyglod
---
drivers/net/i40e/i40e_fdir.c | 5 -
1 file changed, 4 insertions(+),
On Tue, May 24, 2016 at 09:57:35AM +0300, Panu Matilainen wrote:
> On 05/20/2016 08:24 PM, Neil Horman wrote:
> > Modify the PMD_REGISTER_DRIVER macro, adding a name argument to it. The
> > addition of a name argument creates a token that can be used for subsequent
> > macros in the creation of
On 05/23/2016 04:55 PM, Neil Horman wrote:
> 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
> -Original Message-
> From: Stephen Hemminger [mailto:stephen at networkplumber.org]
> Sent: Monday, May 23, 2016 11:25 PM
> To: Pattan, Reshma
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v4 3/9] librte_ether: add new fields to
> rte_eth_dev_info struct
>
> On Mon, 23 May
The PMD?s are now being inited in the EAL init call for the code and that API
is not required anymore. If you are adding a new driver, make sure it is up to
date with any changes in the driver code too.
Regards,
Keith
>Hi,
>
>I have rte_pmd_init_all used in my code. While upgrading from DPDK
Hi,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Wenzhuo Lu
> Sent: Thursday, May 5, 2016 5:11 AM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH 0/4] automatic link recovery on ixgbe/igb VF
>
> Now if the PF link is down and up, VF doesn't handle
You can just check how it does in any of example applications (e.g. l2fwd).
Regards,
Helin
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of shiva m
> Sent: Tuesday, May 24, 2016 1:33 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] rte_pmd_init_all is missing
> -Original Message-
> From: Mrozowicz, SlawomirX
> Sent: Monday, May 23, 2016 8:25 PM
> To: Zhang, Helin ; Wu, Jingjing intel.com>
> Cc: dev at dpdk.org; Mrozowicz, SlawomirX
> Subject: [PATCH] i40e: Unchecked return value
>
> Calling i40e_switch_tx_queue without checking return
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
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 -
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
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 -
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 +
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
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 ++-
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 -
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
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
---
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
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
94 matches
Mail list logo
