Branch: refs/heads/2.0
Home: https://github.com/Linaro/odp
Commit: d37aa7f1c0480273fb0ccfded44a13e3e5a8e42f
https://github.com/Linaro/odp/commit/d37aa7f1c0480273fb0ccfded44a13e3e5a8e42f
Author: Bogdan Pricope
Date: 2017-10-24 (Tue, 24 Oct 2017)
From: Sachin Saxena
Pool info support has been extented to provide minimum and maximum addresses of
application
accessible data for any pool object (event).
Signed-off-by: Sachin Saxena
---
/** Email created from pull request 200
Signed-off-by: Sachin Saxena sachin.sax...@linaro.org
github
/** Email created from pull request 200 (sachin-saxena:api-next)
** https://github.com/Linaro/odp/pull/200
** Patch: https://github.com/Linaro/odp/pull/200.patch
** Base sha:
brbrooks replied on github web page:
include/odp/api/spec/pool.h
line 9
@@ -294,6 +294,22 @@ odp_pool_t odp_pool_lookup(const char *name);
typedef struct odp_pool_info_t {
const char *name; /**< pool name */
odp_pool_param_t params; /**< pool parameters */
+
+
brbrooks replied on github web page:
include/odp/api/spec/pool.h
line 17
@@ -294,6 +294,22 @@ odp_pool_t odp_pool_lookup(const char *name);
typedef struct odp_pool_info_t {
const char *name; /**< pool name */
odp_pool_param_t params; /**< pool parameters */
+
+
From: Bill Fischofer
Add the tests-linux-generic.env file to .gitignore
Signed-off-by: Bill Fischofer
---
/** Email created from pull request 254 (Bill-Fischofer-Linaro:test-gitignore)
** https://github.com/Linaro/odp/pull/254
** Patch:
Add two missing files to .gitignore to reflect restructuring changes to the
test directories.
github
/** Email created from pull request 254 (Bill-Fischofer-Linaro:test-gitignore)
** https://github.com/Linaro/odp/pull/254
** Patch:
From: Bill Fischofer
Add the tests-validation.env file to .gitignore.
Signed-off-by: Bill Fischofer
---
/** Email created from pull request 254 (Bill-Fischofer-Linaro:test-gitignore)
** https://github.com/Linaro/odp/pull/254
** Patch:
From: Bill Fischofer
Restore the ODP_ASSERTS() that assist application debugging that were
lost in the recent packet reference rework. When using --enable-debug
insert guards to detect duplicate packet frees as well as attempts to
modify known shared data sections of
Restore ODP_ASSERT() calls in packet processing lost in the recent restructure.
Also correct a syntax glitch in the previous merge that somehow got past
checkpatch.
github
/** Email created from pull request 180 (Bill-Fischofer-Linaro:pkt-debug)
**
On Mon, Oct 23, 2017 at 2:22 PM, Maxim Uvarov
wrote:
> On 10/23/17 18:23, Francois Ozog wrote:
> > Hi Maxim,
> >
> > Is this regular memory or DMA capable memory?
> >
> > This would mean application would know how, for instance, to handle
> things
> > when memory is
On 10/23/17 18:23, Francois Ozog wrote:
> Hi Maxim,
>
> Is this regular memory or DMA capable memory?
>
> This would mean application would know how, for instance, to handle things
> when memory is lower than a threshold or something similar; or it may
> decide to use a percentage of memory for
Summary of discussion and decision from today's ODP architecture call:
1. Status events (type ODP_EVENT_STATUS) are reserved for use in reporting
out-of-band notifications. The only such usage of this event currently
defined is soft byte/packet expiration notifications on outbound SAs
operating
Hi Maxim,
Is this regular memory or DMA capable memory?
This would mean application would know how, for instance, to handle things
when memory is lower than a threshold or something similar; or it may
decide to use a percentage of memory for packet pools; or many other
policies. I don't see
From: Maxim Uvarov
grep in doxygen log, not in some test.log
Signed-off-by: Maxim Uvarov
---
/** Email created from pull request 253 (muvarov:master_doxyfix)
** https://github.com/Linaro/odp/pull/253
** Patch:
grep in doxygen log, not in some test.log
Signed-off-by: Maxim Uvarov maxim.uva...@linaro.org
github
/** Email created from pull request 253 (muvarov:master_doxyfix)
** https://github.com/Linaro/odp/pull/253
** Patch:
Branch: refs/heads/api-next
Home: https://github.com/Linaro/odp
Commit: e733d6ba6382ccbf9d6c84a6a10bf720c2e16c0d
https://github.com/Linaro/odp/commit/e733d6ba6382ccbf9d6c84a6a10bf720c2e16c0d
Author: Petri Savolainen
Date: 2017-10-23 (Mon, 23 Oct
From: Matias Elo
In zero-copy dpdk pktio mode a matching dpdk mempool is required for each
odp packet pool from which packets are sent using dpdk pktio. This is
required for all packet pools, even for those which are not passed directly
to pktio devices as odp_pktio_open()
From: Matias Elo
Changes:
- Function rte_set_log_level() is now deprecated. Replace it with
rte_log_set_global_level().
- New dependency added to libnuma.
- Pcap-based virtual devices are renamed to net_pcap.
- DPDK pktio devices cannot be restarted after calling
Add missing DPDK mempool initialization zero-copy DKDK pktio and bumb target
DPDK version to 17.08.
github
/** Email created from pull request 252 (matiaselo:dev/bump_dpdk_version)
** https://github.com/Linaro/odp/pull/252
** Patch:
Applications should request the amount of storage they need (possibly as
configured) rather than trying to grab everything they can find "just in
case". Especially in an NFV environment that's not very neighborly behavior.
On Mon, Oct 23, 2017 at 2:44 AM, Dmitry Eremin-Solenikov <
From: Petri Savolainen
Use odp_pool_param_init() to initialize pool parameters.
Signed-off-by: Petri Savolainen
---
/** Email created from pull request 234 (psavol:next-pool-param)
** https://github.com/Linaro/odp/pull/234
** Patch:
From: Petri Savolainen
Remove anonymous union from pool parameter structure.
Union makes it impossible to initialize parameters per
pool type (use other values than all zeros). This change
is not visible to applications (union was anonymous).
Signed-off-by: Petri
From: Petri Savolainen
Packet pool parameters does not require application to specify
the maximum number of packet. Application is more portable, if
it does not restrict max_num, but instead uses this info field
after pool creation.
Signed-off-by: Petri Savolainen
From: Petri Savolainen
Implemented max_num info for packet pools.
Signed-off-by: Petri Savolainen
---
/** Email created from pull request 234 (psavol:next-pool-param)
** https://github.com/Linaro/odp/pull/234
** Patch:
From: Petri Savolainen
Test packet pool subparameters when those are supported.
Signed-off-by: Petri Savolainen
---
/** Email created from pull request 234 (psavol:next-pool-param)
** https://github.com/Linaro/odp/pull/234
** Patch:
From: Petri Savolainen
Additional packet length and number specification gives
more information to implementation about intended packet
length distribution in the pool. This enables e.g. correct
initialization when pool implementation is based on multiple
fixed
From: Petri Savolainen
Added packet pool parameter 'max_num', so that 'num' parameter
can be round up by the implementation. Most implementations have
a fixed segment size, and need this flexibility. For example,
when 'len' is larger than the pool segment size, there
Added packet pool parameters for more fine grained pool
configuration. The basic usage of the parameters is not changed,
except that implementation may now round up 'num' by default.
Application can limit the round up with new 'max_num' parameter.
If application does not limit max_num (but is
> -Original Message-
> From: lng-odp [mailto:lng-odp-boun...@lists.linaro.org] On Behalf Of
> Github ODP bot
> Sent: Friday, October 20, 2017 5:00 PM
> To: lng-odp@lists.linaro.org
> Subject: [lng-odp] [PATCH API-NEXT v1 1/1] api: pktio: clarify
> odp_pktio_config() restrictions
>
>
1). I am not a huge fan of the idea of starting a timer to perform the SA
delete since calculating this timer's timeout will be an issue and will depend
upon the network load at the specific time.
[PLE] I fully agree with you on that. Time-out counters are an ugly solution.
But sometimes
With static_inline.h gone, there is no need to
AC_SUBST([ODP_ABI_COMPAT]). Drop it from configure.ac.
Signed-off-by: Dmitry Eremin-Solenikov
---
configure.ac | 1 -
1 file changed, 1 deletion(-)
diff --git a/configure.ac b/configure.ac
index
There is no point in installing inline headers in non-ABI-compat mode
anymore. They are not included by any other header.
Signed-off-by: Dmitry Eremin-Solenikov
---
platform/linux-generic/Makefile.am | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
Replace static_inline.h header with unconditional defines of _ODP_INLINE
macro (either to 'static inline' or to empty value) depending on the
compilation place.
Signed-off-by: Dmitry Eremin-Solenikov
---
Makefile.inc | 1
Several files consist only (or mostly) of functions compiled only if
ODP is compiled in ABI compatibility mode. Instead of having an ifdef
inside, guard them with if ODP_ABI_COMPAT condition in Makefile.am
Signed-off-by: Dmitry Eremin-Solenikov
---
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am| 6 ++
.../include => include}/odp/api/classification.h | 2 +-
.../include => include}/odp/api/crypto.h | 2 +-
.../include =>
Move more generic headers from platform include dir to common include
dir.
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am | 10 ++
{platform/linux-generic/include => include}/odp/api/chksum.h |
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am | 1 +
.../include => include}/odp/api/ticketlock.h| 11 +++
platform/linux-generic/Makefile.am | 7 +++
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am| 1 +
include/odp/api/sync.h | 28 ++
platform/linux-generic/Makefile.am | 2 +-
.../odp/api =>
Signed-off-by: Dmitry Eremin-Solenikov
---
example/traffic_mgmt/odp_traffic_mgmt.c| 1 +
helper/test/linux/process.c| 2 ++
helper/test/linux/pthread.c| 2 ++
helper/threads.c
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am| 1 +
include/odp/api/byteorder.h| 28 ++
platform/linux-generic/Makefile.am | 3 +-
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am| 1 +
include/odp/api/atomic.h | 28 +++
platform/linux-generic/Makefile.am | 3 +-
.../odp/api/abi/atomic.h}
Rewrite atomic_types.h/atomic_inlines.h to clearly separate simple
(common) and locked 64-bit cases. This is allows us to ease switching of
atomic header to abi setup.
Signed-off-by: Dmitry Eremin-Solenikov
---
.../include/odp/api/plat/atomic_inlines.h
Start simplifying platform headers by using same include paths for
ABI-compat and non-ABI-compat modes. In either case there will be
header (provided either by generic includes, or
by platform itself). In case of non-ABI-compat mode, this header
provides runtime-optimized versions of some
Rework more platform headers to use odp/api/abi/ subdir instead of
odp/api/plat/.
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am| 8 +
.../include => include}/odp/api/barrier.h | 6 ++--
Change defines guarding inclusion of ODP API spec files from ODP_API_FOO
to ODP_API_SPEC_FOO, as they are placed in odp/api/spec/foo.h path.
Signed-off-by: Dmitry Eremin-Solenikov
---
include/odp/api/spec/align.h | 4 ++--
Move the rest of ODP_SCHED_GROUP_* defines to schedule_types.h to remove
duplication between headers.
Signed-off-by: Dmitry Eremin-Solenikov
---
include/odp/api/abi-default/schedule.h | 8
include/odp/api/abi-default/schedule_types.h | 8
Reduce amount of directory levels by moving default ABI headers to
odp/api/abi-default.
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am| 82 +++---
.../default/api/abi => api/abi-default}/atomic.h
Provide a set of platform-neutral ABI files. They are mostly modified
versions of linux-generic api files.
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am| 188 -
Signed-off-by: Dmitry Eremin-Solenikov
---
include/odp/api/spec/classification.h | 11 ---
platform/linux-generic/include/odp/api/classification.h | 13 -
2 files changed, 24 deletions(-)
diff --git
Signed-off-by: Dmitry Eremin-Solenikov
---
include/odp/api/spec/queue.h | 5 -
platform/linux-generic/include/odp/api/queue.h | 12
2 files changed, 17 deletions(-)
diff --git a/include/odp/api/spec/queue.h
There is no need to install ABI headers (or to have them in include
path) in non-ABI-compat mode, they should not be used at all. Still
provide default ABI headers, because platform may depend on them.
Signed-off-by: Dmitry Eremin-Solenikov
---
Makefile.inc
Install ABI headers directly to odp/api/abi, removing the need for extra
symlink.
Signed-off-by: Dmitry Eremin-Solenikov
---
include/Makefile.am | 12 +---
1 file changed, 1 insertion(+), 11 deletions(-)
diff --git a/include/Makefile.am
Signed-off-by: Dmitry Eremin-Solenikov
---
Makefile.inc | 12
example/Makefile.inc | 12
helper/Makefile.am | 12
Signed-off-by: Dmitry Eremin-Solenikov
---
configure.ac | 1 +
1 file changed, 1 insertion(+)
diff --git a/configure.ac b/configure.ac
index 4bc77500c279..255de960c180 100644
--- a/configure.ac
+++ b/configure.ac
@@ -271,6 +271,7 @@ AC_ARG_ENABLE([abi-compat],
This is a copy of https://github.com/Linaro/odp/pull/250 (currently at
v5), since mailer scripts won't send huge patch series.
This patch serie is based on discussions during SFO17 with Petri and
other colleagues.
My main goals were:
- providing full set of ABI headers to be used by other
__odp_force is not part of ODP API, so it should not be used directly.
Signed-off-by: Dmitry Eremin-Solenikov
---
platform/linux-generic/odp_ipsec.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/platform/linux-generic/odp_ipsec.c
__odp_force is not part of ODP API, so it should not be used directly.
Signed-off-by: Dmitry Eremin-Solenikov
---
helper/include/odp/helper/chksum.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/helper/include/odp/helper/chksum.h
Signed-off-by: Dmitry Eremin-Solenikov
---
.travis.yml | 4
1 file changed, 4 insertions(+)
diff --git a/.travis.yml b/.travis.yml
index fdeeefd113ea..6196f91d41de 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -74,6 +74,10 @@ env:
-
Signed-off-by: Dmitry Eremin-Solenikov
---
.travis.yml | 8
1 file changed, 8 insertions(+)
diff --git a/.travis.yml b/.travis.yml
index 49b922e3cdba..fdeeefd113ea 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -244,6 +244,14 @@ jobs:
sachin-saxena replied on github web page:
include/odp/api/spec/pool.h
@@ -294,6 +294,19 @@ odp_pool_t odp_pool_lookup(const char *name);
typedef struct odp_pool_info_t {
const char *name; /**< pool name */
odp_pool_param_t params; /**< pool parameters */
+ /**
Signed-off-by: Sachin Saxena sachin.sax...@linaro.org
github
/** Email created from pull request 200 (sachin-saxena:api-next)
** https://github.com/Linaro/odp/pull/200
** Patch: https://github.com/Linaro/odp/pull/200.patch
** Base sha:
From: Sachin Saxena
Signed-off-by: Sachin Saxena
---
/** Email created from pull request 200 (sachin-saxena:api-next)
** https://github.com/Linaro/odp/pull/200
** Patch: https://github.com/Linaro/odp/pull/200.patch
** Base sha:
From: Petri Savolainen
Clean up shm header file. Remove [in] Doxygen tags as function
parameters are by default input. Move @note text to the main
text block, as it's usual API specification text.
Signed-off-by: Petri Savolainen
---
/**
From: Petri Savolainen
Call ones complement functions. Actual validation of
correctness of the sum is to be done, when an implementation
exist which sets the sum on packet input.
Signed-off-by: Petri Savolainen
---
/** Email created
From: Petri Savolainen
Added packet metadata for ones complement sum over IP
payload in a packet. Some NICs calculate the sum
during packet input (at least for IP fragments) and
store the value into the packet descriptor. This offloads
L4 checksum calculation for IP
From: Petri Savolainen
Added dummy implementation of ones complement calls. Linux
generic implementation does not calculate the sum for all
incoming packets as it would be wasteful in SW. It's better
to wait until application asks it with odp_chksum_ones_comp16().
Added packet metadata for ones complement sum over IP
payload in a packet. Some NICs calculate the sum
during packet input (at least for IP fragments) and
store the value into the packet descriptor. This offloads
L4 checksum calculation for IP fragments as SW does not
need sum all payload data,
From: Petri Savolainen
Implement the new block level debug print function.
Signed-off-by: Petri Savolainen
---
/** Email created from pull request 246 (psavol:next-shm-print)
** https://github.com/Linaro/odp/pull/246
** Patch:
From: Petri Savolainen
Use ODP_PRINT instead of ODP_DBG in the print function.
The idea of a print function is that it prints information
when application calls it. ODP_DBG does not print anything
by default.
Format printted output to be a readable table of shm
From: Petri Savolainen
Added function to print debugging information about the
specified shared memory block. Info is implementation defined
but should include information about allocated memory addresses,
pages, etc.
Signed-off-by: Petri Savolainen
Add SHM block level debug print function. This is similar to e.g.
odp_pool_print(). All SHM blocks print function (odp_shm_print_all()) cannot
print as much details on every block. Cleaned also SHM header file and
odp_shm_print_all() implementation while touching the files.
From: Bogdan Pricope
Make modular framework/subsystems symbols visible
Signed-off-by: Bogdan Pricope
---
/** Email created from pull request 251
(bogdanPricope:2_0_symbols_visibility_pr)
** https://github.com/Linaro/odp/pull/251
**
Make modular framework/subsystems symbols visible
Signed-off-by: Bogdan Pricope bogdan.pric...@linaro.org
github
/** Email created from pull request 251
(bogdanPricope:2_0_symbols_visibility_pr)
** https://github.com/Linaro/odp/pull/251
** Patch:
Hi,
I was on vacation and went through this mail thread just now.
I wasn't sure where to inject my comments hence I am providing them at
the top for clarity.
1). I am not a huge fan of the idea of starting a timer to perform the
SA delete since calculating this timer's timeout will be an issue
On 23/10/17 10:39, Maxim Uvarov wrote:
> It might be reasonable to add also api call to get return free memory. So
> that application can adjust pools /buffers size according to hardware or VM
> settings. Which might be good fit for NFV set up.
> Any opinions on that?
It would depend on the
It might be reasonable to add also api call to get return free memory. So
that application can adjust pools /buffers size according to hardware or VM
settings. Which might be good fit for NFV set up.
Any opinions on that?
Maxim.
77 matches
Mail list logo