Branch: refs/heads/api-next
Home: https://github.com/Linaro/odp
Commit: 73bc46197ec0617878034ac793dcc96ed41eb9a1
https://github.com/Linaro/odp/commit/73bc46197ec0617878034ac793dcc96ed41eb9a1
Author: Maxim Uvarov
Date: 2017-10-06 (Fri, 06 Oct 2017)
Changed paths:
M .trav
Branch: refs/heads/master
Home: https://github.com/Linaro/odp
Commit: 73bc46197ec0617878034ac793dcc96ed41eb9a1
https://github.com/Linaro/odp/commit/73bc46197ec0617878034ac793dcc96ed41eb9a1
Author: Maxim Uvarov
Date: 2017-10-06 (Fri, 06 Oct 2017)
Changed paths:
M .travis
On 6 October 2017 at 22:26, Bill Fischofer
wrote:
> PR #225 [1] is a good example of the sort of packaging/distro problem ODP
> wants to solve. DPDK currently requires special compile flags to determine
> the microarchitecture it's running on. As a result, Travis broke when they
> deployed new te
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
platform/linux-generic/odp_pool.c
line 6
@@ -32,6 +32,10 @@
#define CACHE_BURST32
#define RING_SIZE_MIN (2 * CACHE_BURST)
+/* Make sure packet buffers don't cross page boundaries starting from this page
+ * size. */
+#defi
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
platform/linux-generic/odp_pool.c
line 80
@@ -403,6 +430,12 @@ static odp_pool_t pool_create(const char *name,
odp_pool_param_t *params,
block_size = ROUNDUP_CACHE_LINE(hdr_size + align + headroom +
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
platform/linux-generic/odp_pool.c
line 40
@@ -232,19 +237,40 @@ static void init_buffers(pool_t *pool)
uint32_t mask;
int type;
uint32_t seg_size;
+ uint64_t page_size = 0;
+ int skipped_blocks =
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
platform/linux-generic/odp_pool.c
line 40
@@ -232,19 +237,40 @@ static void init_buffers(pool_t *pool)
uint32_t mask;
int type;
uint32_t seg_size;
+ uint64_t page_size = 0;
+ int skipped_blocks =
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
platform/linux-generic/odp_pool.c
line 6
@@ -32,6 +32,10 @@
#define CACHE_BURST32
#define RING_SIZE_MIN (2 * CACHE_BURST)
+/* Make sure packet buffers don't cross page boundaries starting from this page
+ * size. */
+#defi
PR #225 [1] is a good example of the sort of packaging/distro problem ODP
wants to solve. DPDK currently requires special compile flags to determine
the microarchitecture it's running on. As a result, Travis broke when they
deployed new test machines that had a different microarchitecture. To get
a
From: Dmitry Eremin-Solenikov
- Drop unused $(srcdir)
- Drop unused $(EXEEXT)
- Drop unused dist_ prefixes
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull
From: Dmitry Eremin-Solenikov
- Drop unused $(srcdir)
- Drop unused $(EXEEXT)
- Drop unused dist_ prefixes
- Merge noinst_HEADERS into proper _SOURCES
- Merge -static into AM_LDFLAGS
- Drop extra _CFLAGS adding -I which is already present on include path
Signed-off-by: Dmitry Eremin-Solenikov
-
From: Dmitry Eremin-Solenikov
Referencing header in _HEADERS variable means it will be included into
distribution. Move static_inline.h header to nodist_*_HEADERS to stop
including it into tarball.
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-de
From: Dmitry Eremin-Solenikov
There is no point in using $(srcdir) when referencing files inside
source directory. Just drop that to simplify Makefile.am.
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/2
From: Dmitry Eremin-Solenikov
Due to the way Automake interacts with autoconf substitutions, ODP
tarballs included ABI headers only for the current directory. Include
all ABI headers into tarball.
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
From: Dmitry Eremin-Solenikov
There is little point in handling common headers in platform-specific
dir. Move respective definitions to include/Makefie.am.
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
From: Dmitry Eremin-Solenikov
All tests in test/ are now common to all platforms, so no point in
specifying that via (the only) common_plat subdirectory inside test dir.
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Li
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
From: Dmitry Eremin-Solenikov
Stop calling common_plat validation tests from linux-generic tests
subdir. Instead move their handling to proper Makefile.am.
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
From: Dmitry Eremin-Solenikov
While we are at it, run 128bit atomic check only if __int128 is
available for the target platform.
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://gith
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
Update ODP buildsystem partially basing on ideas discussed during Linaro
Connect.
Simplify platform configuration scripts. Rework some of the checks as Autoconf
macros
Merge tests back to platform directory, to fully separate generic and
platform-specific code
Rework handling of includes, move
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
From: Dmitry Eremin-Solenikov
Signed-off-by: Dmitry Eremin-Solenikov
---
/** Email created from pull request 213 (lumag:plat-def)
** https://github.com/Linaro/odp/pull/213
** Patch: https://github.com/Linaro/odp/pull/213.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge co
From: Maxim Uvarov
as described here build dpdk for Sandybridge
https://stackoverflow.com/questions/37386738/
error-when-running-dpdk-app-on-valgrind
to make it work in Travis
Signed-off-by: Maxim Uvarov
---
/** Email created from pull request 225 (muvarov:master_dpdkgeneric)
** https://github
v2: define RTE_MACHINE in install stage, not on configure. There is really
works.
disable cpu optimisations for code run in Travis env.
By default dpdk uses -march=corei7-avx with generates
specific instructions and execution fails with:
Please check that RTE_MACHINE is set correctly.
More info
From: Maxim Uvarov
some issues happened with dpdk, needed to be more
clear what we build. Turn off dpdk cache for now.
Signed-off-by: Maxim Uvarov
Reviewed-by: Bill Fischofer
---
/** Email created from pull request 225 (muvarov:master_dpdkgeneric)
** https://github.com/Linaro/odp/pull/225
**
Hi Petri,
DPDK modules are not allowing DPDK to adapt to underlying architecture. It
is just pugable HW.
The DDF will deal with that.
The problem we need to solve is architecture adaptation. where the CORE of
the application changes. there is a single DPDK core with drivers. ODP has
different CO
We'll only got for rpm? deb packaging?
Regards
Ilias
On 6 October 2017 at 20:18, Maxim Uvarov wrote:
> http://pkgs.fedoraproject.org/cgit/rpms/dpdk.git/tree/dpdk.spec
>
> On 6 October 2017 at 20:17, Maxim Uvarov wrote:
>
> >
> >
> > On 6 October 2017 at 20:05, Honnappa Nagarahalli <
> > honnap
http://pkgs.fedoraproject.org/cgit/rpms/dpdk.git/tree/dpdk.spec
On 6 October 2017 at 20:17, Maxim Uvarov wrote:
>
>
> On 6 October 2017 at 20:05, Honnappa Nagarahalli <
> honnappa.nagaraha...@linaro.org> wrote:
>
>> Any experts on how is the packaging done for DPDK?
>>
>>
> ./pkg/dpdk.spec ?
>
On 6 October 2017 at 20:05, Honnappa Nagarahalli <
honnappa.nagaraha...@linaro.org> wrote:
> Any experts on how is the packaging done for DPDK?
>
>
./pkg/dpdk.spec ?
> On 6 October 2017 at 08:36, Savolainen, Petri (Nokia - FI/Espoo)
> wrote:
> >> > No, I'm pointing that the more there's common
Matias Elo(matiaselo) replied on github web page:
platform/linux-generic/odp_pool.c
line 40
@@ -232,19 +237,40 @@ static void init_buffers(pool_t *pool)
uint32_t mask;
int type;
uint32_t seg_size;
+ uint64_t page_size = 0;
+ int skipped_blocks = 0;
+
+ if
Matias Elo(matiaselo) replied on github web page:
platform/linux-generic/odp_pool.c
line 6
@@ -32,6 +32,10 @@
#define CACHE_BURST32
#define RING_SIZE_MIN (2 * CACHE_BURST)
+/* Make sure packet buffers don't cross page boundaries starting from this page
+ * size. */
+#define CONT_PAGE_SIZ
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
platform/linux-generic/odp_pool.c
line 80
@@ -403,6 +430,12 @@ static odp_pool_t pool_create(const char *name,
odp_pool_param_t *params,
block_size = ROUNDUP_CACHE_LINE(hdr_size + align + headroom +
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
platform/linux-generic/odp_pool.c
line 6
@@ -32,6 +32,10 @@
#define CACHE_BURST32
#define RING_SIZE_MIN (2 * CACHE_BURST)
+/* Make sure packet buffers don't cross page boundaries starting from this page
+ * size. */
+#defi
Any experts on how is the packaging done for DPDK?
On 6 October 2017 at 08:36, Savolainen, Petri (Nokia - FI/Espoo)
wrote:
>> > No, I'm pointing that the more there's common core SW, the more there
>> > are trade-offs and the less direct HW access == less performance. For
>> > optimal performanc
Matias Elo(matiaselo) replied on github web page:
platform/linux-generic/odp_pool.c
line 80
@@ -403,6 +430,12 @@ static odp_pool_t pool_create(const char *name,
odp_pool_param_t *params,
block_size = ROUNDUP_CACHE_LINE(hdr_size + align + headroom +
From: Ilias Apalodimas
Signed-off-by: Ilias Apalodimas
---
/** Email created from pull request 214 (apalos:bug_3249)
** https://github.com/Linaro/odp/pull/214
** Patch: https://github.com/Linaro/odp/pull/214.patch
** Base sha: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
** Merge commit sha: 57f
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
platform/linux-generic/odp_pool.c
line 40
@@ -232,19 +237,40 @@ static void init_buffers(pool_t *pool)
uint32_t mask;
int type;
uint32_t seg_size;
+ uint64_t page_size = 0;
+ int skipped_blocks =
v2: define RTE_MACHINE in install stage, not on configure. There is really
works.
disable cpu optimisations for code run in Travis env.
By default dpdk uses -march=corei7-avx with generates
specific instructions and execution fails with:
Please check that RTE_MACHINE is set correctly.
More info
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
.travis.yml
@@ -156,7 +156,7 @@ install:
git -c advice.detachedHead=false clone -q --depth=1
--single-branch --branch=v17.02 http://dpdk.org/git/dpdk dpdk
pushd dpdk
git log --oneline --decora
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
.travis.yml
line 5
@@ -57,7 +57,7 @@ cache:
ccache: true
pip: true
directories:
-- dpdk
+#- dpdk
Comment:
I assume we'll turn this back on before merging if this fix works
From: Ilias Apalodimas
PR#171 broke some of the functionality we had on getting cpu speed.
This patch reverts parts of that commit adding a fallback when
/sys/devices/system/cpu/cpuX/cpufreq/ is not available.
Solves https://bugs.linaro.org/show_bug.cgi?id=3249
Signed-off-by: Ilias Apalodimas
From: Maxim Uvarov
some issues happened with dpdk, needed to be more
clear what we build. Turn off dpdk cache for now.
Signed-off-by: Maxim Uvarov
---
/** Email created from pull request 225 (muvarov:master_dpdkgeneric)
** https://github.com/Linaro/odp/pull/225
** Patch: https://github.com/Li
From: Maxim Uvarov
disable cpu optimisations for code run in Travis env.
By default dpdk uses -march=corei7-avx with generates
specific instructions and execution fails with:
Please check that RTE_MACHINE is set correctly.
That is caused by buggy dpdk makefile:
MACHINE_CFLAGS = -march=native
SS
PR#171 broke some of the functionality we had on getting cpu speed.
This patch reverts parts of that commit adding a fallback when
/sys/devices/system/cpu/cpuX/cpufreq/ is not available.
Solves https://bugs.linaro.org/show_bug.cgi?id=3249
Signed-off-by: Ilias Apalodimas ilias.apalodi...@linaro.org
Bill Fischofer(Bill-Fischofer-Linaro) replied on github web page:
.travis.yml
line 5
@@ -57,7 +57,7 @@ cache:
ccache: true
pip: true
directories:
-- dpdk
+#- dpdk
Comment:
I assume we'll turn this back on before merging if this fix works
From: Matias Elo
Increase maximum number of workers to 128. Also, fixes a buffer overflow
which happened when system's maximum thread count was larger than the
value of MAX_WORKERS.
Signed-off-by: Matias Elo
---
/** Email created from pull request 226 (matiaselo:fix/pktio_perf)
** https://git
Increase maximum number of workers to 128. Also, fixes a buffer overflow
which happened when system's maximum thread count was larger than the
value of MAX_WORKERS.
Signed-off-by: Matias Elo matias@nokia.com
github
/** Email created from pull request 22
From: Maxim Uvarov
disable cpu optimisations for code run in Travis env.
By default dpdk uses -march=corei7-avx with generates
specific instructions and execution fails with:
Please check that RTE_MACHINE is set correctly.
More info here:
https://github.com/Linaro/odp/pull/218
Signed-off-by: Max
disable cpu optimisations for code run in Travis env.
By default dpdk uses -march=corei7-avx with generates
specific instructions and execution fails with:
Please check that RTE_MACHINE is set correctly.
More info here:
#218
disabling dpdk caching in Travis for now.
github---
From: Maxim Uvarov
some issues happend with dpdk, needed to be more
clear what we build. Turn off dpdk cache for now.
Signed-off-by: Maxim Uvarov
---
/** Email created from pull request 225 (muvarov:master_dpdkgeneric)
** https://github.com/Linaro/odp/pull/225
** Patch: https://github.com/Lin
> > No, I'm pointing that the more there's common core SW, the more there
> > are trade-offs and the less direct HW access == less performance. For
> > optimal performance, the amount of common core SW is zero.
>
> Yes this is sort of the ideal but I doubt this type of installation
> will be accep
From: Petri Savolainen
Added test for a 1517 bytes long UDP packet. The checksum is
checked using both a single call as well as multiple consecutive
calls (to emulate checksumming over fragments).
Signed-off-by: Petri Savolainen
---
/** Email created from pull request 224 (psavol:next-ip-checks
From: Petri Savolainen
Initialize left_over variable, so that the pad byte
is initialized to zero.
Signed-off-by: Petri Savolainen
---
/** Email created from pull request 224 (psavol:next-ip-checksum)
** https://github.com/Linaro/odp/pull/224
** Patch: https://github.com/Linaro/odp/pull/224.p
Added test for a 1517 bytes long UDP packet.
The checkpatch error can be ignored. It erroneously complains about function
declaration, when it's an array declaration with alignment attribute and data
for automatic initialization.
ERROR: open brace '{' following function declarations go on the nex
On 5 October 2017 at 12:09, Savolainen, Petri (Nokia - FI/Espoo)
wrote:
> No HTML mails, please.
>
>
> From: Bill Fischofer [mailto:bill.fischo...@linaro.org]
> Sent: Wednesday, October 04, 2017 3:55 PM
> To: Savolainen, Petri (Nokia - FI/Espoo)
> Cc: Andriy Berestovskyy ; Ola
> Liljedahl ; lng-
Added test for a 1517 bytes long UDP packet.
The checkpatch error can be ignored. It erroneously complains about function
declaration, when it's an array declaration with alignment attribute and data
for automatic initialization.
ERROR: open brace '{' following function declarations go on the nex
From: Petri Savolainen
Added test for a 1517 bytes long UDP packet. The checksum is
checked using both a single call as well as multiple consecutive
calls (to emulate checksumming over fragments).
Signed-off-by: Petri Savolainen
---
/** Email created from pull request 224 (psavol:next-ip-checks
When packet pool memory page size is >= CONT_PAGE_SIZE, leave those memory
buffers unused which cross a page boundary. This is required by zero-copy
dpdk pktio (NIC drivers don't support buffers which cross page boundaries).
Signed-off-by: Matias Elo matias@nokia.com
github
From: Matias Elo
When packet pool memory page size is >= CONT_PAGE_SIZE, leave those memory
buffers unused which cross a page boundary. This is required by zero-copy
dpdk pktio (NIC drivers don't support buffers which cross page boundaries).
Signed-off-by: Matias Elo
---
/** Email created from
Branch: refs/heads/api-next
Home: https://github.com/Linaro/odp
Commit: 34884aa91524c4329e4ea1a9b312538d8f7dd187
https://github.com/Linaro/odp/commit/34884aa91524c4329e4ea1a9b312538d8f7dd187
Author: Mykyta Iziumtsev
Date: 2017-10-06 (Fri, 06 Oct 2017)
Changed paths:
M p
Branch: refs/heads/master
Home: https://github.com/Linaro/odp
Commit: a63f25ff2994b2df78c24f1f8b63d0e06628eb68
https://github.com/Linaro/odp/commit/a63f25ff2994b2df78c24f1f8b63d0e06628eb68
Author: Dmitry Eremin-Solenikov
Date: 2017-10-06 (Fri, 06 Oct 2017)
Changed paths:
From: Maxim Uvarov
changing dpdk version and not clearing build can lead
to reference to old cached not supported dpdk version.
Needed to remove cache if version was changed.
Signed-off-by: Maxim Uvarov
---
/** Email created from pull request 222 (muvarov:master_cleandpdkcache)
** https://gith
changing dpdk version and not clearing build can lead
to reference to old cached not supported dpdk version.
Needed to remove cache if version was changed.
Signed-off-by: Maxim Uvarov maxim.uva...@linaro.org
github
/** Email created from pull request 222 (mu
70 matches
Mail list logo