Re: [lng-odp] Crypto_operation_multi API?

On 10 May 2016 at 21:31, Nikhil Agarwal wrote: > One use case that will benefit from it will be fat ipsec tunnel use case, > where majority of packets coming from one pktio interface belongs to same > sessions. In general, it can be used in case where we are doing

Re: [lng-odp] ODP Addressing Model

On 05/09 16:54:26, Bill Fischofer wrote: > The purpose of this thread is to summarize the discussions we've had over > the past week on this topic and to foster discussion. The goal here is to > reach consensus on a complete specification for ODP in this area that can > be added to the User Guide

Re: [lng-odp] ODP Addressing Model

On 10 May 2016 at 20:33, Bill Fischofer wrote: > > > On Tue, May 10, 2016 at 3:17 AM, Savolainen, Petri (Nokia - FI/Espoo) < > petri.savolai...@nokia.com> wrote: > >> >> >> >> >> *From:* lng-odp [mailto:lng-odp-boun...@lists.linaro.org] *On Behalf Of *Bill >> Fischofer

[lng-odp] [PATCH] CHANGELOG: Add details for docs and arch changes

Signed-off-by: Mike Holmes --- CHANGELOG | 12 1 file changed, 12 insertions(+) diff --git a/CHANGELOG b/CHANGELOG index d9c66ed..b9d12f5 100644 --- a/CHANGELOG +++ b/CHANGELOG @@ -1,3 +1,15 @@ +== OpenDataPlane (1.10.1.0) + +=== New Features +

Re: [lng-odp] [PATCH 1/8] doc: api guide: remove incorrect include

On 9 May 2016 at 17:05, Bill Fischofer wrote: > For this series: Reviewed-and-tested-by: Bill Fischofer < > bill.fischo...@linaro.org> > merged > > On Mon, May 9, 2016 at 3:43 PM, Mike Holmes > wrote: > >> The api guide should not reference

[lng-odp] [PATCH 2/2] linux-generic: test: move pktio_env to common dir

pktio_env is used more widely than by just pktio testing. Move it to a more central location removing the need for a special case for performance tests where the pktio dir has to be traversed during build even when the directory is not used. Signed-off-by: Mike Holmes ---

[lng-odp] [PATCH 1/2] linux-generic: test: improve locating pktio_env

Rather than limit pktio_env to the current directory, find it on the path just as is done with pktio_main Signed-off-by: Mike Holmes --- platform/linux-generic/test/pktio/pktio_run | 9 ++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git

Re: [lng-odp] ODP Addressing Model

On Tue, May 10, 2016 at 3:17 AM, Savolainen, Petri (Nokia - FI/Espoo) < petri.savolai...@nokia.com> wrote: > > > > > *From:* lng-odp [mailto:lng-odp-boun...@lists.linaro.org] *On Behalf Of *Bill > Fischofer > *Sent:* Tuesday, May 10, 2016 12:54 AM > *To:* LNG ODP Mailman List

Re: [lng-odp] [API-NEXT, RFC, 1/1] api: cpu: performance profiling start/stop

We didn't get around to discussing this during today's public call, but we'll try to cover this during tomorrow's ARCH call. As I noted earlier, the question of thread assignment to cores becomes complicated in virtual environment when it's not clear that a (virtual) core necessarily implies that

Re: [lng-odp] [PATCHv2 1/2] doc: userguide: add timer/timeout state diagram

I looked at them and I liked the added color so reworked my diagram adding that and posted a v3 for it. I'm not sure trying to separate into two diagrams really helps but perhaps this reworked diagram is easier to follow? On Tue, May 10, 2016 at 3:45 AM, Christophe Milard <

[lng-odp] [PATCHv3 2/2] doc: userguide: add timer and timeout event section to user guide

Signed-off-by: Ivan Khoronzhuk Signed-off-by: Bill Fischofer --- doc/users-guide/Makefile.am| 1 + doc/users-guide/users-guide-timer.adoc | 97 ++ doc/users-guide/users-guide.adoc | 2 + 3

[lng-odp] [PATCHv3 1/2] doc: userguide: add timer/timeout state diagram

Signed-off-by: Bill Fischofer --- doc/images/.gitignore | 1 + doc/images/timer_fsm.gv | 55 + doc/users-guide/Makefile.am | 1 + 3 files changed, 57 insertions(+) create mode 100644 doc/images/timer_fsm.gv diff

Re: [lng-odp] [API-NEXT PATCH] api: system: specify default huge page size

As we discussed today, I believe doc clarifications like this should be fine for RC3. Petri to confirm. On Tue, May 10, 2016 at 10:34 AM, Maxim Uvarov wrote: > System can have different sizes for huge pages. Clarify that > this api returns default system huge page size,

Re: [lng-odp] Crypto_operation_multi API?

I think there's general agreement that the crypto APIs should be improved for Tiger Moth in a few areas such as adding support for additional capabilities (e.g., Barry's PKI proposal) as well as flow routing for protocol offload. We just need specific proposals to discuss. It's not too soon to

Re: [lng-odp] [PATCH] linux-generic: timer: change assert to account for padding

On Tue, May 10, 2016 at 11:19 AM, Maxim Uvarov wrote: > Ola, can you please review this patch? > > Bill, Bala, > > I am not sure that this change is correct. > > There is 2 things: > 1. Align on 16 bytes: > The entire struct will be aligned on a 16 byte boundary if

Re: [lng-odp] [PATCHv2] linux-generic: queue: avoid leakage on queue create

Merged, Maxim. On 05/05/16 16:19, Bill Fischofer wrote: I believe Maxim reviewed it, however perhaps not in this thread. On Thu, May 5, 2016 at 8:02 AM, Mike Holmes > wrote: Does this need a review ? On 16 April 2016 at 14:21,

[lng-odp] [Bug 2186] odp_queue_create() fails to clean up if schedule_queue_init fails

https://bugs.linaro.org/show_bug.cgi?id=2186 Maxim Uvarov changed: What|Removed |Added CC|

[lng-odp] run checkpatch automatically

Hi, As Mike raised my attention, I'm sloppy in that, so I decided to integrate it with my git wrapper. It is a shell script in my home directory which precedes git so I can enforce stuff not configurable for git. git hooks can do similar stuff, but they have to be created for each repository,

Re: [lng-odp] ODP Addressing Model

On 9 May 2016 23:54, "Bill Fischofer" wrote: > > The purpose of this thread is to summarize the discussions we've had over the past week on this topic and to foster discussion. The goal here is to reach consensus on a complete specification for ODP in this area that can

Re: [lng-odp] Crypto_operation_multi API?

One use case that will benefit from it will be fat ipsec tunnel use case, where majority of packets coming from one pktio interface belongs to same sessions. In general, it can be used in case where we are doing batch processing using burst rx/tx apis(odp_pktio_recv/tx with more than one packet)

[lng-odp] [API-NEXT PATCH] api: system: specify default huge page size

System can have different sizes for huge pages. Clarify that this api returns default system huge page size, which is usually 2Mb on x86 reachable with: grep Hugepagesize: /proc/meminfo Signed-off-by: Maxim Uvarov --- It might be too late for current Monarch rename api

Re: [lng-odp] Crypto_operation_multi API?

Do you have any specific changes or additions to crypto API? or any specific use-case which might benefit from the additional API? Regards, Bala On 10 May 2016 at 20:50, Nikhil Agarwal wrote: > Hi All, > > Since we have multi flavor for packet Rx/Tx APIs on queues

[lng-odp] Crypto_operation_multi API?

Hi All, Since we have multi flavor for packet Rx/Tx APIs on queues and scheduler, shall we have have a similar API for crypto operation? It might save significant CPU cycles for implementation having hardware engines. Any thoughts? Regards Nikhil ___

[lng-odp] [PATCHv2 1/2] linux-generic: use default huge page size

odp_shm_reserve() relays on huge page size to round up requested size. If 1 Gb pages present than parser takes it first as first alphabetical file name in sysfs. That lead to issue where all allocations wants 1 GB HP. This patch takes system default huge pages which are usually 2Mb, and has to be

[lng-odp] [PATCHv2 2/2] linux-generic: drop odp_ prefix for internal cpuinfo

A little bit code clean up to drop odp_ prefix from internal things and rename huge_pages to default_huge_pages internal struct. Signed-off-by: Maxim Uvarov --- platform/linux-generic/arch/default/odp_sysinfo_parse.c | 2 +-

[lng-odp] [PATCHv2 0/2] linux-generic: use default huge page size

A little bit more clean up things in patch v2. Looks like we also need to do some renames for internal things to drop odp_ prefix. So that only api functions should start with odp_, no any internal things. Is there any volunteer to do bunch of renames? :) Thanks, Maxim. Maxim Uvarov (2):

[lng-odp] [PATCH] linux-generic: use default huge page size

odp_shm_reserve() relays on huge page size to round up requested size. If 1 Gb pages present than parser takes it first as first alphabetical file name in sysfs. That lead to issue where all allocations wants 1 GB HP. This patch takes system default huge pages which are usually 2Mb, and has to be

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

On 05/10/16 15:53, nousi wrote: value is zero in all the huge page parameters. That is not the problem. Problem is that we have call: /** * Huge page size in bytes * * @return Huge page size in bytes */ uint64_t odp_sys_huge_page_size(void); which worked well for only one huge page

Re: [lng-odp] [API-NEXT, RFC, 1/1] api: cpu: performance profiling start/stop

Hi, Petri While we can continue processor-related discussions in Bill's new comprehensive email thread, about ODP-427 of how to guarantee locality of odp_cpu_xxx() APIs, can we make a decision between two choices in tomorrow's ARCH meeting? *Choice one: *constraint to ODP thread concept: every

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

some more debug logs. #ifdef MAP_HUGETLB huge_sz = odp_sys_huge_page_size(); <--- 1 GB here need_huge_page = (huge_sz && alloc_size > page_sz); /* munmap for huge pages requires sizes round up by page */ alloc_hp_size = (size + align + (huge_sz - 1)) & (-huge_sz); #endif *o/p:

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

value is zero in all the huge page parameters. root@ubuntu-15-10:/sys/kernel/mm/hugepages/hugepages-1048576kB# ls -l total 0 -r--r--r-- 1 root root 4096 May 10 18:15 free_hugepages -rw-r--r-- 1 root root 4096 May 10 18:15 nr_hugepages -rw-r--r-- 1 root root 4096 May 10 18:15

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

In my case there is 1 Gb huge page in: /sys/kernel/mm/hugepages and odp_shm_reserve() just rounds all allocations to 1 GB: #ifdef MAP_HUGETLB huge_sz = odp_sys_huge_page_size(); <--- 1 GB here need_huge_page = (huge_sz && alloc_size > page_sz); /* munmap for huge pages requires

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

Thanks, I see this error also - reason is to huge allocations 1024 MB for each queue. Let me think where we fail... Maxim. On 05/10/16 15:07, nousi wrote: Hi Maxim, Thanks for your support. mamp return below error; "mmap: Cannot allocate memory" "mount -t hugetlbfs none /mnt/hugetlbfs"

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

Hi Maxim, Thanks for your support. mamp return below error; "mmap: Cannot allocate memory" "mount -t hugetlbfs none /mnt/hugetlbfs" also does not help. Huge page allocation success in below two calls after that it fails. 1) odp_thread_globals 2) odp_buffer_pools Please have a look at the

Re: [lng-odp] [PATCH] linux-generic: correctly set arch as undefined

merged with fix for make distcheck On 9 May 2016 at 10:50, Bill Fischofer wrote: > > > On Mon, May 9, 2016 at 6:59 AM, Mike Holmes > wrote: > >> arch is defined as an architecture flag so that optimisations or >> necessary differences can be

Re: [lng-odp] [PATCH] linux-generic: arch: add missing cache line size definition for linux

On 10 May 2016 at 05:01, Mike Holmes wrote: > > > On 4 May 2016 at 12:47, Bill Fischofer wrote: > >> Use a default cache line size of 64 for otherwise unrecognized linux >> platforms. This resolves bug

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

Does: mount -t hugetlbfs none /mnt/hugetlbfs help? Maxim. On 10 May 2016 at 13:16, Maxim Uvarov wrote: > looks like you have enough free HP. Which error returns mmap() on try > with HP? > > On 10 May 2016 at 11:57, nousi wrote: > >>

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

looks like you have enough free HP. Which error returns mmap() on try with HP? On 10 May 2016 at 11:57, nousi wrote: > linaro@ubuntu-15-10:~$ cat /proc/sys/vm/nr_hugepages > 1024 > linaro@ubuntu-15-10:~$ > > linaro@ubuntu-15-10:~$ cat /proc/meminfo > MemTotal:

Re: [lng-odp] [PATCH] linux-generic: arch: add missing cache line size definition for linux

On 4 May 2016 at 12:47, Bill Fischofer wrote: > Use a default cache line size of 64 for otherwise unrecognized linux > platforms. This resolves bug https://bugs.linaro.org/show_bug.cgi?id=2218 > > Reported-by: Matias Elo > Signed-off-by: Bill

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

linaro@ubuntu-15-10:~$ cat /proc/sys/vm/nr_hugepages 1024 linaro@ubuntu-15-10:~$ linaro@ubuntu-15-10:~$ cat /proc/meminfo MemTotal:8061836 kB MemFree: 470516 kB MemAvailable:1901932 kB Buffers: 92600 kB Cached: 1939696 kB SwapCached: 7516 kB Active:

Re: [lng-odp] [PATCHv2 1/2] doc: userguide: add timer/timeout state diagram

Hi Bill, Just sent you an RFC with 2 FSMs. Had to fight a bit with graphviz. Not finished yet. I will continue working on it if you think that makes sense... Christophe. On 9 May 2016 at 17:15, Bill Fischofer wrote: > As we discussed in our call, we can both play

Re: [lng-odp] ODP Addressing Model

From: lng-odp [mailto:lng-odp-boun...@lists.linaro.org] On Behalf Of Bill Fischofer Sent: Tuesday, May 10, 2016 12:54 AM To: LNG ODP Mailman List Subject: [lng-odp] ODP Addressing Model The purpose of this thread is to summarize the discussions we've had over the

Re: [lng-odp] Fwd: No huge pages, fall back to normal pages

Please put output for: cat /proc/meminfo cat /proc/sys/vm/nr_hugepages Thank you, Maxim. On 10 May 2016 at 08:36, nousi wrote: > mmap is failing in "odp_shm_reserve" function (odp_queue_init_global() > ---> odp_shm_reserve() ---> odp_shm_reserve()) > > > debug logs: > >

Re: [lng-odp] [PATCH] linux-generic: timer: change assert to account for padding

Reviewed-by: Balasubramanian Manoharan On 10 May 2016 at 07:25, Bill Fischofer wrote: > The tick_buf_t struct may be larger than 16 bytes when a lock char is > needed so correct the ODP_STATIC_ASSERT to reflect this. This addresses > bug