date:20241111

On Mon, Nov 11, 2024 at 03:11:03PM +0100, Thomas Monjalon wrote:
> 31/10/2024 04:08, Andre Muezerie:
> > -Required version is Meson 0.57.
> > -
> > -Versions starting from 0.58 are unusable with LLVM toolchain
> > -because of a `Meson issue 
> > `_.
> > +Meson version 1.5.2 or newer is recommended.
> 
> I would prefer to know which minimal version is supported.
> Giving the latest version is not so helpful.
> 
> PS: when sending a new version, please reply to the very first version
> to avoid excessive indent in mail list, thanks.
> 
> 
Sorry for the late reply. 

The Meson issue previously mentioned in the documentation took a significant 
amount of time to get fixed. The fix only showed up in the main tree on August 
7: 
https://github.com/mesonbuild/meson/pull/11715#event-13793516174

Meson 1.5.2 was the first release containing that fix:
https://github.com/mesonbuild/meson/releases

That's the reason we added that specific version information to the patch, not
just because it was the newest. But I see how it looked like that was the 
reason.

I'll change the wording a little to make it more clear, and reply to the first 
version
as suggested.

Thanks for all the suggestions. They are greatly appreciated.

Re: [PATCH] net/iavf: add segment-length check to Tx prep

2024-11-11 Thread Medvedkin, Vladimir


Acked-by: Vladimir Medvedkin 

On 11/11/2024 16:42, Bruce Richardson wrote:

In the Tx prep function, the metadata checks were only checking the
packet length and ignoring the data length. For single-buffer packets we
can quickly check that the data length is the packet length.

Fixes: 19ee91c6bd9a ("net/iavf: check illegal packet sizes")
Cc: sta...@dpdk.org

Reported-by: Padraig Connolly 
Signed-off-by: Bruce Richardson 
---
  drivers/net/iavf/iavf_rxtx.c | 6 +-
  1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c
index 4850b9e381..6a093c6746 100644
--- a/drivers/net/iavf/iavf_rxtx.c
+++ b/drivers/net/iavf/iavf_rxtx.c
@@ -3677,7 +3677,11 @@ iavf_prep_pkts(__rte_unused void *tx_queue, struct 
rte_mbuf **tx_pkts,
return i;
}
  
-		if (m->pkt_len < IAVF_TX_MIN_PKT_LEN) {

+   /* valid packets are greater than min size, and single-buffer 
pkts
+* must have data_len == pkt_len
+*/
+   if (m->pkt_len < IAVF_TX_MIN_PKT_LEN ||
+   (m->nb_segs == 1 && m->data_len != m->pkt_len)) 
{
rte_errno = EINVAL;
return i;
}


--
Regards,
Vladimir

Re: [PATCH] power: fix a typo in the PM QoS guide

2024-11-11 Thread lihuisong (C)




在 2024/11/11 21:11, Thomas Monjalon 写道:

11/11/2024 13:52, Huisong Li:

-Some services are latency sensitive and very except the low resume time,
+Some services are latency sensitive and request the low resume time,

I suppose it should be "request a low resume time".

Ack, thanks




.

Re: [PATCH v10 00/10] net/zxdh: introduce net zxdh driver

07/11/2024 11:32, Junlong Wang:
> >On 11/6/2024 12:40 AM, Ferruh Yigit wrote:
> >> For series,
> >> Acked-by: Ferruh Yigit 
> >> 
> >> Series applied to dpdk-next-net/main, thanks.
> >>
> 
> > Hi Junlong,
> 
> > It seems we missed to mark driver as experimental, I will update it in
> > next-net.
> 
>   Sorry, I'm too careless, I will pay more attention next time.
>   Thank you very much.

I'm removing the useless #ifdef __cplusplus while pulling in main,
as we are trying to clean them in the repo.

Re: Re: [PATCH v1 2/2] /usertools/dpdk-devbind:add the binding and unbinding of platform device

2024-11-11 Thread liwench...@phytium.com.cn

Dear Stephen, You agree this pacth,  it means I don't need modify it and commit 
again?



liwench...@phytium.com.cn
 
From: Stephen Hemminger
Date: 2024-11-06 11:55
To: liwencheng
CC: dev
Subject: Re: [PATCH v1 2/2] /usertools/dpdk-devbind:add the binding and 
unbinding of platform device
On Fri,  1 Nov 2024 10:07:20 +
liwencheng  wrote:
 
> +def unbind_platform_one(dev_name):
> +filename = "/sys/bus/platform/devices/%s/driver" % dev_name
> +
> +if exists(filename):
> +try:
> +f = open(os.path.join(filename, "unbind"), "w")
> +except OSError as err:
> +sys.exit("Error: unbind failed for %s - Cannot open %s: %s" %
> + (dev_name, os.path.join(filename, "unbind"), err))
> +f.write(dev_name)
> +f.close()
> +filename = "/sys/bus/platform/devices/%s/driver_override" % dev_name
> +try:
> +f = open(filename, "w")
> +except OSError as err:
> +sys.exit("Error: unbind failed for %s - Cannot open %s: %s" %
> + (dev_name, filename, err))
> +f.write("")
> +f.close()
> +print("Successfully unbind platform device %s" % dev_name)
> +
> +
 
Looks fine, but probably worth the effort in the future to update to current
Python style guidelines. Applies to all of devbind.
 
Acked-by: Stephen Hemminger 
 
* Module dpdk-devbind
usertools/dpdk-devbind.py:1:0: C0114: Missing module docstring 
(missing-module-docstring)
usertools/dpdk-devbind.py:1:0: C0103: Module name "dpdk-devbind" doesn't 
conform to snake_case naming style (invalid-name)
usertools/dpdk-devbind.py:104:0: C0103: Constant name "loaded_modules" doesn't 
conform to UPPER_CASE naming style (invalid-name)
usertools/dpdk-devbind.py:107:0: C0103: Constant name "b_flag" doesn't conform 
to UPPER_CASE naming style (invalid-name)
usertools/dpdk-devbind.py:108:0: C0103: Constant name "status_flag" doesn't 
conform to UPPER_CASE naming style (invalid-name)
usertools/dpdk-devbind.py:109:0: C0103: Constant name "force_flag" doesn't 
conform to UPPER_CASE naming style (invalid-name)
usertools/dpdk-devbind.py:110:0: C0103: Constant name "noiommu_flag" doesn't 
conform to UPPER_CASE naming style (invalid-name)
usertools/dpdk-devbind.py:115:0: C0116: Missing function or method docstring 
(missing-function-docstring)
usertools/dpdk-devbind.py:116:4: W0603: Using the global statement 
(global-statement)
usertools/dpdk-devbind.py:142:17: W1514: Using open without explicitly 
specifying an encoding (unspecified-encoding)
usertools/dpdk-devbind.py:150:0: C0116: Missing function or method docstring 
(missing-function-docstring)
usertools/dpdk-devbind.py:151:4: W0601: Global variable 'platform_devices' 
undefined at the module level (global-variable-undefined)
usertools/dpdk-devbind.py:157:0: C0116: Missing function or method docstring 
(missing-function-docstring)
usertools/dpdk-devbind.py:171:4: W0603: Using the global statement 
(global-statement)
usertools/dpdk-devbind.py:212:33: C0209: Formatting a regular string which 
could be an f-string (consider-using-f-string)
usertools/dpdk-devbind.py:226:4: W0603: Using the global statement 
(global-statement)
usertools/dpdk-devbind.py:234:4: W0602: Using global for 'devices' but no 
assignment is done (global-variable-not-assigned)
usertools/dpdk-devbind.py:235:4: W0602: Using global for 'dpdk_drivers' but no 
assignment is done (global-variable-not-assigned)
usertools/dpdk-devbind.py:246:31: C0201: Consider iterating the dictionary 
directly instead of calling .keys() (consider-iterating-dictionary)
usertools/dpdk-devbind.py:248:31: C0201: Consider iterating the dictionary 
directly instead of calling .keys() (consider-iterating-dictionary)
usertools/dpdk-devbind.py:278:4: C0206: Consider iterating with .items() 
(consider-using-dict-items)
usertools/dpdk-devbind.py:278:13: C0201: Consider iterating the dictionary 
directly instead of calling .keys() (consider-iterating-dictionary)
usertools/dpdk-devbind.py:299:51: C0209: Formatting a regular string which 
could be an f-string (consider-using-f-string)
usertools/dpdk-devbind.py:230:0: R0912: Too many branches (21/12) 
(too-many-branches)
usertools/dpdk-devbind.py:311:0: C0116: Missing function or method docstring 
(missing-function-docstring)
usertools/dpdk-devbind.py:312:4: C0200: Consider using enumerate instead of 
iterating with range and len (consider-using-enumerate)
usertools/dpdk-devbind.py:312:4: R1702: Too many nested blocks (6/5) 
(too-many-nested-blocks)
usertools/dpdk-devbind.py:343:4: C0206: Consider iterating with .items() 
(consider-using-dict-items)
usertools/dpdk-devbind.py:343:13: C0201: Consider iterating the dictionary 
directly instead of calling .keys() (consider-iterating-dictionary)
usertools/dpdk-devbind.py:352:21: C0209: Formatting a regular string which 
could be an f-string (consider-using-f-string)
usertools/dpdk-devbind.py:356:0: C0116: Missing function or method docstrin

Re: [PATCH v2 1/2] config/arm: strict use of -mcpu for supported CPUs

Please check the CI results, there are some failures.


29/10/2024 02:59, Wathsala Vithanage:
> Arm recommends using -mcpu over -march and march-extensions when the
> compiler supports the target CPU (neoverse-n1 etc.). Arm build so far
> has been an amalgam of -mcpu and -march. When march is in use, it has
> been the case so far to silently fall back to a downgraded march when
> the compiler does not support the requested march. This is unnecessary
> and confusing to an end user who could be building DPDK with a
> particular ISA version and performance expectations in mind.
> 
> This patch aims to rectify both the above issues. For part numbers that
> has a corresponding -mcpu, this patch removes all references to march
> and march_features from meson.build. For those SOCs that do not have a
> corresponding -mcpu, a pseudo cpu name in the format mcpu_ is
> introduced and referenced in the mcpu field in the part number
> dictionary of the part_number_config dictionary. The definition of the
> mcpu_ must be provided in the mcpu_defs dictionary.
> Each mcpu_ dictionary in the mcpu_defs have a march field and
> a march_extensions field to construct the optimal compiler setting for
> an SOC that has no corresponding -mcpu value. This patch alters the
> behavior of the build system such that it will no longer silently fall
> back to an older ISA version, it will only perform what's specified in
> the build dictionaries, if the compiler does not support the specified
> configuration it will fail with an error message.
> 
> How to add a new SOC to the build system with these changes?
> If compiler supports mcpu follow the usual practice but without march
> and march_features fields in the part number dictionary. If mcpu is not
> available or buggy (misses certain features) for some reason follow the
> same process but set mcpu field to a string in the form 'mcpu_foo'
> (pseudo mcpu mentioned earlier). Then in the mcpu_defs dictionary add
> mcpu_foo dictionary as shown.
> 
> 'mcpu_foo': {
>'march': 'armv8.2-a',
>'march_extensions': [rcpc]
> }
> 
> march_extensions is a comma separated list of march extensions supported
> by the compiler such as sve, crypto etc. Empty match_extensions allowed
> as use of such extensions are optional.
> 
> Signed-off-by: Wathsala Vithanage 
> Reviewed-by: Dhruv Tripathi

Re: [PATCH v3 0/2] examples/l3fwd fixes for ACL mode

07/11/2024 19:50, Konstantin Ananyev:
> v2 -> v3
> - rebased against main
> - fixed check-patch warnings (grammar)
> 
> v1 -> v2
> As Song Jiale pointed outprevious fix is not enough to fix
> the problem he is observing with l3fwd in ACl mode:
> https://bugs.dpdk.org/show_bug.cgi?id=1502
> This is a second attempt to fix it.
> 
> Konstantin Ananyev (2):
>   examples/l3fwd: fix read beyond array bondaries
>   examples/l3fwd: fix read beyond array boundaries in ACL mode

Applied, thanks.
Note: your mailmap change is squashed here as it does not bring
a high value in the git history :)

Re: [PATCH v5] doc: modernize build instructions on Windows

31/10/2024 04:08, Andre Muezerie:
> -Required version is Meson 0.57.
> -
> -Versions starting from 0.58 are unusable with LLVM toolchain
> -because of a `Meson issue 
> `_.
> +Meson version 1.5.2 or newer is recommended.

I would prefer to know which minimal version is supported.
Giving the latest version is not so helpful.

PS: when sending a new version, please reply to the very first version
to avoid excessive indent in mail list, thanks.

[PATCH] net/iavf: add segment-length check to Tx prep

2024-11-11 Thread Bruce Richardson

In the Tx prep function, the metadata checks were only checking the
packet length and ignoring the data length. For single-buffer packets we
can quickly check that the data length is the packet length.

Fixes: 19ee91c6bd9a ("net/iavf: check illegal packet sizes")
Cc: sta...@dpdk.org

Reported-by: Padraig Connolly 
Signed-off-by: Bruce Richardson 
---
 drivers/net/iavf/iavf_rxtx.c | 6 +-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c
index 4850b9e381..6a093c6746 100644
--- a/drivers/net/iavf/iavf_rxtx.c
+++ b/drivers/net/iavf/iavf_rxtx.c
@@ -3677,7 +3677,11 @@ iavf_prep_pkts(__rte_unused void *tx_queue, struct 
rte_mbuf **tx_pkts,
return i;
}
 
-   if (m->pkt_len < IAVF_TX_MIN_PKT_LEN) {
+   /* valid packets are greater than min size, and single-buffer 
pkts
+* must have data_len == pkt_len
+*/
+   if (m->pkt_len < IAVF_TX_MIN_PKT_LEN ||
+   (m->nb_segs == 1 && m->data_len != m->pkt_len)) 
{
rte_errno = EINVAL;
return i;
}
-- 
2.43.0

Re: [PATCH v2 1/2] config/arm: strict use of -mcpu for supported CPUs

11/11/2024 18:57, Wathsala Wathawana Vithanage:
> 
> > Subject: Re: [PATCH v2 1/2] config/arm: strict use of -mcpu for supported
> > CPUs
> > 
> > Please check the CI results, there are some failures.
> > 
> 
> Hi Thomas,
> 
> CI is failing because gcc version 11 does not enable crypto extension for 
> thunderx2,
> with -mcpu=thunderx2t99 unless +crypto feature modifier is used.
> Goal of this patch is to drop use of such feature modifiers when an -mcpu is 
> available
> for a given platform.
> This issue is fixed in gcc version 13 for thunderx2t99. 

We cannot afford dropping the support of GCC 11.
Which solution do you propose?

Re: [PATCH v6] doc: modernize build instructions on Windows

11/11/2024 17:14, Andre Muezerie:
> +The minimal Meson supported version is 1.5.2.

OK looks good, thanks.

[PATCH v7 00/21] remove use of VLAs for Windows

As per guidance technical board meeting 2024/04/17. This series
removes the use of VLAs from code built for Windows for all 3
toolchains. If there are additional opportunities to convert VLAs
to regular C arrays please provide the details for incorporation
into the series.

MSVC does not support VLAs, replace VLAs with standard C arrays
or alloca(). alloca() is available for all toolchain/platform
combinations officially supported by DPDK.

v7:
 * remove use of VLA from new file which sneaked in during review

v6:
 * remove use of VLA from new test code added recently
 * fix title for patch 08/20

v5:
 * add patches for net/ice, net/ixgbe and gro
   from Konstantin Ananyev from
https://patchwork.dpdk.org/project/dpdk/list/?series=31972&archive=both&state=*
 * address debug_autotest ASan failure
 * address array-bound error in bitset_autotest with gcc-13

v4:
 * rebase and adapt for changes made in main since v3 was sent
 * use fixed maximum array size instead of VLA when doable

v3:
 * address checkpatch/check git log warnings (minor typos)

v2:
 * replace patches for ethdev, hash, rcu and include new
   patches for eal from Konstantin Ananyev
   from https://patchwork.dpdk.org/project/dpdk/list/?series=31781

Andre Muezerie (3):
  test: remove use of VLAs for Windows built code in bitset tests
  app/testpmd: remove use of VLAs for Windows built code in
shared_rxq_fwd
  hash: remove use of VLAs by using standard arrays

Konstantin Ananyev (10):
  eal/linux: remove use of VLAs
  eal/common: remove use of VLAs
  ethdev: remove use of VLAs for Windows built code
  hash: remove use of VLAs for Windows built code
  hash/thash: remove use of VLAs for Windows built
  rcu: remove use of VLAs for Windows built code
  gro: fix overwrite unprocessed packets
  gro: remove use of VLAs
  net/ixgbe: remove use of VLAs
  net/ice: remove use of VLAs

Tyler Retzlaff (8):
  eal: include header required for alloca
  app/testpmd: remove use of VLAs for Windows built
  test: remove use of VLAs for Windows built code
  common/idpf: remove use of VLAs for Windows built code
  net/i40e: remove use of VLAs for Windows built code
  common/mlx5: remove use of VLAs for Windows built code
  net/mlx5: remove use of VLAs for Windows built code
  build: enable vla warnings on Windows built code

 app/test-pmd/cmdline.c|   2 +-
 app/test-pmd/cmdline_flow.c   |  15 +-
 app/test-pmd/config.c |  16 +-
 app/test-pmd/shared_rxq_fwd.c |   2 +-
 app/test/meson.build  |   5 +
 app/test/test.c   |   2 +-
 app/test/test_bitset.c|  55 +++---
 app/test/test_cmdline_string.c|   2 +-
 app/test/test_cryptodev.c |  34 ++--
 app/test/test_cryptodev_blockcipher.c |   4 +-
 app/test/test_cryptodev_crosscheck.c  |   2 +-
 app/test/test_dmadev.c|   9 +-
 app/test/test_hash.c  |  14 +-
 app/test/test_lcore_var_perf.c|   2 +-
 app/test/test_mempool.c   |  25 +--
 app/test/test_reassembly_perf.c   |   4 +-
 app/test/test_reorder.c   |  48 ++---
 app/test/test_service_cores.c |   9 +-
 app/test/test_thash.c |   7 +-
 config/meson.build|   4 +
 drivers/common/idpf/idpf_common_rxtx.c|   2 +-
 drivers/common/idpf/idpf_common_rxtx_avx512.c |   6 +-
 drivers/common/mlx5/mlx5_common.h |   4 +-
 drivers/common/mlx5/mlx5_devx_cmds.c  |   7 +-
 drivers/net/i40e/i40e_testpmd.c   |   5 +-
 drivers/net/ice/ice_rxtx.c|  18 +-
 drivers/net/ice/ice_rxtx.h|   2 +
 drivers/net/ixgbe/ixgbe_ethdev.c  |  21 +-
 drivers/net/ixgbe/ixgbe_rxtx_vec_common.h |   4 +-
 drivers/net/mlx5/mlx5.c   |   5 +-
 drivers/net/mlx5/mlx5_flow.c  |   6 +-
 lib/eal/common/eal_common_proc.c  |   5 +-
 lib/eal/linux/eal_interrupts.c|  32 ++-
 lib/eal/linux/include/rte_os.h|   1 +
 lib/eal/windows/include/rte_os.h  |   1 +
 lib/ethdev/rte_ethdev.c   | 183 +++---
 lib/gro/rte_gro.c |  42 ++--
 lib/hash/rte_cuckoo_hash.c|   4 +-
 lib/hash/rte_thash.c  |   2 +-
 lib/hash/rte_thash.h  |   8 +
 lib/hash/rte_thash_gf2_poly_math.c|   7 +-
 lib/rcu/rte_rcu_qsbr.c|   7 +-
 lib/rcu/rte_rcu_qsbr.h|   5 +
 43 files changed, 377 insertions(+), 261 deletions(-)

--
2.34.1

[PATCH v7 02/21] eal/linux: remove use of VLAs

From: Konstantin Ananyev 

1) ./lib/eal/linux/eal_interrupts.c:1073:16
: warning: ISO C90 forbids variable length array ‘events’

MSVC does not support VLAs. Use alloca() to allocate the memory on
the stack.

2) ./lib/eal/linux/eal_interrupts.c:1319:16
: warning: ISO C90 forbids variable length array ‘evs’

make eal_epoll_wait() use a fixed size array and use it though multiple
iterations to process up to @maxevents events.
Note that technically it is not one to one replacement, as here we might
reduce number of events returned by first call to epoll_wait(..., timeout);

Signed-off-by: Konstantin Ananyev 
---
 lib/eal/linux/eal_interrupts.c | 32 +++-
 1 file changed, 27 insertions(+), 5 deletions(-)

diff --git a/lib/eal/linux/eal_interrupts.c b/lib/eal/linux/eal_interrupts.c
index 6436f796eb..23039964fc 100644
--- a/lib/eal/linux/eal_interrupts.c
+++ b/lib/eal/linux/eal_interrupts.c
@@ -34,6 +34,8 @@
 #define EAL_INTR_EPOLL_WAIT_FOREVER (-1)
 #define NB_OTHER_INTR   1
 
+#define MAX_ITER_EVNUM RTE_EVENT_ETH_INTR_RING_SIZE
+
 static RTE_DEFINE_PER_LCORE(int, _epfd) = -1; /**< epoll fd per thread */
 
 /**
@@ -1070,7 +1072,7 @@ eal_intr_process_interrupts(struct epoll_event *events, 
int nfds)
 static void
 eal_intr_handle_interrupts(int pfd, unsigned totalfds)
 {
-   struct epoll_event events[totalfds];
+   struct epoll_event *events = alloca(sizeof(struct epoll_event) * 
totalfds);
int nfds = 0;
 
for(;;) {
@@ -1316,8 +1318,9 @@ static int
 eal_epoll_wait(int epfd, struct rte_epoll_event *events,
   int maxevents, int timeout, bool interruptible)
 {
-   struct epoll_event evs[maxevents];
int rc;
+   uint32_t i, k, n, num;
+   struct epoll_event evs[MAX_ITER_EVNUM];
 
if (!events) {
EAL_LOG(ERR, "rte_epoll_event can't be NULL");
@@ -1328,12 +1331,31 @@ eal_epoll_wait(int epfd, struct rte_epoll_event *events,
if (epfd == RTE_EPOLL_PER_THREAD)
epfd = rte_intr_tls_epfd();
 
+   num = maxevents;
+   n = RTE_MIN(RTE_DIM(evs), num);
+
+   /* Process events in chunks of MAX_ITER_EVNUM */
+
while (1) {
-   rc = epoll_wait(epfd, evs, maxevents, timeout);
+   rc = epoll_wait(epfd, evs, n, timeout);
if (likely(rc > 0)) {
+
/* epoll_wait has at least one fd ready to read */
-   rc = eal_epoll_process_event(evs, rc, events);
-   break;
+   for (i = 0, k = 0; rc > 0;) {
+   k += rc;
+   rc = eal_epoll_process_event(evs, rc,
+   events + i);
+   i += rc;
+
+   /*
+* try to read more events that are already
+* available (up to maxevents in total).
+*/
+   n = RTE_MIN(RTE_DIM(evs), num - k);
+   rc = (n == 0) ? 0 : epoll_wait(epfd, evs, n, 0);
+   }
+   return i;
+
} else if (rc < 0) {
if (errno == EINTR) {
if (interruptible)
-- 
2.34.1

[PATCH v7 04/21] ethdev: remove use of VLAs for Windows built code

From: Konstantin Ananyev 

1) ./lib/ethdev/rte_ethdev.c:3244:16
: warning: ISO C90 forbids variable length array ‘xstats_names’
2) ./lib/ethdev/rte_ethdev.c:3345:17
: warning: ISO C90 forbids variable length array ‘ids_copy’
3) ./lib/ethdev/rte_ethdev.c:3538:16
: warning: ISO C90 forbids variable length array ‘xstats’
4) ./lib/ethdev/rte_ethdev.c:3554:17
: warning: ISO C90 forbids variable length array ‘ids_copy’

For 1) and 3) - just replaced VLA with arrays allocated from heap.
As I understand xstats extraction belongs to control-path, so extra
calloc/free is hopefully acceptable.
Also ethdev xstats already doing that within
rte_eth_xstats_get_id_by_name().
For 2) and 4) changed the code to use fixed size array and call
appropriate devops function several times, if needed.

Signed-off-by: Konstantin Ananyev 
---
 lib/ethdev/rte_ethdev.c | 183 +---
 1 file changed, 113 insertions(+), 70 deletions(-)

diff --git a/lib/ethdev/rte_ethdev.c b/lib/ethdev/rte_ethdev.c
index 6413c54e3b..09cc4764c3 100644
--- a/lib/ethdev/rte_ethdev.c
+++ b/lib/ethdev/rte_ethdev.c
@@ -36,6 +36,8 @@
 #include "ethdev_trace.h"
 #include "sff_telemetry.h"
 
+#define ETH_XSTATS_ITER_NUM0x100
+
 struct rte_eth_dev rte_eth_devices[RTE_MAX_ETHPORTS];
 
 /* public fast-path API */
@@ -3308,7 +3310,8 @@ int
 rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name,
uint64_t *id)
 {
-   int cnt_xstats, idx_xstat;
+   int cnt_xstats, idx_xstat, rc;
+   struct rte_eth_xstat_name *xstats_names;
 
RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
 
@@ -3334,26 +3337,33 @@ rte_eth_xstats_get_id_by_name(uint16_t port_id, const 
char *xstat_name,
}
 
/* Get id-name lookup table */
-   struct rte_eth_xstat_name xstats_names[cnt_xstats];
+   xstats_names = calloc(cnt_xstats, sizeof(xstats_names[0]));
+   if (xstats_names == NULL) {
+   RTE_ETHDEV_LOG_LINE(ERR, "Can't allocate memory");
+   return -ENOMEM;
+   }
 
if (cnt_xstats != rte_eth_xstats_get_names_by_id(
port_id, xstats_names, cnt_xstats, NULL)) {
RTE_ETHDEV_LOG_LINE(ERR, "Cannot get xstats lookup");
+   free(xstats_names);
return -1;
}
 
+   rc = -EINVAL;
for (idx_xstat = 0; idx_xstat < cnt_xstats; idx_xstat++) {
if (!strcmp(xstats_names[idx_xstat].name, xstat_name)) {
*id = idx_xstat;
 
rte_eth_trace_xstats_get_id_by_name(port_id,
xstat_name, *id);
-
-   return 0;
+   rc = 0;
+   break;
};
}
 
-   return -EINVAL;
+   free(xstats_names);
+   return rc;
 }
 
 /* retrieve basic stats names */
@@ -3399,6 +3409,38 @@ eth_basic_stats_get_names(struct rte_eth_dev *dev,
return cnt_used_entries;
 }
 
+static int
+eth_xstats_get_by_name_by_id(struct rte_eth_dev *dev, const uint64_t *ids,
+   struct rte_eth_xstat_name *xstats_names, uint32_t size,
+   uint32_t basic_count)
+{
+   int32_t rc;
+   uint32_t i, k, m, n;
+   uint64_t ids_copy[ETH_XSTATS_ITER_NUM];
+
+   m = 0;
+   for (n = 0; n != size; n += k) {
+
+   k = RTE_MIN(size - n, RTE_DIM(ids_copy));
+
+   /*
+* Convert ids to xstats ids that PMD knows.
+* ids known by user are basic + extended stats.
+*/
+   for (i = 0; i < k; i++)
+   ids_copy[i] = ids[n + i] - basic_count;
+
+   rc = (*dev->dev_ops->xstats_get_names_by_id)(dev, ids_copy,
+   xstats_names + m, k);
+   if (rc < 0)
+   return rc;
+   m += rc;
+   }
+
+   return m;
+}
+
+
 /* retrieve ethdev extended statistics names */
 int
 rte_eth_xstats_get_names_by_id(uint16_t port_id,
@@ -3406,9 +3448,8 @@ rte_eth_xstats_get_names_by_id(uint16_t port_id,
uint64_t *ids)
 {
struct rte_eth_xstat_name *xstats_names_copy;
-   unsigned int no_basic_stat_requested = 1;
-   unsigned int no_ext_stat_requested = 1;
unsigned int expected_entries;
+   unsigned int nb_basic_stats;
unsigned int basic_count;
struct rte_eth_dev *dev;
unsigned int i;
@@ -3434,27 +3475,18 @@ rte_eth_xstats_get_names_by_id(uint16_t port_id,
if (ids && !xstats_names)
return -EINVAL;
 
-   if (ids && dev->dev_ops->xstats_get_names_by_id != NULL && size > 0) {
-   uint64_t ids_copy[size];
-
-   for (i = 0; i < size; i++) {
-   if (ids[i] < basic_count) {
-   no_basic_stat_requested = 0;
-   break;
-   }
-
-

[PATCH v7 05/21] hash: remove use of VLAs for Windows built code

From: Konstantin Ananyev 

1) ./lib/hash/rte_cuckoo_hash.c:2362:9
: warning: ISO C90 forbids variable length array ‘positions’
2) ../lib/hash/rte_cuckoo_hash.c:2478:9
: warning: ISO C90 forbids variable length array ‘positions’

Both rte_hash_lookup_bulk_data() and
rte_hash_lookup_with_hash_bulk_data() expect
@num_keys <= RTE_HASH_LOOKUP_BULK_MAX.
So, for both cases it should be safe to replace VLA with fixed size
array.

Signed-off-by: Konstantin Ananyev 
---
 lib/hash/rte_cuckoo_hash.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/hash/rte_cuckoo_hash.c b/lib/hash/rte_cuckoo_hash.c
index 9575e8aa0c..fc93182efe 100644
--- a/lib/hash/rte_cuckoo_hash.c
+++ b/lib/hash/rte_cuckoo_hash.c
@@ -2418,7 +2418,7 @@ rte_hash_lookup_bulk_data(const struct rte_hash *h, const 
void **keys,
(num_keys > RTE_HASH_LOOKUP_BULK_MAX) ||
(hit_mask == NULL)), -EINVAL);
 
-   int32_t positions[num_keys];
+   int32_t positions[RTE_HASH_LOOKUP_BULK_MAX];
 
__rte_hash_lookup_bulk(h, keys, num_keys, positions, hit_mask, data);
 
@@ -2534,7 +2534,7 @@ rte_hash_lookup_with_hash_bulk_data(const struct rte_hash 
*h,
(num_keys > RTE_HASH_LOOKUP_BULK_MAX) ||
(hit_mask == NULL)), -EINVAL);
 
-   int32_t positions[num_keys];
+   int32_t positions[RTE_HASH_LOOKUP_BULK_MAX];
 
__rte_hash_lookup_with_hash_bulk(h, keys, sig, num_keys,
positions, hit_mask, data);
-- 
2.34.1

[PATCH v7 03/21] eal/common: remove use of VLAs

From: Konstantin Ananyev 

1) ../lib/eal/common/eal_common_proc.c:695:15
: warning: variable length array used

As msg->num_fds should not exceed RTE_MP_MAX_FD_NUM, replaced
it with fixed size array.

Signed-off-by: Konstantin Ananyev 
---
 lib/eal/common/eal_common_proc.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/lib/eal/common/eal_common_proc.c b/lib/eal/common/eal_common_proc.c
index d24093937c..201973c5db 100644
--- a/lib/eal/common/eal_common_proc.c
+++ b/lib/eal/common/eal_common_proc.c
@@ -692,7 +692,8 @@ send_msg(const char *dst_path, struct rte_mp_msg *msg, int 
type)
struct sockaddr_un dst;
struct mp_msg_internal m;
int fd_size = msg->num_fds * sizeof(int);
-   char control[CMSG_SPACE(fd_size)];
+   const int32_t control_sz = CMSG_SPACE(fd_size);
+   char control[CMSG_SPACE(sizeof(msg->fds))];
 
m.type = type;
memcpy(&m.msg, msg, sizeof(*msg));
@@ -712,7 +713,7 @@ send_msg(const char *dst_path, struct rte_mp_msg *msg, int 
type)
msgh.msg_iov = &iov;
msgh.msg_iovlen = 1;
msgh.msg_control = control;
-   msgh.msg_controllen = sizeof(control);
+   msgh.msg_controllen = control_sz;
 
cmsg = CMSG_FIRSTHDR(&msgh);
cmsg->cmsg_len = CMSG_LEN(fd_size);
-- 
2.34.1

[PATCH v7 01/21] eal: include header required for alloca

From: Tyler Retzlaff 

Include alloca.h for Linux and malloc.h for Windows to get declaration
of alloca().

Signed-off-by: Tyler Retzlaff 
---
 lib/eal/linux/include/rte_os.h   | 1 +
 lib/eal/windows/include/rte_os.h | 1 +
 2 files changed, 2 insertions(+)

diff --git a/lib/eal/linux/include/rte_os.h b/lib/eal/linux/include/rte_os.h
index 35c07c70cb..20eff0409a 100644
--- a/lib/eal/linux/include/rte_os.h
+++ b/lib/eal/linux/include/rte_os.h
@@ -10,6 +10,7 @@
  * which is not supported natively or named differently in Linux.
  */
 
+#include 
 #include 
 #include 
 
diff --git a/lib/eal/windows/include/rte_os.h b/lib/eal/windows/include/rte_os.h
index 9d69467aaa..d09adeb3b4 100644
--- a/lib/eal/windows/include/rte_os.h
+++ b/lib/eal/windows/include/rte_os.h
@@ -13,6 +13,7 @@
 #include 
 #include 
 #include 
+#include 
 
 #include 
 
-- 
2.34.1

[PATCH v7 11/21] net/ice: remove use of VLAs

From: Konstantin Ananyev 

../drivers/net/ice/ice_rxtx.c:1871:29: warning:
variable length array used [-Wvla]

Here VLA is used as a temp array for mbufs that will be used as a split
RX data buffers.
As at any given time only one thread can do RX from particular queue,
at rx_queue_setup() we can allocate extra space for that array, and then
safely use it at RX fast-path.

Signed-off-by: Konstantin Ananyev 
Acked-by: Anatoly Burakov 
---
 drivers/net/ice/ice_rxtx.c | 18 --
 drivers/net/ice/ice_rxtx.h |  2 ++
 2 files changed, 14 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ice/ice_rxtx.c b/drivers/net/ice/ice_rxtx.c
index 024d97cb46..a52a759031 100644
--- a/drivers/net/ice/ice_rxtx.c
+++ b/drivers/net/ice/ice_rxtx.c
@@ -1176,7 +1176,7 @@ ice_rx_queue_setup(struct rte_eth_dev *dev,
struct ice_vsi *vsi = pf->main_vsi;
struct ice_rx_queue *rxq;
const struct rte_memzone *rz;
-   uint32_t ring_size;
+   uint32_t ring_size, tlen;
uint16_t len;
int use_def_burst_func = 1;
uint64_t offloads;
@@ -1284,9 +1284,14 @@ ice_rx_queue_setup(struct rte_eth_dev *dev,
/* always reserve more for bulk alloc */
len = (uint16_t)(nb_desc + ICE_RX_MAX_BURST);
 
+   /* allocate extra entries for SW split buffer */
+   tlen = ((rxq->offloads & RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT) != 0) ?
+   rxq->rx_free_thresh : 0;
+   tlen += len;
+
/* Allocate the software ring. */
rxq->sw_ring = rte_zmalloc_socket(NULL,
- sizeof(struct ice_rx_entry) * len,
+ sizeof(struct ice_rx_entry) * tlen,
  RTE_CACHE_LINE_SIZE,
  socket_id);
if (!rxq->sw_ring) {
@@ -1295,6 +1300,8 @@ ice_rx_queue_setup(struct rte_eth_dev *dev,
return -ENOMEM;
}
 
+   rxq->sw_split_buf = (tlen == len) ? NULL : rxq->sw_ring + len;
+
ice_reset_rx_queue(rxq);
rxq->q_set = true;
dev->data->rx_queues[queue_idx] = rxq;
@@ -1873,7 +1880,6 @@ ice_rx_alloc_bufs(struct ice_rx_queue *rxq)
uint64_t dma_addr;
int diag, diag_pay;
uint64_t pay_addr;
-   struct rte_mbuf *mbufs_pay[rxq->rx_free_thresh];
 
/* Allocate buffers in bulk */
alloc_idx = (uint16_t)(rxq->rx_free_trigger -
@@ -1888,7 +1894,7 @@ ice_rx_alloc_bufs(struct ice_rx_queue *rxq)
 
if (rxq->offloads & RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT) {
diag_pay = rte_mempool_get_bulk(rxq->rxseg[1].mp,
-   (void *)mbufs_pay, rxq->rx_free_thresh);
+   (void *)rxq->sw_split_buf, rxq->rx_free_thresh);
if (unlikely(diag_pay != 0)) {
PMD_RX_LOG(ERR, "Failed to get payload mbufs in bulk");
return -ENOMEM;
@@ -1913,8 +1919,8 @@ ice_rx_alloc_bufs(struct ice_rx_queue *rxq)
rxdp[i].read.hdr_addr = 0;
rxdp[i].read.pkt_addr = dma_addr;
} else {
-   mb->next = mbufs_pay[i];
-   pay_addr = 
rte_cpu_to_le_64(rte_mbuf_data_iova_default(mbufs_pay[i]));
+   mb->next = rxq->sw_split_buf[i].mbuf;
+   pay_addr = 
rte_cpu_to_le_64(rte_mbuf_data_iova_default(mb->next));
rxdp[i].read.hdr_addr = dma_addr;
rxdp[i].read.pkt_addr = pay_addr;
}
diff --git a/drivers/net/ice/ice_rxtx.h b/drivers/net/ice/ice_rxtx.h
index f7276cfc9f..d0f0b6c1d2 100644
--- a/drivers/net/ice/ice_rxtx.h
+++ b/drivers/net/ice/ice_rxtx.h
@@ -139,6 +139,8 @@ struct ice_rx_queue {
uint32_t hw_time_high; /* high 32 bits of timestamp */
uint32_t hw_time_low; /* low 32 bits of timestamp */
uint64_t hw_time_update; /* SW time of HW record updating */
+   struct ice_rx_entry *sw_split_buf;
+   /* address of temp buffer for RX split mbufs */
struct rte_eth_rxseg_split rxseg[ICE_RX_MAX_NSEG];
uint32_t rxseg_nb;
bool ts_enable; /* if rxq timestamp is enabled */
-- 
2.34.1

[PATCH v7 16/21] common/mlx5: remove use of VLAs for Windows built code

From: Tyler Retzlaff 

MSVC does not support VLAs, replace VLAs with standard C arrays
or alloca(). alloca() is available for all toolchain/platform
combinations officially supported by DPDK.

Signed-off-by: Tyler Retzlaff 
---
 drivers/common/mlx5/mlx5_common.h| 4 ++--
 drivers/common/mlx5/mlx5_devx_cmds.c | 7 +++
 2 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/drivers/common/mlx5/mlx5_common.h 
b/drivers/common/mlx5/mlx5_common.h
index 1abd1e8239..f29f06a86e 100644
--- a/drivers/common/mlx5/mlx5_common.h
+++ b/drivers/common/mlx5/mlx5_common.h
@@ -125,10 +125,10 @@ mlx5_fp_debug_enabled(void)
 /* Allocate a buffer on the stack and fill it with a printf format string. */
 #define MKSTR(name, ...) \
int mkstr_size_##name = snprintf(NULL, 0, "" __VA_ARGS__); \
-   char name[mkstr_size_##name + 1]; \
+   char *name = alloca(mkstr_size_##name + 1); \
\
memset(name, 0, mkstr_size_##name + 1); \
-   snprintf(name, sizeof(name), "" __VA_ARGS__)
+   snprintf(name, mkstr_size_##name + 1, "" __VA_ARGS__)
 
 enum {
PCI_VENDOR_ID_MELLANOX = 0x15b3,
diff --git a/drivers/common/mlx5/mlx5_devx_cmds.c 
b/drivers/common/mlx5/mlx5_devx_cmds.c
index a75f011750..804ee67cd6 100644
--- a/drivers/common/mlx5/mlx5_devx_cmds.c
+++ b/drivers/common/mlx5/mlx5_devx_cmds.c
@@ -284,10 +284,9 @@ mlx5_devx_cmd_flow_counter_query(struct mlx5_devx_obj *dcs,
 void *cmd_comp,
 uint64_t async_id)
 {
-   int out_len = MLX5_ST_SZ_BYTES(query_flow_counter_out) +
-   MLX5_ST_SZ_BYTES(traffic_counter);
-   uint32_t out[out_len];
+   uint32_t out[MLX5_ST_SZ_BYTES(query_flow_counter_out) + 
MLX5_ST_SZ_BYTES(traffic_counter)];
uint32_t in[MLX5_ST_SZ_DW(query_flow_counter_in)] = {0};
+   const int out_len = RTE_DIM(out);
void *stats;
int rc;
 
@@ -346,7 +345,7 @@ mlx5_devx_cmd_mkey_create(void *ctx,
int klm_num = attr->klm_num;
int in_size_dw = MLX5_ST_SZ_DW(create_mkey_in) +
 (klm_num ? RTE_ALIGN(klm_num, 4) : 0) * MLX5_ST_SZ_DW(klm);
-   uint32_t in[in_size_dw];
+   uint32_t *in = alloca(sizeof(uint32_t) * in_size_dw);
uint32_t out[MLX5_ST_SZ_DW(create_mkey_out)] = {0};
void *mkc;
struct mlx5_devx_obj *mkey = mlx5_malloc(MLX5_MEM_ZERO, sizeof(*mkey),
-- 
2.34.1

[PATCH v7 06/21] hash/thash: remove use of VLAs for Windows built

From: Konstantin Ananyev 

1) ./lib/hash/rte_thash.c:774:9
: warning: ISO C90 forbids variable length array ‘tmp_tuple’

>From my understanding, tuple size here should never exceed
sizeof(union rte_thash_tuple), so it should be safe to replace VLA with
fixed size array.

Signed-off-by: Konstantin Ananyev 
---
 lib/hash/rte_thash.c | 2 +-
 lib/hash/rte_thash.h | 8 
 2 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/lib/hash/rte_thash.c b/lib/hash/rte_thash.c
index 336c228e64..842e3ad85d 100644
--- a/lib/hash/rte_thash.c
+++ b/lib/hash/rte_thash.c
@@ -761,7 +761,7 @@ rte_thash_adjust_tuple(struct rte_thash_ctx *ctx,
uint32_t desired_value, unsigned int attempts,
rte_thash_check_tuple_t fn, void *userdata)
 {
-   uint32_t tmp_tuple[tuple_len / sizeof(uint32_t)];
+   uint32_t tmp_tuple[RTE_THASH_MAX_L4_LEN];
unsigned int i, j, ret = 0;
uint32_t hash, adj_bits;
const uint8_t *hash_key;
diff --git a/lib/hash/rte_thash.h b/lib/hash/rte_thash.h
index c0af5968df..427246ad2e 100644
--- a/lib/hash/rte_thash.h
+++ b/lib/hash/rte_thash.h
@@ -121,6 +121,14 @@ __rte_internal
 uint32_t
 thash_get_rand_poly(uint32_t poly_degree);
 
+/**
+ * maximum length in dwords of input tuple to
+ * calculate hash of ipv(4|6) header +
+ * transport header
+ */
+#define RTE_THASH_MAX_L4_LEN   \
+   ((sizeof(union rte_thash_tuple)) / sizeof(uint32_t))
+
 /**
  * Prepare special converted key to use with rte_softrss_be()
  * @param orig
-- 
2.34.1

[PATCH v7 08/21] gro: fix overwrite unprocessed packets

From: Konstantin Ananyev 

gro_vxlan_tcp4_tbl_timeout_flush() is called without taking into account
that first entries in pkts[] can be already occupied by
un-processed packets.

Fixes: 74080d7dcf31 ("gro: support IPv6 for TCP")
Cc: sta...@dpdk.org

Signed-off-by: Konstantin Ananyev 
Acked-by: Ferruh Yigit 
---
 lib/gro/rte_gro.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/gro/rte_gro.c b/lib/gro/rte_gro.c
index d824eebd93..db86117609 100644
--- a/lib/gro/rte_gro.c
+++ b/lib/gro/rte_gro.c
@@ -327,7 +327,7 @@ rte_gro_reassemble_burst(struct rte_mbuf **pkts,
/* Flush all packets from the tables */
if (do_vxlan_tcp_gro) {
i += gro_vxlan_tcp4_tbl_timeout_flush(&vxlan_tcp_tbl,
-   0, pkts, nb_pkts);
+   0, &pkts[i], nb_pkts - i);
}
 
if (do_vxlan_udp_gro) {
-- 
2.34.1

[PATCH v7 07/21] rcu: remove use of VLAs for Windows built code

From: Konstantin Ananyev 

1) ./lib/rcu/rte_rcu_qsbr.c:359:9
: warning: ISO C90 forbids variable length array ‘data’ [-Wvla]
2) ./lib/rcu/rte_rcu_qsbr.c:422:9
: warning: ISO C90 forbids variable length array ‘data’ [-Wvla]

In both cases we allocate VLA for one element from RCU deferred queue.
Right now, size of element in RCU queue is not limited by API.
The approach is to introduce some reasonable limitation on RCU DQ
element size.
Choose 128B for now.
With that in place we can replace both VLA occurencies with fixed size
array.

Note that such change need to be treated as API change.
So can be applied only at 24.11.

Signed-off-by: Konstantin Ananyev 
---
 lib/rcu/rte_rcu_qsbr.c | 7 ---
 lib/rcu/rte_rcu_qsbr.h | 5 +
 2 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/lib/rcu/rte_rcu_qsbr.c b/lib/rcu/rte_rcu_qsbr.c
index 40d7c566c8..09a14a15f1 100644
--- a/lib/rcu/rte_rcu_qsbr.c
+++ b/lib/rcu/rte_rcu_qsbr.c
@@ -245,7 +245,8 @@ rte_rcu_qsbr_dq_create(const struct 
rte_rcu_qsbr_dq_parameters *params)
if (params == NULL || params->free_fn == NULL ||
params->v == NULL || params->name == NULL ||
params->size == 0 || params->esize == 0 ||
-   (params->esize % 4 != 0)) {
+   (params->esize % 4 != 0) ||
+   params->esize > RTE_QSBR_ESIZE_MAX) {
RCU_LOG(ERR, "Invalid input parameter");
rte_errno = EINVAL;
 
@@ -323,7 +324,7 @@ int rte_rcu_qsbr_dq_enqueue(struct rte_rcu_qsbr_dq *dq, 
void *e)
return 1;
}
 
-   char data[dq->esize];
+   char data[RTE_QSBR_ESIZE_MAX + __RTE_QSBR_TOKEN_SIZE];
dq_elem = (__rte_rcu_qsbr_dq_elem_t *)data;
/* Start the grace period */
dq_elem->token = rte_rcu_qsbr_start(dq->v);
@@ -386,7 +387,7 @@ rte_rcu_qsbr_dq_reclaim(struct rte_rcu_qsbr_dq *dq, 
unsigned int n,
 
cnt = 0;
 
-   char data[dq->esize];
+   char data[RTE_QSBR_ESIZE_MAX + __RTE_QSBR_TOKEN_SIZE];
/* Check reader threads quiescent state and reclaim resources */
while (cnt < n &&
rte_ring_dequeue_bulk_elem_start(dq->r, &data,
diff --git a/lib/rcu/rte_rcu_qsbr.h b/lib/rcu/rte_rcu_qsbr.h
index 550fadf56a..abcbd78914 100644
--- a/lib/rcu/rte_rcu_qsbr.h
+++ b/lib/rcu/rte_rcu_qsbr.h
@@ -86,6 +86,11 @@ struct __rte_cache_aligned rte_rcu_qsbr_cnt {
 #define __RTE_QSBR_CNT_MAX ((uint64_t)~0)
 #define __RTE_QSBR_TOKEN_SIZE sizeof(uint64_t)
 
+/**
+ * Max allowable size (in bytes) of each element in the defer queue
+ */
+#define RTE_QSBR_ESIZE_MAX (2 * RTE_CACHE_LINE_MIN_SIZE)
+
 /* RTE Quiescent State variable structure.
  * This structure has two elements that vary in size based on the
  * 'max_threads' parameter.
-- 
2.34.1

[PATCH v7 10/21] net/ixgbe: remove use of VLAs

From: Konstantin Ananyev 

1) ../drivers/net/ixgbe/ixgbe_ethdev.c:3556:46: warning:
variable length array used [-Wvla]
2) ../drivers/net/ixgbe/ixgbe_ethdev.c:3739:23: warning:
variable length array used [-Wvla]
3) ../drivers/net/ixgbe/ixgbe_rxtx_vec_common.h:17:24: warning:
variable length array used [-Wvla]

For first two cases: in fact ixgbe_xstats_calc_num() always returns
constant value, so it should be safe to replace that function invocation
just with a macro that performs same calculations.

For case #3: reassemble_packets() is invoked only by
ixgbe_recv_scattered_burst_vec().
And in turn, ixgbe_recv_scattered_burst_vec() operates only on fixed
max amount of packets per call: RTE_IXGBE_MAX_RX_BURST.
So, it should be safe to replace VLA with fixed size array.

Signed-off-by: Konstantin Ananyev 
Acked-by: Anatoly Burakov 
---
 drivers/net/ixgbe/ixgbe_ethdev.c  | 21 +
 drivers/net/ixgbe/ixgbe_rxtx_vec_common.h |  4 +++-
 2 files changed, 16 insertions(+), 9 deletions(-)

diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c b/drivers/net/ixgbe/ixgbe_ethdev.c
index 7da2ccf6a8..fea468140b 100644
--- a/drivers/net/ixgbe/ixgbe_ethdev.c
+++ b/drivers/net/ixgbe/ixgbe_ethdev.c
@@ -3435,11 +3435,16 @@ ixgbe_dev_stats_reset(struct rte_eth_dev *dev)
 }
 
 /* This function calculates the number of xstats based on the current config */
+
+#define IXGBE_XSTATS_CALC_NUM  \
+   (IXGBE_NB_HW_STATS + IXGBE_NB_MACSEC_STATS + \
+   (IXGBE_NB_RXQ_PRIO_STATS * IXGBE_NB_RXQ_PRIO_VALUES) + \
+   (IXGBE_NB_TXQ_PRIO_STATS * IXGBE_NB_TXQ_PRIO_VALUES))
+
 static unsigned
-ixgbe_xstats_calc_num(void) {
-   return IXGBE_NB_HW_STATS + IXGBE_NB_MACSEC_STATS +
-   (IXGBE_NB_RXQ_PRIO_STATS * IXGBE_NB_RXQ_PRIO_VALUES) +
-   (IXGBE_NB_TXQ_PRIO_STATS * IXGBE_NB_TXQ_PRIO_VALUES);
+ixgbe_xstats_calc_num(void)
+{
+   return IXGBE_XSTATS_CALC_NUM;
 }
 
 static int ixgbe_dev_xstats_get_names(__rte_unused struct rte_eth_dev *dev,
@@ -3555,8 +3560,8 @@ static int ixgbe_dev_xstats_get_names_by_id(
}
 
uint16_t i;
-   uint16_t size = ixgbe_xstats_calc_num();
-   struct rte_eth_xstat_name xstats_names_copy[size];
+   struct rte_eth_xstat_name xstats_names_copy[IXGBE_XSTATS_CALC_NUM];
+   const uint16_t size = RTE_DIM(xstats_names_copy);
 
ixgbe_dev_xstats_get_names_by_id(dev, NULL, xstats_names_copy,
size);
@@ -3738,8 +3743,8 @@ ixgbe_dev_xstats_get_by_id(struct rte_eth_dev *dev, const 
uint64_t *ids,
}
 
uint16_t i;
-   uint16_t size = ixgbe_xstats_calc_num();
-   uint64_t values_copy[size];
+   uint64_t values_copy[IXGBE_XSTATS_CALC_NUM];
+   const uint16_t size = RTE_DIM(values_copy);
 
ixgbe_dev_xstats_get_by_id(dev, NULL, values_copy, size);
 
diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_common.h 
b/drivers/net/ixgbe/ixgbe_rxtx_vec_common.h
index a4d9ec9b08..c1cf0a581a 100644
--- a/drivers/net/ixgbe/ixgbe_rxtx_vec_common.h
+++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_common.h
@@ -14,11 +14,13 @@ static inline uint16_t
 reassemble_packets(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_bufs,
   uint16_t nb_bufs, uint8_t *split_flags)
 {
-   struct rte_mbuf *pkts[nb_bufs]; /*finished pkts*/
+   struct rte_mbuf *pkts[RTE_IXGBE_MAX_RX_BURST]; /*finished pkts*/
struct rte_mbuf *start = rxq->pkt_first_seg;
struct rte_mbuf *end =  rxq->pkt_last_seg;
unsigned int pkt_idx, buf_idx;
 
+   RTE_ASSERT(nb_bufs <= RTE_DIM(pkts));
+
for (buf_idx = 0, pkt_idx = 0; buf_idx < nb_bufs; buf_idx++) {
if (end != NULL) {
/* processing a split packet */
-- 
2.34.1

[PATCH v7 17/21] net/mlx5: remove use of VLAs for Windows built code

From: Tyler Retzlaff 

MSVC does not support VLAs, replace VLAs with standard C arrays
or alloca(). alloca() is available for all toolchain/platform
combinations officially supported by DPDK.

Signed-off-by: Tyler Retzlaff 
---
 drivers/net/mlx5/mlx5.c  | 5 ++---
 drivers/net/mlx5/mlx5_flow.c | 6 +++---
 2 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 52b90e6ff3..ab6fe6d753 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -1598,14 +1598,13 @@ void
 mlx5_rt_timestamp_config(struct mlx5_dev_ctx_shared *sh,
 struct mlx5_hca_attr *hca_attr)
 {
-   uint32_t dw_cnt = MLX5_ST_SZ_DW(register_mtutc);
-   uint32_t reg[dw_cnt];
+   uint32_t reg[MLX5_ST_SZ_DW(register_mtutc)];
int ret = ENOTSUP;
 
if (hca_attr->access_register_user)
ret = mlx5_devx_cmd_register_read(sh->cdev->ctx,
  MLX5_REGISTER_ID_MTUTC, 0,
- reg, dw_cnt);
+ reg, RTE_DIM(reg));
if (!ret) {
uint32_t ts_mode;
 
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 9c43201e05..ab58fa875e 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -1479,8 +1479,8 @@ mlx5_flow_item_acceptable(const struct rte_eth_dev *dev,
  "mask/last without a spec is not"
  " supported");
if (item->spec && item->last && !range_accepted) {
-   uint8_t spec[size];
-   uint8_t last[size];
+   uint8_t *spec = alloca(size);
+   uint8_t *last = alloca(size);
unsigned int i;
int ret;
 
@@ -8474,7 +8474,7 @@ mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
.type = RTE_FLOW_ITEM_TYPE_END,
},
};
-   uint16_t queue[priv->reta_idx_n];
+   uint16_t *queue = alloca(sizeof(uint16_t) * priv->reta_idx_n);
struct rte_flow_action_rss action_rss = {
.func = RTE_ETH_HASH_FUNCTION_DEFAULT,
.level = 0,
-- 
2.34.1

[PATCH v7 09/21] gro: remove use of VLAs

From: Konstantin Ananyev 

../lib/gro/rte_gro.c:182:34: warning:
variable length array used [-Wvla]
../lib/gro/rte_gro.c:363:34: warning:
variable length array used [-Wvla]

In both cases the pattern is the same: we use unprocess_pkts[nb_pkts] to
collect un-used by GRO packets, and then copy them to the start of
input/output pkts[] array.
In both cases, we can safely copy pkts[i] into already
processed entry at the same array, i.e. into pkts[unprocess_num].
Such change eliminates need of temporary VLA: unprocess_pkts[nb_pkts].

Signed-off-by: Konstantin Ananyev 
Acked-by: Ferruh Yigit 
---
 lib/gro/rte_gro.c | 40 ++--
 1 file changed, 14 insertions(+), 26 deletions(-)

diff --git a/lib/gro/rte_gro.c b/lib/gro/rte_gro.c
index db86117609..6d5aadf32a 100644
--- a/lib/gro/rte_gro.c
+++ b/lib/gro/rte_gro.c
@@ -179,7 +179,6 @@ rte_gro_reassemble_burst(struct rte_mbuf **pkts,
struct gro_vxlan_udp4_item vxlan_udp_items[RTE_GRO_MAX_BURST_ITEM_NUM]
= {{{0}} };
 
-   struct rte_mbuf *unprocess_pkts[nb_pkts];
uint32_t item_num;
int32_t ret;
uint16_t i, unprocess_num = 0, nb_after_gro = nb_pkts;
@@ -275,7 +274,7 @@ rte_gro_reassemble_burst(struct rte_mbuf **pkts,
/* Merge successfully */
nb_after_gro--;
else if (ret < 0)
-   unprocess_pkts[unprocess_num++] = pkts[i];
+   pkts[unprocess_num++] = pkts[i];
} else if (IS_IPV4_VXLAN_UDP4_PKT(pkts[i]->packet_type) &&
do_vxlan_udp_gro) {
ret = gro_vxlan_udp4_reassemble(pkts[i],
@@ -284,7 +283,7 @@ rte_gro_reassemble_burst(struct rte_mbuf **pkts,
/* Merge successfully */
nb_after_gro--;
else if (ret < 0)
-   unprocess_pkts[unprocess_num++] = pkts[i];
+   pkts[unprocess_num++] = pkts[i];
} else if (IS_IPV4_TCP_PKT(pkts[i]->packet_type) &&
do_tcp4_gro) {
ret = gro_tcp4_reassemble(pkts[i], &tcp_tbl, 0);
@@ -292,7 +291,7 @@ rte_gro_reassemble_burst(struct rte_mbuf **pkts,
/* merge successfully */
nb_after_gro--;
else if (ret < 0)
-   unprocess_pkts[unprocess_num++] = pkts[i];
+   pkts[unprocess_num++] = pkts[i];
} else if (IS_IPV4_UDP_PKT(pkts[i]->packet_type) &&
do_udp4_gro) {
ret = gro_udp4_reassemble(pkts[i], &udp_tbl, 0);
@@ -300,7 +299,7 @@ rte_gro_reassemble_burst(struct rte_mbuf **pkts,
/* merge successfully */
nb_after_gro--;
else if (ret < 0)
-   unprocess_pkts[unprocess_num++] = pkts[i];
+   pkts[unprocess_num++] = pkts[i];
} else if (IS_IPV6_TCP_PKT(pkts[i]->packet_type) &&
do_tcp6_gro) {
ret = gro_tcp6_reassemble(pkts[i], &tcp6_tbl, 0);
@@ -308,21 +307,15 @@ rte_gro_reassemble_burst(struct rte_mbuf **pkts,
/* merge successfully */
nb_after_gro--;
else if (ret < 0)
-   unprocess_pkts[unprocess_num++] = pkts[i];
+   pkts[unprocess_num++] = pkts[i];
} else
-   unprocess_pkts[unprocess_num++] = pkts[i];
+   pkts[unprocess_num++] = pkts[i];
}
 
if ((nb_after_gro < nb_pkts)
 || (unprocess_num < nb_pkts)) {
-   i = 0;
-   /* Copy unprocessed packets */
-   if (unprocess_num > 0) {
-   memcpy(&pkts[i], unprocess_pkts,
-   sizeof(struct rte_mbuf *) *
-   unprocess_num);
-   i = unprocess_num;
-   }
+
+   i = unprocess_num;
 
/* Flush all packets from the tables */
if (do_vxlan_tcp_gro) {
@@ -360,7 +353,6 @@ rte_gro_reassemble(struct rte_mbuf **pkts,
uint16_t nb_pkts,
void *ctx)
 {
-   struct rte_mbuf *unprocess_pkts[nb_pkts];
struct gro_ctx *gro_ctx = ctx;
void *tcp_tbl, *udp_tbl, *vxlan_tcp_tbl, *vxlan_udp_tbl, *tcp6_tbl;
uint64_t current_time;
@@ -396,33 +388,29 @@ rte_gro_reassemble(struct rte_mbuf **pkts,
do_vxlan_tcp_gro) {
if (gro_vxlan_tcp4_reassemble(pkts[i], vxlan_tcp_tbl,

[PATCH v7 12/21] app/testpmd: remove use of VLAs for Windows built

From: Tyler Retzlaff 

MSVC does not support VLAs, replace VLAs with standard C arrays
or alloca(). alloca() is available for all toolchain/platform
combinations officially supported by DPDK.

Signed-off-by: Tyler Retzlaff 
---
 app/test-pmd/cmdline.c  |  2 +-
 app/test-pmd/cmdline_flow.c | 15 ++-
 app/test-pmd/config.c   | 16 +---
 3 files changed, 20 insertions(+), 13 deletions(-)

diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index 7e0666e9f6..2897e44c34 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -13274,7 +13274,7 @@ cmd_set_port_ptypes_parsed(
return;
}
 
-   uint32_t ptypes[ret];
+   uint32_t *ptypes = alloca(sizeof(uint32_t) * ret);
 
ret = rte_eth_dev_set_ptypes(port_id, ptype_mask, ptypes, ret);
if (ret < 0) {
diff --git a/app/test-pmd/cmdline_flow.c b/app/test-pmd/cmdline_flow.c
index 1e4f2ebc55..fb0c4f838d 100644
--- a/app/test-pmd/cmdline_flow.c
+++ b/app/test-pmd/cmdline_flow.c
@@ -11707,8 +11707,7 @@ parse_hex(struct context *ctx, const struct token 
*token,
char tmp[16]; /* Ought to be enough. */
int ret;
unsigned int hexlen = len;
-   unsigned int length = 256;
-   uint8_t hex_tmp[length];
+   uint8_t hex_tmp[256];
 
/* Arguments are expected. */
if (!arg_data)
@@ -11735,7 +11734,7 @@ parse_hex(struct context *ctx, const struct token 
*token,
str += 2;
hexlen -= 2;
}
-   if (hexlen > length)
+   if (hexlen > RTE_DIM(hex_tmp))
goto error;
ret = parse_hex_string(str, hex_tmp, &hexlen);
if (ret < 0)
@@ -11868,10 +11867,13 @@ parse_ipv4_addr(struct context *ctx, const struct 
token *token,
void *buf, unsigned int size)
 {
const struct arg *arg = pop_args(ctx);
-   char str2[len + 1];
+   char str2[INET_ADDRSTRLEN];
struct in_addr tmp;
int ret;
 
+   /* Length is longer than the max length an IPv4 address can have. */
+   if (len >= INET_ADDRSTRLEN)
+   return -1;
/* Argument is expected. */
if (!arg)
return -1;
@@ -11914,11 +11916,14 @@ parse_ipv6_addr(struct context *ctx, const struct 
token *token,
void *buf, unsigned int size)
 {
const struct arg *arg = pop_args(ctx);
-   char str2[len + 1];
+   char str2[INET6_ADDRSTRLEN];
struct rte_ipv6_addr tmp;
int ret;
 
(void)token;
+   /* Length is longer than the max length an IPv6 address can have. */
+   if (len >= INET6_ADDRSTRLEN)
+   return -1;
/* Argument is expected. */
if (!arg)
return -1;
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index 88770b4dfc..9ff2b8817f 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -1802,7 +1802,8 @@ port_flow_configure(portid_t port_id,
 {
struct rte_port *port;
struct rte_flow_error error;
-   const struct rte_flow_queue_attr *attr_list[nb_queue];
+   const struct rte_flow_queue_attr **attr_list =
+   alloca(sizeof(struct rte_flow_queue_attr *) * nb_queue);
int std_queue;
 
if (port_id_is_invalid(port_id, ENABLED_WARN) ||
@@ -2616,10 +2617,10 @@ port_flow_template_table_create(portid_t port_id, 
uint32_t id,
int ret;
uint32_t i;
struct rte_flow_error error;
-   struct rte_flow_pattern_template
-   *flow_pattern_templates[nb_pattern_templates];
-   struct rte_flow_actions_template
-   *flow_actions_templates[nb_actions_templates];
+   struct rte_flow_pattern_template **flow_pattern_templates =
+   alloca(sizeof(struct rte_flow_pattern_template *) * 
nb_pattern_templates);
+   struct rte_flow_actions_template **flow_actions_templates =
+   alloca(sizeof(struct rte_flow_actions_template *) * 
nb_actions_templates);
 
if (port_id_is_invalid(port_id, ENABLED_WARN) ||
port_id == (portid_t)RTE_PORT_ALL)
@@ -5529,7 +5530,7 @@ parse_port_list(const char *list, unsigned int *values, 
unsigned int maxsize)
char *end = NULL;
int min, max;
int value, i;
-   unsigned int marked[maxsize];
+   unsigned int *marked = alloca(sizeof(unsigned int) * maxsize);
 
if (list == NULL || values == NULL)
return 0;
@@ -7270,7 +7271,8 @@ show_macs(portid_t port_id)
if (eth_dev_info_get_print_err(port_id, &dev_info))
return;
 
-   struct rte_ether_addr addr[dev_info.max_mac_addrs];
+   struct rte_ether_addr *addr =
+   alloca(sizeof(struct rte_ether_addr) * dev_info.max_mac_addrs);
rc = rte_eth_macaddrs_get(port_id, addr, dev_info.max_mac_addrs);
if (rc < 0)
return;
-- 
2.34.1

[PATCH v7 14/21] common/idpf: remove use of VLAs for Windows built code

From: Tyler Retzlaff 

MSVC does not support VLAs, replace VLAs with standard C arrays
or alloca(). alloca() is available for all toolchain/platform
combinations officially supported by DPDK.

Signed-off-by: Tyler Retzlaff 
Acked-by: Bruce Richardson 
---
 drivers/common/idpf/idpf_common_rxtx.c| 2 +-
 drivers/common/idpf/idpf_common_rxtx_avx512.c | 6 --
 2 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/common/idpf/idpf_common_rxtx.c 
b/drivers/common/idpf/idpf_common_rxtx.c
index a04e54ce26..e04ab40fa2 100644
--- a/drivers/common/idpf/idpf_common_rxtx.c
+++ b/drivers/common/idpf/idpf_common_rxtx.c
@@ -569,7 +569,7 @@ idpf_split_rx_bufq_refill(struct idpf_rx_queue *rx_bufq)
uint16_t nb_refill = rx_bufq->rx_free_thresh;
uint16_t nb_desc = rx_bufq->nb_rx_desc;
uint16_t next_avail = rx_bufq->rx_tail;
-   struct rte_mbuf *nmb[rx_bufq->rx_free_thresh];
+   struct rte_mbuf **nmb = alloca(sizeof(struct rte_mbuf *) * 
rx_bufq->rx_free_thresh);
uint64_t dma_addr;
uint16_t delta;
int i;
diff --git a/drivers/common/idpf/idpf_common_rxtx_avx512.c 
b/drivers/common/idpf/idpf_common_rxtx_avx512.c
index b8450b03ae..63e10c542f 100644
--- a/drivers/common/idpf/idpf_common_rxtx_avx512.c
+++ b/drivers/common/idpf/idpf_common_rxtx_avx512.c
@@ -1002,7 +1002,8 @@ idpf_tx_singleq_free_bufs_avx512(struct idpf_tx_queue 
*txq)
uint32_t n;
uint32_t i;
int nb_free = 0;
-   struct rte_mbuf *m, *free[txq->rs_thresh];
+   struct rte_mbuf *m;
+   struct rte_mbuf **free = alloca(sizeof(struct rte_mbuf *) * 
txq->rs_thresh);
 
/* check DD bits on threshold descriptor */
if ((txq->tx_ring[txq->next_dd].qw1 &
@@ -1326,7 +1327,8 @@ idpf_tx_splitq_free_bufs_avx512(struct idpf_tx_queue *txq)
uint32_t n;
uint32_t i;
int nb_free = 0;
-   struct rte_mbuf *m, *free[txq->rs_thresh];
+   struct rte_mbuf *m;
+   struct rte_mbuf **free = alloca(sizeof(struct rte_mbuf *) * 
txq->rs_thresh);
 
n = txq->rs_thresh;
 
-- 
2.34.1

[PATCH v7 13/21] test: remove use of VLAs for Windows built code

From: Tyler Retzlaff 

MSVC does not support VLAs, replace VLAs with standard C arrays
or alloca(). alloca() is available for all toolchain/platform
combinations officially supported by DPDK.

Signed-off-by: Tyler Retzlaff 
---
 app/test/test.c   |  2 +-
 app/test/test_cmdline_string.c|  2 +-
 app/test/test_cryptodev.c | 34 +--
 app/test/test_cryptodev_blockcipher.c |  4 +--
 app/test/test_cryptodev_crosscheck.c  |  2 +-
 app/test/test_dmadev.c|  9 +++--
 app/test/test_hash.c  | 14 
 app/test/test_mempool.c   | 25 +++---
 app/test/test_reorder.c   | 48 +++
 app/test/test_service_cores.c |  9 +++--
 app/test/test_thash.c |  7 ++--
 11 files changed, 82 insertions(+), 74 deletions(-)

diff --git a/app/test/test.c b/app/test/test.c
index 680351f6a3..fd653cbbfd 100644
--- a/app/test/test.c
+++ b/app/test/test.c
@@ -105,7 +105,7 @@ int
 main(int argc, char **argv)
 {
struct cmdline *cl;
-   char *tests[argc]; /* store an array of tests to run */
+   char **tests = alloca(sizeof(char *) * argc); /* store an array of 
tests to run */
int test_count = 0;
int i;
char *extra_args;
diff --git a/app/test/test_cmdline_string.c b/app/test/test_cmdline_string.c
index 97516c9400..e1cf86020f 100644
--- a/app/test/test_cmdline_string.c
+++ b/app/test/test_cmdline_string.c
@@ -40,7 +40,7 @@ struct string_elt_str string_elt_strs[] = {
 #if (CMDLINE_TEST_BUFSIZE < STR_TOKEN_SIZE) \
 || (CMDLINE_TEST_BUFSIZE < STR_MULTI_TOKEN_SIZE)
 #undef CMDLINE_TEST_BUFSIZE
-#define CMDLINE_TEST_BUFSIZE RTE_MAX(STR_TOKEN_SIZE, STR_MULTI_TOKEN_SIZE)
+#define CMDLINE_TEST_BUFSIZE RTE_MAX_T(STR_TOKEN_SIZE, STR_MULTI_TOKEN_SIZE, 
size_t)
 #endif
 
 struct string_nb_str {
diff --git a/app/test/test_cryptodev.c b/app/test/test_cryptodev.c
index 25eef342b0..888a3bfa21 100644
--- a/app/test/test_cryptodev.c
+++ b/app/test/test_cryptodev.c
@@ -2885,7 +2885,7 @@ create_wireless_algo_hash_session(uint8_t dev_id,
enum rte_crypto_auth_operation op,
enum rte_crypto_auth_algorithm algo)
 {
-   uint8_t hash_key[key_len];
+   uint8_t *hash_key = alloca(key_len);
 
struct crypto_testsuite_params *ts_params = &testsuite_params;
struct crypto_unittest_params *ut_params = &unittest_params;
@@ -2921,7 +2921,7 @@ create_wireless_algo_cipher_session(uint8_t dev_id,
const uint8_t *key, const uint8_t key_len,
uint8_t iv_len)
 {
-   uint8_t cipher_key[key_len];
+   uint8_t *cipher_key = alloca(key_len);
struct crypto_testsuite_params *ts_params = &testsuite_params;
struct crypto_unittest_params *ut_params = &unittest_params;
 
@@ -3073,7 +3073,7 @@ create_wireless_cipher_auth_session(uint8_t dev_id,
const struct wireless_test_data *tdata)
 {
const uint8_t key_len = tdata->key.len;
-   uint8_t cipher_auth_key[key_len];
+   uint8_t *cipher_auth_key = alloca(key_len);
 
struct crypto_testsuite_params *ts_params = &testsuite_params;
struct crypto_unittest_params *ut_params = &unittest_params;
@@ -9077,7 +9077,7 @@ create_aead_session(uint8_t dev_id, enum 
rte_crypto_aead_algorithm algo,
const uint16_t aad_len, const uint8_t auth_len,
uint8_t iv_len)
 {
-   uint8_t aead_key[key_len];
+   uint8_t *aead_key = alloca(key_len);
 
struct crypto_testsuite_params *ts_params = &testsuite_params;
struct crypto_unittest_params *ut_params = &unittest_params;
@@ -12991,7 +12991,7 @@ test_cryptodev_error_recover_helper(uint8_t dev_id, 
const void *test_data, bool
struct crypto_testsuite_params *ts_params = &testsuite_params;
struct crypto_unittest_params *ut_params = &unittest_params;
const struct blockcipher_test_data *tdata = test_data;
-   uint8_t cipher_key[tdata->cipher_key.len];
+   uint8_t *cipher_key = alloca(tdata->cipher_key.len);
struct rte_crypto_sym_op *sym_op = NULL;
struct rte_crypto_op *op = NULL;
char *dst;
@@ -13345,7 +13345,7 @@ static int
 test_AES_GCM_auth_encryption_fail_aad_corrupt(void)
 {
struct aead_test_data tdata;
-   uint8_t aad[gcm_test_case_7.aad.len];
+   uint8_t *aad = alloca(gcm_test_case_7.aad.len);
int res;
 
RTE_LOG(INFO, USER1, "This is a negative test, errors are expected\n");
@@ -13734,7 +13734,7 @@ static int
 test_AES_GCM_auth_decryption_fail_aad_corrupt(void)
 {
struct aead_test_data tdata;
-   uint8_t aad[gcm_test_case_7.aad.len];
+   uint8_t *aad = alloca(gcm_test_case_7.aad.len);
int res;
 
memcpy(&tdata, &gcm_test_case_7, sizeof(struct aead_test_data));
@@ -13986,7 +13986,7 @@ test_authenticated_encryption_sessionless(
int retval;
uint8_t *ciphertext, *auth_tag;
uint16_t pla

[PATCH v7 19/21] test: remove use of VLAs for Windows built code in bitset tests

1) MSVC does not support VLAs. Use standard fixed C arrays of
maximum size required instead.

2) ../usr/lib/gcc/x86_64-redhat-linux/13/include/emmintrin.h:742:8:
error: array subscript 9 is outside array bounds of 'uint64_t[16]'
{aka 'long unsigned int[16]'} [-Werror=array-bounds=]
   3695 742 | *__P = __B;

Compile with -Wno-array-bounds to avoid false positives when
using gcc version 11 or newer (gcc compiler bug/limitation).

Signed-off-by: Andre Muezerie 
---
 app/test/meson.build|  5 +++
 app/test/test_bitset.c  | 55 +++--
 app/test/test_lcore_var_perf.c  |  2 +-
 app/test/test_reassembly_perf.c |  4 +--
 4 files changed, 39 insertions(+), 27 deletions(-)

diff --git a/app/test/meson.build b/app/test/meson.build
index 40f22a54d5..e7419a611b 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -249,6 +249,11 @@ foreach d:optional_deps
 endif
 endforeach
 
+# Bugzilla ID: 678
+if (toolchain == 'gcc' and cc.version().version_compare('>=11.0.0'))
+cflags += '-Wno-array-bounds'
+endif
+
 if cc.has_argument('-Wno-format-truncation')
 cflags += '-Wno-format-truncation'
 endif
diff --git a/app/test/test_bitset.c b/app/test/test_bitset.c
index 50b8bf0da4..f58c491d8f 100644
--- a/app/test/test_bitset.c
+++ b/app/test/test_bitset.c
@@ -99,11 +99,13 @@ typedef void clear_fun(uint64_t *bitset, size_t bit_num);
 typedef void assign_fun(uint64_t *bitset, size_t bit_num, bool value);
 typedef void flip_fun(uint64_t *bitset, size_t bit_num);
 
+#define RAND_SET_MAX_SIZE (1000)
+
 static int
 test_set_clear_size(test_fun test_fun, set_fun set_fun, clear_fun clear_fun, 
size_t size)
 {
size_t i;
-   bool reference[size];
+   bool reference[RAND_SET_MAX_SIZE];
uint64_t *bitset;
 
rand_bool_ary(reference, size);
@@ -132,7 +134,6 @@ test_set_clear_size(test_fun test_fun, set_fun set_fun, 
clear_fun clear_fun, siz
 }
 
 #define RAND_ITERATIONS (1)
-#define RAND_SET_MAX_SIZE (1000)
 
 static int
 test_set_clear_fun(test_fun test_fun, set_fun set_fun, clear_fun clear_fun)
@@ -168,7 +169,7 @@ test_flip_size(test_fun test_fun, assign_fun assign_fun, 
flip_fun flip_fun, size
rand_bitset(bitset, size);
 
for (i = 0; i < size; i++) {
-   RTE_BITSET_DECLARE(reference, size);
+   RTE_BITSET_DECLARE(reference, RAND_SET_MAX_SIZE);
 
rte_bitset_copy(reference, bitset, size);
 
@@ -288,7 +289,7 @@ static int
 test_find_size(size_t size, bool set)
 {
uint64_t *bitset;
-   bool reference[size];
+   bool reference[RAND_SET_MAX_SIZE];
size_t i;
 
bitset = alloc_bitset(size);
@@ -388,8 +389,8 @@ record_match(ssize_t match_idx, size_t size, int *calls)
 static int
 test_foreach_size(ssize_t size, bool may_wrap, bool set)
 {
-   bool reference[size];
-   int calls[size];
+   bool reference[RAND_SET_MAX_SIZE];
+   int calls[RAND_SET_MAX_SIZE];
uint64_t *bitset;
ssize_t i;
ssize_t start_bit;
@@ -633,17 +634,19 @@ test_define(void)
 typedef void bitset_op(uint64_t *dst, const uint64_t *a, const uint64_t *b, 
size_t bit_num);
 typedef bool bool_op(bool a, bool b);
 
+#define LOGIC_MAX_SET_SIZE 200
+
 static int
 test_logic_op(bitset_op bitset_op, bool_op bool_op)
 {
-   const size_t size = 1 + rte_rand_max(200);
-   RTE_BITSET_DECLARE(bitset_a, size);
-   RTE_BITSET_DECLARE(bitset_b, size);
-   RTE_BITSET_DECLARE(bitset_d, size);
+   const size_t size = 1 + rte_rand_max(LOGIC_MAX_SET_SIZE);
+   RTE_BITSET_DECLARE(bitset_a, LOGIC_MAX_SET_SIZE);
+   RTE_BITSET_DECLARE(bitset_b, LOGIC_MAX_SET_SIZE);
+   RTE_BITSET_DECLARE(bitset_d, LOGIC_MAX_SET_SIZE);
 
-   bool ary_a[size];
-   bool ary_b[size];
-   bool ary_d[size];
+   bool ary_a[LOGIC_MAX_SET_SIZE];
+   bool ary_b[LOGIC_MAX_SET_SIZE];
+   bool ary_d[LOGIC_MAX_SET_SIZE];
 
rand_bool_ary(ary_a, size);
rand_bool_ary(ary_b, size);
@@ -708,14 +711,14 @@ test_complement(void)
for (i = 0; i < RAND_ITERATIONS; i++) {
const size_t size = 1 + rte_rand_max(RAND_SET_MAX_SIZE - 1);
 
-   RTE_BITSET_DECLARE(src, size);
+   RTE_BITSET_DECLARE(src, RAND_SET_MAX_SIZE);
 
rand_bitset(src, size);
 
bool bit_idx = rte_rand_max(size);
bool bit_value = rte_bitset_test(src, bit_idx);
 
-   RTE_BITSET_DECLARE(dst, size);
+   RTE_BITSET_DECLARE(dst, RAND_SET_MAX_SIZE);
 
rte_bitset_complement(dst, src, size);
 
@@ -726,6 +729,8 @@ test_complement(void)
return TEST_SUCCESS;
 }
 
+#define SHIFT_SET_MAX_SIZE 500
+
 static int
 test_shift(bool right)
 {
@@ -734,12 +739,12 @@ test_shift(bool right)
const char *direction = right ? "right" : "left";
 
for (i = 0; i < 1; i++) {
-   const int size = 1 + (int)rte_rand_max(500);
+   cons

[PATCH v7 21/21] hash: remove use of VLAs by using standard arrays

MSVC does not support VLAs, replace VLAs with standard C arrays.

Signed-off-by: Andre Muezerie 
---
 lib/hash/rte_thash_gf2_poly_math.c | 7 ---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/lib/hash/rte_thash_gf2_poly_math.c 
b/lib/hash/rte_thash_gf2_poly_math.c
index 1c62974e71..cf7c7d396c 100644
--- a/lib/hash/rte_thash_gf2_poly_math.c
+++ b/lib/hash/rte_thash_gf2_poly_math.c
@@ -8,6 +8,7 @@
 #include 
 #include 
 
+#define MAX_POLY_DEGREE 32
 #define MAX_TOEPLITZ_KEY_LENGTH 64
 RTE_LOG_REGISTER_SUFFIX(thash_poly_logtype, thash_poly, INFO);
 #define RTE_LOGTYPE_HASH thash_poly_logtype
@@ -149,7 +150,7 @@ gf2_pow(uint32_t a, uint32_t pow, uint32_t r, int degree)
 static uint32_t
 __thash_get_rand_poly(int poly_degree)
 {
-   uint32_t roots[poly_degree];
+   uint32_t roots[MAX_POLY_DEGREE];
uint32_t rnd;
uint32_t ret_poly = 0;
int i, j;
@@ -194,7 +195,7 @@ __thash_get_rand_poly(int poly_degree)
 * Get coefficients of the polynomial for
 * (x - roots[0])(x - roots[1])...(x - roots[n])
 */
-   uint32_t poly_coefficients[poly_degree + 1];
+   uint32_t poly_coefficients[MAX_POLY_DEGREE + 1];
for (i = 0; i <= poly_degree; i++)
poly_coefficients[i] = 0;
 
@@ -247,7 +248,7 @@ thash_get_rand_poly(uint32_t poly_degree)
 {
uint32_t ret_poly;
 
-   if (poly_degree > 32) {
+   if (poly_degree > MAX_POLY_DEGREE) {
HASH_LOG(ERR, "Wrong polynomial degree %d, must be in range [1, 
32]", poly_degree);
return 0;
}
-- 
2.34.1

[PATCH v7 15/21] net/i40e: remove use of VLAs for Windows built code

From: Tyler Retzlaff 

MSVC does not support VLAs, replace VLAs with standard C arrays
or alloca(). alloca() is available for all toolchain/platform
combinations officially supported by DPDK.

Signed-off-by: Tyler Retzlaff 
Reviewed-by: Bruce Richardson 
---
 drivers/net/i40e/i40e_testpmd.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/net/i40e/i40e_testpmd.c b/drivers/net/i40e/i40e_testpmd.c
index b6ef5d6e42..21f596297b 100644
--- a/drivers/net/i40e/i40e_testpmd.c
+++ b/drivers/net/i40e/i40e_testpmd.c
@@ -2168,8 +2168,7 @@ cmd_ptype_mapping_get_parsed(void *parsed_result,
 {
struct cmd_ptype_mapping_get_result *res = parsed_result;
int ret = -ENOTSUP;
-   int max_ptype_num = 256;
-   struct rte_pmd_i40e_ptype_mapping mapping[max_ptype_num];
+   struct rte_pmd_i40e_ptype_mapping mapping[256];
uint16_t count;
int i;
 
@@ -2178,7 +2177,7 @@ cmd_ptype_mapping_get_parsed(void *parsed_result,
 
ret = rte_pmd_i40e_ptype_mapping_get(res->port_id,
mapping,
-   max_ptype_num,
+   RTE_DIM(mapping),
&count,
res->valid_only);
switch (ret) {
-- 
2.34.1

[PATCH v7 20/21] app/testpmd: remove use of VLAs for Windows built code in shared_rxq_fwd

MSVC does not support VLAs, replace VLAs with standard C arrays
or alloca(). alloca() is available for all toolchain/platform
combinations officially supported by DPDK.

Signed-off-by: Andre Muezerie 
---
 app/test-pmd/shared_rxq_fwd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/app/test-pmd/shared_rxq_fwd.c b/app/test-pmd/shared_rxq_fwd.c
index 623d62da88..b85830b90e 100644
--- a/app/test-pmd/shared_rxq_fwd.c
+++ b/app/test-pmd/shared_rxq_fwd.c
@@ -92,7 +92,7 @@ forward_shared_rxq(struct fwd_stream *fs, uint16_t nb_rx,
 static bool
 shared_rxq_fwd(struct fwd_stream *fs)
 {
-   struct rte_mbuf *pkts_burst[nb_pkt_per_burst];
+   struct rte_mbuf *pkts_burst[MAX_PKT_BURST];
uint16_t nb_rx;
 
nb_rx = common_fwd_stream_receive(fs, pkts_burst, nb_pkt_per_burst);
-- 
2.34.1

[PATCH v7 18/21] build: enable vla warnings on Windows built code

From: Tyler Retzlaff 

MSVC does not support optional C11 VLAs. When building for Windows
enable -Wvla so that mingw and clang also fail if a VLA is used.

Signed-off-by: Tyler Retzlaff 
---
 config/meson.build | 4 
 1 file changed, 4 insertions(+)

diff --git a/config/meson.build b/config/meson.build
index 6aaad6d8a4..ebca19b4e5 100644
--- a/config/meson.build
+++ b/config/meson.build
@@ -342,6 +342,10 @@ if cc.get_id() == 'intel'
 warning_flags += '-diag-disable=@0@'.format(i)
 endforeach
 endif
+# no VLAs in code built on Windows
+if is_windows
+warning_flags += '-Wvla'
+endif
 foreach arg: warning_flags
 if cc.has_argument(arg)
 add_project_arguments(arg, language: 'c')
-- 
2.34.1

Re: [RFC] ip_frag: support IPv6 reassembly with extensions

14/02/2024 04:51, Stephen Hemminger:
> On Tue, 13 Feb 2024 12:47:27 +0100
>  wrote:
> 
> > +/*
> > + * Function to crawl through the extension header stack.
> > + * This function breaks as soon a the fragment header is
> > + * found and returns the total length the traversed exts
> > + * and the last extension before the fragment header
> > + */
> > +static inline uint32_t
> > +ip_frag_get_last_exthdr(struct rte_ipv6_hdr *ip_hdr, uint8_t **last_ext)
> > +{
> > +   uint32_t total_len = 0;
> > +   size_t ext_len = 0;
> > +   *last_ext = (uint8_t *)(ip_hdr + 1);
> > +   int next_proto = ip_hdr->proto;
> > +
> > +   while (next_proto != IPPROTO_FRAGMENT &&
> > +   (next_proto = rte_ipv6_get_next_ext(
> > +   *last_ext, next_proto, &ext_len)) >= 0) {
> > +
> > +   total_len += ext_len;
> > +
> > +   if (next_proto == IPPROTO_FRAGMENT)
> > +   return total_len;
> > +
> > +   *last_ext += ext_len;
> > +   }
> > +
> > +   return total_len;
> > +}
> 
> Doing endless loop like this opens up DoS attacks.
> Better to use rte_next_skip_ip6_ext() or do similar limited loop.

There was no reply to this interesting comment?

Re: [PATCH v2 1/3] graph: avoid global node ID counter

2024-11-11 Thread Robin Jarry


, Nov 11, 2024 at 08:34:

From: Kiran Kumar K 

The node id is determined based on a global variable that is
incremented every time a node is created. Adding changes to
remove the global counter. Make sure that the node list is always
ordered by increasing node ids. When creating a new node, pick a free
id which is not allocated.

Signed-off-by: Kiran Kumar K 


Hi Kiran,

Thanks for the fix.

I think most of the accesses to the node_list is racy. Maybe we need to 
add a lock to protect that linked list from concurrent access?



---
 lib/graph/graph_private.h |  8 
 lib/graph/node.c  | 81 +--
 2 files changed, 69 insertions(+), 20 deletions(-)

diff --git a/lib/graph/graph_private.h b/lib/graph/graph_private.h
index da48d73587..fdaf5649b8 100644
--- a/lib/graph/graph_private.h
+++ b/lib/graph/graph_private.h
@@ -29,14 +29,6 @@ extern int rte_graph_logtype;
 #define graph_info(...) GRAPH_LOG(INFO, __VA_ARGS__)
 #define graph_dbg(...) GRAPH_LOG(DEBUG, __VA_ARGS__)
 
-#define ID_CHECK(id, id_max)   \

-   do {   \
-   if ((id) >= (id_max)) {\
-   rte_errno = EINVAL;\
-   goto fail; \
-   }  \
-   } while (0)
-
 #define SET_ERR_JMP(err, where, fmt, ...)  
\
do {   \
graph_err(fmt, ##__VA_ARGS__); \
diff --git a/lib/graph/node.c b/lib/graph/node.c
index 63db629da8..0f382d744c 100644
--- a/lib/graph/node.c
+++ b/lib/graph/node.c
@@ -15,9 +15,51 @@
 #include "graph_private.h"
 
 static struct node_head node_list = STAILQ_HEAD_INITIALIZER(node_list);

-static rte_node_t node_id;
 
-#define NODE_ID_CHECK(id) ID_CHECK(id, node_id)

+static struct node *
+node_from_id(rte_node_t id)
+{
+   struct node *node;
+
+   STAILQ_FOREACH(node, &node_list, next) {
+   if (node->id == id)
+   return node;
+   }
+   rte_errno = EINVAL;
+   return NULL;
+}
+
+static rte_node_t
+next_next_free_id(void)
+{
+   struct node *node;
+   rte_node_t id = 0;
+
+   STAILQ_FOREACH(node, &node_list, next) {
+   if (id < node->id)
+   break;
+   id = node->id + 1;
+   }
+   return id;
+}
+
+static void
+node_insert_ordered(struct node *node)
+{
+   struct node *after, *g;
+
+   after = NULL;
+   STAILQ_FOREACH(g, &node_list, next) {
+   if (g->id < node->id)
+   after = g;
+   else if (g->id > node->id)
+   break;
+   }
+   if (after == NULL)
+   STAILQ_INSERT_HEAD(&node_list, node, next);
+   else
+   STAILQ_INSERT_AFTER(&node_list, after, node, next);
+}
 
 /* Private functions */

 struct node_head *
@@ -116,10 +158,10 @@ __rte_node_register(const struct rte_node_register *reg)
}
 
 	node->lcore_id = RTE_MAX_LCORE;

-   node->id = node_id++;
+   node->id = next_next_free_id();
 
-	/* Add the node at tail */

-   STAILQ_INSERT_TAIL(&node_list, node, next);
+   /* Add the node in ordered list */
+   node_insert_ordered(node);
graph_spinlock_unlock();
 
 	return node->id;

@@ -194,7 +236,9 @@ rte_node_clone(rte_node_t id, const char *name)
 {
struct node *node;
 
-	NODE_ID_CHECK(id);

+   if (node_from_id(id) == NULL)
+   goto fail;
+
STAILQ_FOREACH(node, &node_list, next)
if (node->id == id)
return node_clone(node, name);
@@ -220,7 +264,8 @@ rte_node_id_to_name(rte_node_t id)
 {
struct node *node;
 
-	NODE_ID_CHECK(id);

+   if (node_from_id(id) == NULL)
+   goto fail;
STAILQ_FOREACH(node, &node_list, next)
if (node->id == id)
return node->name;
@@ -234,7 +279,8 @@ rte_node_edge_count(rte_node_t id)
 {
struct node *node;
 
-	NODE_ID_CHECK(id);

+   if (node_from_id(id) == NULL)
+   goto fail;
STAILQ_FOREACH(node, &node_list, next)
if (node->id == id)
return node->nb_edges;
@@ -303,7 +349,8 @@ rte_node_edge_shrink(rte_node_t id, rte_edge_t size)
rte_edge_t rc = RTE_EDGE_ID_INVALID;
struct node *node;
 
-	NODE_ID_CHECK(id);

+   if (node_from_id(id) == NULL)
+   goto fail;
graph_spinlock_lock();
 
 	STAILQ_FOREACH(node, &node_list, next) {

@@ -330,7 +377,8 @@ rte_node_edge_update(rte_node_t id, rte_edge_t from, const 
char **next_nodes,
rte_edge_t rc = RTE_EDGE_ID_I

[PATCH v9 3/7] ring: make copying functions generic

Note upfront: that change doesn't introduce any functional
or performance changes.
It is just a code-reordering for:
 - improve code modularity and re-usability
 - ability in future to re-use the same code to introduce new functionality

There is no real need for enqueue_elems()/dequeue_elems()
to get pointer to actual rte_ring structure, instead it is enough to pass
a pointer to actual elements buffer inside the ring.
In return, we'll get a copying functions that could be used for other
queueing abstractions that do have circular ring buffer inside.

Signed-off-by: Konstantin Ananyev 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 lib/ring/rte_ring_elem_pvt.h | 115 ---
 1 file changed, 67 insertions(+), 48 deletions(-)

diff --git a/lib/ring/rte_ring_elem_pvt.h b/lib/ring/rte_ring_elem_pvt.h
index 3a83668a08..6eafae121f 100644
--- a/lib/ring/rte_ring_elem_pvt.h
+++ b/lib/ring/rte_ring_elem_pvt.h
@@ -17,12 +17,14 @@
 #endif
 
 static __rte_always_inline void
-__rte_ring_enqueue_elems_32(struct rte_ring *r, const uint32_t size,
-   uint32_t idx, const void *obj_table, uint32_t n)
+__rte_ring_enqueue_elems_32(void *ring_table, const void *obj_table,
+   uint32_t size, uint32_t idx, uint32_t n)
 {
unsigned int i;
-   uint32_t *ring = (uint32_t *)&r[1];
+
+   uint32_t *ring = (uint32_t *)ring_table;
const uint32_t *obj = (const uint32_t *)obj_table;
+
if (likely(idx + n <= size)) {
for (i = 0; i < (n & ~0x7); i += 8, idx += 8) {
ring[idx] = obj[i];
@@ -60,14 +62,14 @@ __rte_ring_enqueue_elems_32(struct rte_ring *r, const 
uint32_t size,
 }
 
 static __rte_always_inline void
-__rte_ring_enqueue_elems_64(struct rte_ring *r, uint32_t prod_head,
-   const void *obj_table, uint32_t n)
+__rte_ring_enqueue_elems_64(void *ring_table, const void *obj_table,
+   uint32_t size, uint32_t idx, uint32_t n)
 {
unsigned int i;
-   const uint32_t size = r->size;
-   uint32_t idx = prod_head & r->mask;
-   uint64_t *ring = (uint64_t *)&r[1];
+
+   uint64_t *ring = (uint64_t *)ring_table;
const unaligned_uint64_t *obj = (const unaligned_uint64_t *)obj_table;
+
if (likely(idx + n <= size)) {
for (i = 0; i < (n & ~0x3); i += 4, idx += 4) {
ring[idx] = obj[i];
@@ -93,14 +95,14 @@ __rte_ring_enqueue_elems_64(struct rte_ring *r, uint32_t 
prod_head,
 }
 
 static __rte_always_inline void
-__rte_ring_enqueue_elems_128(struct rte_ring *r, uint32_t prod_head,
-   const void *obj_table, uint32_t n)
+__rte_ring_enqueue_elems_128(void *ring_table, const void *obj_table,
+   uint32_t size, uint32_t idx, uint32_t n)
 {
unsigned int i;
-   const uint32_t size = r->size;
-   uint32_t idx = prod_head & r->mask;
-   rte_int128_t *ring = (rte_int128_t *)&r[1];
+
+   rte_int128_t *ring = (rte_int128_t *)ring_table;
const rte_int128_t *obj = (const rte_int128_t *)obj_table;
+
if (likely(idx + n <= size)) {
for (i = 0; i < (n & ~0x1); i += 2, idx += 2)
memcpy((void *)(ring + idx),
@@ -126,37 +128,47 @@ __rte_ring_enqueue_elems_128(struct rte_ring *r, uint32_t 
prod_head,
  * single and multi producer enqueue functions.
  */
 static __rte_always_inline void
-__rte_ring_enqueue_elems(struct rte_ring *r, uint32_t prod_head,
-   const void *obj_table, uint32_t esize, uint32_t num)
+__rte_ring_do_enqueue_elems(void *ring_table, const void *obj_table,
+   uint32_t size, uint32_t idx, uint32_t esize, uint32_t num)
 {
/* 8B and 16B copies implemented individually to retain
 * the current performance.
 */
if (esize == 8)
-   __rte_ring_enqueue_elems_64(r, prod_head, obj_table, num);
+   __rte_ring_enqueue_elems_64(ring_table, obj_table, size,
+   idx, num);
else if (esize == 16)
-   __rte_ring_enqueue_elems_128(r, prod_head, obj_table, num);
+   __rte_ring_enqueue_elems_128(ring_table, obj_table, size,
+   idx, num);
else {
-   uint32_t idx, scale, nr_idx, nr_num, nr_size;
+   uint32_t scale, nr_idx, nr_num, nr_size;
 
/* Normalize to uint32_t */
scale = esize / sizeof(uint32_t);
nr_num = num * scale;
-   idx = prod_head & r->mask;
nr_idx = idx * scale;
-   nr_size = r->size * scale;
-   __rte_ring_enqueue_elems_32(r, nr_size, nr_idx,
-   obj_table, nr_num);
+   nr_size = size * scale;
+   __rte_ring_enqueue_elems_32(ring_table, obj_table, nr_size,
+   nr_idx, nr_num);
}
 }
 
 static __rte_always_inline void
-__rte_ring_dequeue_elems_32(struct rte_ring *r, const uint32

[PATCH] power: fix a typo in the PM QoS guide

The typo in the guide is hard to understand. Necessary to fix it.

Fixes: dd6fd75bf662 ("power: introduce PM QoS API on CPU wide")

Signed-off-by: Huisong Li 
---
 doc/guides/prog_guide/power_man.rst | 2 +-
 lib/power/rte_power_qos.h   | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/doc/guides/prog_guide/power_man.rst 
b/doc/guides/prog_guide/power_man.rst
index 22e6e4fe1d..024670a9b4 100644
--- a/doc/guides/prog_guide/power_man.rst
+++ b/doc/guides/prog_guide/power_man.rst
@@ -118,7 +118,7 @@ based on this CPU resume latency in their idle task.
 
 The deeper the idle state, the lower the power consumption,
 but the longer the resume time.
-Some services are latency sensitive and very except the low resume time,
+Some services are latency sensitive and request the low resume time,
 like interrupt packet receiving mode.
 
 Applications can set and get the CPU resume latency with
diff --git a/lib/power/rte_power_qos.h b/lib/power/rte_power_qos.h
index 7a8dab9272..a6d3677409 100644
--- a/lib/power/rte_power_qos.h
+++ b/lib/power/rte_power_qos.h
@@ -24,7 +24,7 @@ extern "C" {
  * 
https://www.kernel.org/doc/html/latest/admin-guide/abi-testing.html?highlight=pm_qos_resume_latency_us#abi-sys-devices-power-pm-qos-resume-latency-us
  *
  * The deeper the idle state, the lower the power consumption, but the
- * longer the resume time. Some service are delay sensitive and very except the
+ * longer the resume time. Some service are delay sensitive and request the
  * low resume time, like interrupt packet receiving mode.
  *
  * In these case, per-CPU PM QoS API can be used to control this CPU's idle
-- 
2.22.0

Re: [RFC] ip_frag: support IPv6 reassembly with extensions

11/11/2024 13:59, Konstantin Ananyev:
> 
> > 14/02/2024 04:51, Stephen Hemminger:
> > > On Tue, 13 Feb 2024 12:47:27 +0100
> > >  wrote:
> > >
> > > > +/*
> > > > + * Function to crawl through the extension header stack.
> > > > + * This function breaks as soon a the fragment header is
> > > > + * found and returns the total length the traversed exts
> > > > + * and the last extension before the fragment header
> > > > + */
> > > > +static inline uint32_t
> > > > +ip_frag_get_last_exthdr(struct rte_ipv6_hdr *ip_hdr, uint8_t 
> > > > **last_ext)
> > > > +{
> > > > +   uint32_t total_len = 0;
> > > > +   size_t ext_len = 0;
> > > > +   *last_ext = (uint8_t *)(ip_hdr + 1);
> > > > +   int next_proto = ip_hdr->proto;
> > > > +
> > > > +   while (next_proto != IPPROTO_FRAGMENT &&
> > > > +   (next_proto = rte_ipv6_get_next_ext(
> > > > +   *last_ext, next_proto, &ext_len)) >= 0) {
> > > > +
> > > > +   total_len += ext_len;
> > > > +
> > > > +   if (next_proto == IPPROTO_FRAGMENT)
> > > > +   return total_len;
> > > > +
> > > > +   *last_ext += ext_len;
> > > > +   }
> > > > +
> > > > +   return total_len;
> > > > +}
> > >
> > > Doing endless loop like this opens up DoS attacks.
> > > Better to use rte_next_skip_ip6_ext() or do similar limited loop.
> > 
> > There was no reply to this interesting comment?
> 
> I think there is a limit is the latest version for that patch:
> https://patchwork.dpdk.org/project/dpdk/patch/20241015082133.3910533-1-vignesh.purushotham.srini...@ericsson.com/

Even if there is a new version, we should reply to comments.

> It is also has an ACK from me...
> Though looking at it once again - we'd better have an extra check here to
> make sure that total_len would not exceed mbuf->data_len. 

I suppose it can wait 25.03.

Re: [PATCH] power: fix a typo in the PM QoS guide

11/11/2024 13:52, Huisong Li:
> -Some services are latency sensitive and very except the low resume time,
> +Some services are latency sensitive and request the low resume time,

I suppose it should be "request a low resume time".

Re: [PATCH v1] mailmap: change primary mail address

07/11/2024 20:10, Konstantin Ananyev:
> use working mail address as a main one.
> 
> Signed-off-by: Konstantin Ananyev 
> ---
> -Konstantin Ananyev  
>  
> +Konstantin Ananyev  
> 

You must keep your old email address as well to associate your history.

Applied with this change, thanks.

[PATCH v10 1/7] test/ring: fix failure with custom number of lcores

ring_stress_autotest fails to initialize the ring when RTE_MAX_LCORE value
is not a number of 2.
There is a flaw in calculation required number of elements in the ring.
Fix it by aligning number of elements to next power of 2.

Fixes: bf28df24e915 ("test/ring: add contention stress test")
Cc: sta...@dpdk.org

Reported-by: Phanendra Vukkisala 
Signed-off-by: Konstantin Ananyev 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 app/test/test_ring_stress_impl.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/app/test/test_ring_stress_impl.h b/app/test/test_ring_stress_impl.h
index 8b0bfb11fe..ee5274aeef 100644
--- a/app/test/test_ring_stress_impl.h
+++ b/app/test/test_ring_stress_impl.h
@@ -297,7 +297,7 @@ mt1_init(struct rte_ring **rng, void **data, uint32_t num)
*data = elm;
 
/* alloc ring */
-   nr = 2 * num;
+   nr = rte_align32pow2(2 * num);
sz = rte_ring_get_memsize(nr);
r = rte_zmalloc(NULL, sz, alignof(typeof(*r)));
if (r == NULL) {
-- 
2.35.3

[PATCH v10 2/7] ring: common functions for 'move head' ops

Note upfront: that change doesn't introduce any functional or
performance changes.
It is just a code-reordering for:
 - code deduplication
 - ability in future to re-use the same code to introduce new functionality

For each sync mode corresponding move_prod_head() and
move_cons_head() are nearly identical to each other,
the only differences are:
 - do we need to use a @capacity to calculate number of entries or not.
 - what we need to update (prod/cons) and what is used as
   read-only counterpart.
So instead of having 2 copies of nearly identical functions,
introduce a new common one that could be used by both functions:
move_prod_head() and move_cons_head().

As another positive thing - we can get rid of referencing whole rte_ring
structure in that new common sub-function.

Signed-off-by: Konstantin Ananyev 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 lib/ring/rte_ring_c11_pvt.h  | 156 ++-
 lib/ring/rte_ring_elem_pvt.h |  66 +
 lib/ring/rte_ring_generic_pvt.h  | 143 +---
 lib/ring/rte_ring_hts_elem_pvt.h | 107 ++---
 lib/ring/rte_ring_rts_elem_pvt.h | 107 ++---
 5 files changed, 255 insertions(+), 324 deletions(-)

diff --git a/lib/ring/rte_ring_c11_pvt.h b/lib/ring/rte_ring_c11_pvt.h
index 629b2d9288..b9388af0da 100644
--- a/lib/ring/rte_ring_c11_pvt.h
+++ b/lib/ring/rte_ring_c11_pvt.h
@@ -11,6 +11,17 @@
 #ifndef _RTE_RING_C11_PVT_H_
 #define _RTE_RING_C11_PVT_H_
 
+/**
+ * @file rte_ring_c11_pvt.h
+ * It is not recommended to include this file directly,
+ * include  instead.
+ * Contains internal helper functions for MP/SP and MC/SC ring modes.
+ * For more information please refer to .
+ */
+
+/**
+ * @internal This function updates tail values.
+ */
 static __rte_always_inline void
 __rte_ring_update_tail(struct rte_ring_headtail *ht, uint32_t old_val,
uint32_t new_val, uint32_t single, uint32_t enqueue)
@@ -29,40 +40,45 @@ __rte_ring_update_tail(struct rte_ring_headtail *ht, 
uint32_t old_val,
 }
 
 /**
- * @internal This function updates the producer head for enqueue
+ * @internal This is a helper function that moves the producer/consumer head
  *
- * @param r
- *   A pointer to the ring structure
- * @param is_sp
- *   Indicates whether multi-producer path is needed or not
+ * @param d
+ *   A pointer to the headtail structure with head value to be moved
+ * @param s
+ *   A pointer to the counter-part headtail structure. Note that this
+ *   function only reads tail value from it
+ * @param capacity
+ *   Either ring capacity value (for producer), or zero (for consumer)
+ * @param is_st
+ *   Indicates whether multi-thread safe path is needed or not
  * @param n
- *   The number of elements we will want to enqueue, i.e. how far should the
- *   head be moved
+ *   The number of elements we want to move head value on
  * @param behavior
- *   RTE_RING_QUEUE_FIXED:Enqueue a fixed number of items from a ring
- *   RTE_RING_QUEUE_VARIABLE: Enqueue as many items as possible from ring
+ *   RTE_RING_QUEUE_FIXED:Move on a fixed number of items
+ *   RTE_RING_QUEUE_VARIABLE: Move on as many items as possible
  * @param old_head
- *   Returns head value as it was before the move, i.e. where enqueue starts
+ *   Returns head value as it was before the move
  * @param new_head
- *   Returns the current/new head value i.e. where enqueue finishes
- * @param free_entries
- *   Returns the amount of free space in the ring BEFORE head was moved
+ *   Returns the new head value
+ * @param entries
+ *   Returns the number of ring entries available BEFORE head was moved
  * @return
- *   Actual number of objects enqueued.
- *   If behavior == RTE_RING_QUEUE_FIXED, this will be 0 or n only.
+ *   Actual number of objects the head was moved on
+ *   If behavior == RTE_RING_QUEUE_FIXED, this will be 0 or n only
  */
 static __rte_always_inline unsigned int
-__rte_ring_move_prod_head(struct rte_ring *r, unsigned int is_sp,
-   unsigned int n, enum rte_ring_queue_behavior behavior,
-   uint32_t *old_head, uint32_t *new_head,
-   uint32_t *free_entries)
+__rte_ring_headtail_move_head(struct rte_ring_headtail *d,
+   const struct rte_ring_headtail *s, uint32_t capacity,
+   unsigned int is_st, unsigned int n,
+   enum rte_ring_queue_behavior behavior,
+   uint32_t *old_head, uint32_t *new_head, uint32_t *entries)
 {
-   const uint32_t capacity = r->capacity;
-   uint32_t cons_tail;
-   unsigned int max = n;
+   uint32_t stail;
int success;
+   unsigned int max = n;
 
-   *old_head = rte_atomic_load_explicit(&r->prod.head, 
rte_memory_order_relaxed);
+   *old_head = rte_atomic_load_explicit(&d->head,
+   rte_memory_order_relaxed);
do {
/* Reset n to the initial burst count */
n = max;
@@ -73,112 +89,36 @@

[PATCH v10 0/7] Stage-Ordered API and other extensions for ring library

Testing coverage (passed):
x86_64, i686, PPC, ARM   

Would like to express my gratitude to all community members who helped
with testing it on different platforms, in particular:
David Christensen 
Cody Cheng 
Patrick Robb 
Phanendra Vukkisala 
Chengwen Feng 

v9 - v10
- fix clang14 complaints regarding stdatomic (app/test)

v8 -> v9
- fix grammar (Thomas)
- test: use stdatomic API (Thomas)

v7 -> v8
- fixed several nits (spotted by Morten):
  indentation, variable names, comments, etc.

v6 -> v7
- updated Programmer Guide (Jerin, Morten, Stephen)
- fix some functions in public headers without comments (Morten)
- update debug checks, added new macro for that: RTE_SORING_DEBUG
  (disabled by default).

v5 -> v6
- fix problem with ring_stress_autotest (Phanendra)
- added more checks and debug output

v4 -> v5
- fix public API/doc comments from Jerin
- update devtools/build-dict.sh (Stephen)
- fix MSVC warnings
- introduce new test-suite for meson (stress) with
  ring_stress_autotest and soring_stress_autotest in it
- enhance error report in tests
- reorder some sync code in soring and add extra checks
  (for better debuggability)

v3 -> v4:
- fix compilation/doxygen complains (attempt #2)
- updated release notes

v2 -> v3:
- fix compilation/doxygen complains
- dropped patch:
  "examples/l3fwd: make ACL work in pipeline and eventdev modes": [2]
  As was mentioned in the patch desctiption it was way too big,
  controversial and incomplete. If the community is ok to introduce
  pipeline model into the l3fwd, then it is propbably worth to be
  a separate patch series.

v1 -> v2:
- rename 'elmst/objst' to 'meta' (Morten)
- introduce new data-path APIs set: one with both meta{} and objs[],
  second with just objs[] (Morten)
- split data-path APIs into burst/bulk flavours (same as rte_ring)
- added dump function for te_soring and improved dump() for rte_ring.
- dropped patch:
  " ring: minimize reads of the counterpart cache-line"
  - no performance gain observed
  - actually it does change behavior of conventional rte_ring
enqueue/dequeue APIs -
it could return available/free less then actually exist in the ring.
As in some other libs we reliy on that information - it will
introduce problems.

The main aim of these series is to extend ring library with
new API that allows user to create/use Staged-Ordered-Ring (SORING)
abstraction. In addition to that there are few other patches that serve
different purposes:
- first two patches are just code reordering to de-duplicate
  and generalize existing rte_ring code.
- patch #3 extends rte_ring_dump() to correctly print head/tail metadata
  for different sync modes.
- next two patches introduce SORING API into the ring library and
  provide UT for it.

SORING overview
===
Staged-Ordered-Ring (SORING) provides a SW abstraction for 'ordered' queues
with multiple processing 'stages'. It is based on conventional DPDK
rte_ring, re-uses many of its concepts, and even substantial part of
its code.
It can be viewed as an 'extension' of rte_ring functionality.
In particular, main SORING properties:
- circular ring buffer with fixed size objects
- producer, consumer plus multiple processing stages in between.
- allows to split objects processing into multiple stages.
- objects remain in the same ring while moving from one stage to the other,
  initial order is preserved, no extra copying needed.
- preserves the ingress order of objects within the queue across multiple
  stages
- each stage (and producer/consumer) can be served by single and/or
  multiple threads.

- number of stages, size and number of objects in the ring are
 configurable at ring initialization time.

Data-path API provides four main operations:
- enqueue/dequeue works in the same manner as for conventional rte_ring,
  all rte_ring synchronization types are supported.
- acquire/release - for each stage there is an acquire (start) and
  release (finish) operation. After some objects are 'acquired' -
  given thread can safely assume that it has exclusive ownership of
  these objects till it will invoke 'release' for them.
  After 'release', objects can be 'acquired' by next stage and/or dequeued
  by the consumer (in case of last stage).

Expected use-case: applications that uses pipeline model
(probably with multiple stages) for packet processing, when preserving
incoming packet order is important.

The concept of ‘ring with stages’ is similar to DPDK OPDL eventdev PMD [1],
but the internals are different.
In particular, SORING maintains internal array of 'states' for each element
in the ring that is  shared by all threads/processes that access the ring.
That allows 'release' to avoid excessive waits on the tail value and helps
to improve performancei and scalability.
In terms of performance, with our measurements rte_soring and
conventional rte_ring provide nearly identical numbers.
As an example, on our SUT: Intel ICX CPU @ 2.00GHz,
l3fwd (--lookup=acl) in pipeline mode [2] both
rte_r

[PATCH v10 4/7] ring: make dump function more verbose

From: Eimear Morrissey 

The current rte_ring_dump function uses the generic rte_ring_headtail
structure to access head/tail positions. This is incorrect for the RTS
case where the head is stored in a different offset in the union of
structs. Switching to a separate function for each sync type allows
to dump correct head/tail values and extra metadata.

Signed-off-by: Eimear Morrissey 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 .mailmap |  1 +
 app/test/test_ring_stress_impl.h |  1 +
 lib/ring/rte_ring.c  | 87 ++--
 lib/ring/rte_ring.h  | 15 ++
 lib/ring/version.map |  7 +++
 5 files changed, 107 insertions(+), 4 deletions(-)

diff --git a/.mailmap b/.mailmap
index 4894219f2f..dbffe91d8a 100644
--- a/.mailmap
+++ b/.mailmap
@@ -388,6 +388,7 @@ Eduard Serra 
 Edward Makarov 
 Edwin Brossette 
 Eelco Chaudron 
+Eimear Morrissey 
 Elad Nachman 
 Elad Persiko 
 Elena Agostini 
diff --git a/app/test/test_ring_stress_impl.h b/app/test/test_ring_stress_impl.h
index ee5274aeef..f99a7ff675 100644
--- a/app/test/test_ring_stress_impl.h
+++ b/app/test/test_ring_stress_impl.h
@@ -380,6 +380,7 @@ test_mt1(int (*test)(void *))
}
 
lcore_stat_dump(stdout, UINT32_MAX, &arg[mc].stats);
+   rte_ring_dump(stdout, r);
mt1_fini(r, data);
return rc;
 }
diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c
index aebb6d6728..a7d9d6b037 100644
--- a/lib/ring/rte_ring.c
+++ b/lib/ring/rte_ring.c
@@ -364,20 +364,99 @@ rte_ring_free(struct rte_ring *r)
rte_free(te);
 }
 
+static const char *
+ring_get_sync_type(const enum rte_ring_sync_type st)
+{
+   switch (st) {
+   case RTE_RING_SYNC_ST:
+   return "ST";
+   case RTE_RING_SYNC_MT:
+   return "MT";
+   case RTE_RING_SYNC_MT_RTS:
+   return "MT_RTS";
+   case RTE_RING_SYNC_MT_HTS:
+   return "MT_HTS";
+   default:
+   return "unknown";
+   }
+}
+
+static void
+ring_dump_ht_headtail(FILE *f, const char *prefix,
+   const struct rte_ring_headtail *ht)
+{
+   fprintf(f, "%ssync_type=%s\n", prefix,
+   ring_get_sync_type(ht->sync_type));
+   fprintf(f, "%shead=%"PRIu32"\n", prefix, ht->head);
+   fprintf(f, "%stail=%"PRIu32"\n", prefix, ht->tail);
+}
+
+static void
+ring_dump_rts_headtail(FILE *f, const char *prefix,
+   const struct rte_ring_rts_headtail *rts)
+{
+   fprintf(f, "%ssync_type=%s\n", prefix,
+   ring_get_sync_type(rts->sync_type));
+   fprintf(f, "%shead.pos=%"PRIu32"\n", prefix, rts->head.val.pos);
+   fprintf(f, "%shead.cnt=%"PRIu32"\n", prefix, rts->head.val.cnt);
+   fprintf(f, "%stail.pos=%"PRIu32"\n", prefix, rts->tail.val.pos);
+   fprintf(f, "%stail.cnt=%"PRIu32"\n", prefix, rts->tail.val.cnt);
+   fprintf(f, "%shtd_max=%"PRIu32"\n", prefix, rts->htd_max);
+}
+
+static void
+ring_dump_hts_headtail(FILE *f, const char *prefix,
+   const struct rte_ring_hts_headtail *hts)
+{
+   fprintf(f, "%ssync_type=%s\n", prefix,
+   ring_get_sync_type(hts->sync_type));
+   fprintf(f, "%shead=%"PRIu32"\n", prefix, hts->ht.pos.head);
+   fprintf(f, "%stail=%"PRIu32"\n", prefix, hts->ht.pos.tail);
+}
+
+void
+rte_ring_headtail_dump(FILE *f, const char *prefix,
+   const struct rte_ring_headtail *r)
+{
+   if (f == NULL || r == NULL)
+   return;
+
+   prefix = (prefix != NULL) ? prefix : "";
+
+   switch (r->sync_type) {
+   case RTE_RING_SYNC_ST:
+   case RTE_RING_SYNC_MT:
+   ring_dump_ht_headtail(f, prefix, r);
+   break;
+   case RTE_RING_SYNC_MT_RTS:
+   ring_dump_rts_headtail(f, prefix,
+   (const struct rte_ring_rts_headtail *)r);
+   break;
+   case RTE_RING_SYNC_MT_HTS:
+   ring_dump_hts_headtail(f, prefix,
+   (const struct rte_ring_hts_headtail *)r);
+   break;
+   default:
+   RING_LOG(ERR, "Invalid ring sync type detected");
+   }
+}
+
 /* dump the status of the ring on the console */
 void
 rte_ring_dump(FILE *f, const struct rte_ring *r)
 {
+   if (f == NULL || r == NULL)
+   return;
+
fprintf(f, "ring <%s>@%p\n", r->name, r);
fprintf(f, "  flags=%x\n", r->flags);
fprintf(f, "  size=%"PRIu32"\n", r->size);
fprintf(f, "  capacity=%"PRIu32"\n", r->capacity);
-   fprintf(f, "  ct=%"PRIu32"\n", r->cons.tail);
-   fprintf(f, "  ch=%"PRIu32"\n", r->cons.head);
-   fprintf(f, "  pt=%"PRIu32"\n", r->prod.tail);
-   fprintf(f, "  ph=%"PRIu32"\n", r->prod.head);
fprintf(f, "  used=%u\n", rte_ring_count(r));
fprintf(f, "  avail=%u\n", rte_ring_free_count(r));
+
+   rte_ring_headtail_dump(f, "  cons.", &(r->cons));
+

[PATCH v10 7/7] test: add stress test suite

Add a new test suite which purpose is to run 'stress' tests:
main purpose is put a pressure to dpdk sync algorithms
to flag their misbehaving/slowdown/etc.
Right now it consists from just 2 test-cases:
meson test --suite stress-tests --list
DPDK:stress-tests / ring_stress_autotest
DPDK:stress-tests / soring_stress_autotest

These tests are quite time consuming (~15 mins each),
that's another reason to put them into a separate test-suite.

Signed-off-by: Konstantin Ananyev 
Acked-by: Stephen Hemminger 
---
 app/test/suites/meson.build   | 10 ++
 app/test/test.h   |  1 +
 app/test/test_ring_stress.c   |  2 +-
 app/test/test_soring_stress.c |  2 +-
 4 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/app/test/suites/meson.build b/app/test/suites/meson.build
index 191702cf76..e482373330 100644
--- a/app/test/suites/meson.build
+++ b/app/test/suites/meson.build
@@ -5,6 +5,7 @@
 # to complete, so timeout to 10 minutes
 timeout_seconds = 600
 timeout_seconds_fast = 10
+timeout_seconds_stress = 900
 
 test_no_huge_args = ['--no-huge', '-m', '2048']
 has_hugepage = run_command(has_hugepages_cmd, check: true).stdout().strip() != 
'0'
@@ -21,6 +22,7 @@ endif
 # - fast_tests
 # - perf_tests
 # - driver_tests
+# - stress_tests
 test_suites = run_command(get_test_suites_cmd, autotest_sources,
  check: true).stdout().strip().split()
 foreach suite:test_suites
@@ -39,6 +41,14 @@ foreach suite:test_suites
 timeout: timeout_seconds,
 is_parallel: false)
 endforeach
+elif suite_name == 'stress-tests'
+foreach t: suite_tests
+test(t, dpdk_test,
+env: ['DPDK_TEST=' + t],
+timeout: timeout_seconds_stress,
+is_parallel: false,
+suite: suite_name)
+endforeach
 elif suite_name != 'fast-tests'
 # simple cases - tests without parameters or special handling
 foreach t: suite_tests
diff --git a/app/test/test.h b/app/test/test.h
index 15e23d297f..ebc4864bf8 100644
--- a/app/test/test.h
+++ b/app/test/test.h
@@ -208,5 +208,6 @@ void add_test_command(struct test_command *t);
 #define REGISTER_FAST_TEST(cmd, no_huge, ASan, func)  
REGISTER_TEST_COMMAND(cmd, func)
 #define REGISTER_PERF_TEST REGISTER_TEST_COMMAND
 #define REGISTER_DRIVER_TEST REGISTER_TEST_COMMAND
+#define REGISTER_STRESS_TEST REGISTER_TEST_COMMAND
 
 #endif
diff --git a/app/test/test_ring_stress.c b/app/test/test_ring_stress.c
index 1af45e0fc8..82e19b02c3 100644
--- a/app/test/test_ring_stress.c
+++ b/app/test/test_ring_stress.c
@@ -63,4 +63,4 @@ test_ring_stress(void)
return (k != n);
 }
 
-REGISTER_TEST_COMMAND(ring_stress_autotest, test_ring_stress);
+REGISTER_STRESS_TEST(ring_stress_autotest, test_ring_stress);
diff --git a/app/test/test_soring_stress.c b/app/test/test_soring_stress.c
index 334af6a29c..e5655d49cb 100644
--- a/app/test/test_soring_stress.c
+++ b/app/test/test_soring_stress.c
@@ -45,4 +45,4 @@ test_ring_stress(void)
return (k != n);
 }
 
-REGISTER_TEST_COMMAND(soring_stress_autotest, test_ring_stress);
+REGISTER_STRESS_TEST(soring_stress_autotest, test_ring_stress);
-- 
2.35.3

[PATCH v10 3/7] ring: make copying functions generic

Note upfront: that change doesn't introduce any functional
or performance changes.
It is just a code-reordering for:
 - improve code modularity and re-usability
 - ability in future to re-use the same code to introduce new functionality

There is no real need for enqueue_elems()/dequeue_elems()
to get pointer to actual rte_ring structure, instead it is enough to pass
a pointer to actual elements buffer inside the ring.
In return, we'll get a copying functions that could be used for other
queueing abstractions that do have circular ring buffer inside.

Signed-off-by: Konstantin Ananyev 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 lib/ring/rte_ring_elem_pvt.h | 115 ---
 1 file changed, 67 insertions(+), 48 deletions(-)

diff --git a/lib/ring/rte_ring_elem_pvt.h b/lib/ring/rte_ring_elem_pvt.h
index 3a83668a08..6eafae121f 100644
--- a/lib/ring/rte_ring_elem_pvt.h
+++ b/lib/ring/rte_ring_elem_pvt.h
@@ -17,12 +17,14 @@
 #endif
 
 static __rte_always_inline void
-__rte_ring_enqueue_elems_32(struct rte_ring *r, const uint32_t size,
-   uint32_t idx, const void *obj_table, uint32_t n)
+__rte_ring_enqueue_elems_32(void *ring_table, const void *obj_table,
+   uint32_t size, uint32_t idx, uint32_t n)
 {
unsigned int i;
-   uint32_t *ring = (uint32_t *)&r[1];
+
+   uint32_t *ring = (uint32_t *)ring_table;
const uint32_t *obj = (const uint32_t *)obj_table;
+
if (likely(idx + n <= size)) {
for (i = 0; i < (n & ~0x7); i += 8, idx += 8) {
ring[idx] = obj[i];
@@ -60,14 +62,14 @@ __rte_ring_enqueue_elems_32(struct rte_ring *r, const 
uint32_t size,
 }
 
 static __rte_always_inline void
-__rte_ring_enqueue_elems_64(struct rte_ring *r, uint32_t prod_head,
-   const void *obj_table, uint32_t n)
+__rte_ring_enqueue_elems_64(void *ring_table, const void *obj_table,
+   uint32_t size, uint32_t idx, uint32_t n)
 {
unsigned int i;
-   const uint32_t size = r->size;
-   uint32_t idx = prod_head & r->mask;
-   uint64_t *ring = (uint64_t *)&r[1];
+
+   uint64_t *ring = (uint64_t *)ring_table;
const unaligned_uint64_t *obj = (const unaligned_uint64_t *)obj_table;
+
if (likely(idx + n <= size)) {
for (i = 0; i < (n & ~0x3); i += 4, idx += 4) {
ring[idx] = obj[i];
@@ -93,14 +95,14 @@ __rte_ring_enqueue_elems_64(struct rte_ring *r, uint32_t 
prod_head,
 }
 
 static __rte_always_inline void
-__rte_ring_enqueue_elems_128(struct rte_ring *r, uint32_t prod_head,
-   const void *obj_table, uint32_t n)
+__rte_ring_enqueue_elems_128(void *ring_table, const void *obj_table,
+   uint32_t size, uint32_t idx, uint32_t n)
 {
unsigned int i;
-   const uint32_t size = r->size;
-   uint32_t idx = prod_head & r->mask;
-   rte_int128_t *ring = (rte_int128_t *)&r[1];
+
+   rte_int128_t *ring = (rte_int128_t *)ring_table;
const rte_int128_t *obj = (const rte_int128_t *)obj_table;
+
if (likely(idx + n <= size)) {
for (i = 0; i < (n & ~0x1); i += 2, idx += 2)
memcpy((void *)(ring + idx),
@@ -126,37 +128,47 @@ __rte_ring_enqueue_elems_128(struct rte_ring *r, uint32_t 
prod_head,
  * single and multi producer enqueue functions.
  */
 static __rte_always_inline void
-__rte_ring_enqueue_elems(struct rte_ring *r, uint32_t prod_head,
-   const void *obj_table, uint32_t esize, uint32_t num)
+__rte_ring_do_enqueue_elems(void *ring_table, const void *obj_table,
+   uint32_t size, uint32_t idx, uint32_t esize, uint32_t num)
 {
/* 8B and 16B copies implemented individually to retain
 * the current performance.
 */
if (esize == 8)
-   __rte_ring_enqueue_elems_64(r, prod_head, obj_table, num);
+   __rte_ring_enqueue_elems_64(ring_table, obj_table, size,
+   idx, num);
else if (esize == 16)
-   __rte_ring_enqueue_elems_128(r, prod_head, obj_table, num);
+   __rte_ring_enqueue_elems_128(ring_table, obj_table, size,
+   idx, num);
else {
-   uint32_t idx, scale, nr_idx, nr_num, nr_size;
+   uint32_t scale, nr_idx, nr_num, nr_size;
 
/* Normalize to uint32_t */
scale = esize / sizeof(uint32_t);
nr_num = num * scale;
-   idx = prod_head & r->mask;
nr_idx = idx * scale;
-   nr_size = r->size * scale;
-   __rte_ring_enqueue_elems_32(r, nr_size, nr_idx,
-   obj_table, nr_num);
+   nr_size = size * scale;
+   __rte_ring_enqueue_elems_32(ring_table, obj_table, nr_size,
+   nr_idx, nr_num);
}
 }
 
 static __rte_always_inline void
-__rte_ring_dequeue_elems_32(struct rte_ring *r, const uint32

[PATCH v10 5/7] ring/soring: introduce Staged Ordered Ring

Staged-Ordered-Ring (SORING) provides a SW abstraction for 'ordered' queues
with multiple processing 'stages'.
It is based on conventional DPDK rte_ring, re-uses many of its concepts,
and even substantial part of its code.
It can be viewed as an 'extension' of rte_ring functionality.
In particular, main SORING properties:
- circular ring buffer with fixed size objects
- producer, consumer plus multiple processing stages in the middle.
- allows to split objects processing into multiple stages.
- objects remain in the same ring while moving from one stage to the other,
  initial order is preserved, no extra copying needed.
- preserves the ingress order of objects within the queue across multiple
  stages, i.e.:
  at the same stage multiple threads can process objects from the ring in
  any order, but for the next stage objects will always appear in the
  original order.
- each stage (and producer/consumer) can be served by single and/or
  multiple threads.
- number of stages, size and number of objects in the ring are
  configurable at ring initialization time.

Data-path API provides four main operations:
- enqueue/dequeue works in the same manner as for conventional rte_ring,
  all rte_ring synchronization types are supported.
- acquire/release - for each stage there is an acquire (start) and
  release (finish) operation.
  after some objects are 'acquired' - given thread can safely assume that
  it has exclusive possession of these objects till 'release' for them is
  invoked.
  Note that right now user has to release exactly the same number of
  objects that was acquired before.
  After 'release', objects can be 'acquired' by next stage and/or dequeued
  by the consumer (in case of last stage).

Expected use-case: applications that uses pipeline model
(probably with multiple stages) for packet processing, when preserving
incoming packet order is important. I.E.: IPsec processing, etc.

Signed-off-by: Eimear Morrissey 
Signed-off-by: Konstantin Ananyev 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 devtools/build-dict.sh|   1 +
 doc/api/doxy-api-index.md |   1 +
 doc/guides/prog_guide/img/soring-pic1.svg | 635 ++
 doc/guides/prog_guide/ring_lib.rst| 202 +++
 doc/guides/rel_notes/release_24_11.rst|   8 +
 lib/ring/meson.build  |   4 +-
 lib/ring/rte_soring.c | 198 +++
 lib/ring/rte_soring.h | 557 +++
 lib/ring/soring.c | 613 +
 lib/ring/soring.h | 138 +
 lib/ring/version.map  |  19 +
 11 files changed, 2374 insertions(+), 2 deletions(-)
 create mode 100644 doc/guides/prog_guide/img/soring-pic1.svg
 create mode 100644 lib/ring/rte_soring.c
 create mode 100644 lib/ring/rte_soring.h
 create mode 100644 lib/ring/soring.c
 create mode 100644 lib/ring/soring.h

diff --git a/devtools/build-dict.sh b/devtools/build-dict.sh
index a8cac49029..5eb592abfa 100755
--- a/devtools/build-dict.sh
+++ b/devtools/build-dict.sh
@@ -17,6 +17,7 @@ sed '/^..->/d' |
 sed '/^uint->/d' |
 sed "/^doesn'->/d" |
 sed '/^wasn->/d' |
+sed '/^soring->/d' |
 
 # print to stdout
 cat
diff --git a/doc/api/doxy-api-index.md b/doc/api/doxy-api-index.md
index f0193502bc..b2fc24b3e4 100644
--- a/doc/api/doxy-api-index.md
+++ b/doc/api/doxy-api-index.md
@@ -174,6 +174,7 @@ The public API headers are grouped by topics:
   [mbuf](@ref rte_mbuf.h),
   [mbuf pool ops](@ref rte_mbuf_pool_ops.h),
   [ring](@ref rte_ring.h),
+  [soring](@ref rte_soring.h),
   [stack](@ref rte_stack.h),
   [tailq](@ref rte_tailq.h),
   [bitset](@ref rte_bitset.h),
diff --git a/doc/guides/prog_guide/img/soring-pic1.svg 
b/doc/guides/prog_guide/img/soring-pic1.svg
new file mode 100644
index 00..c97e66ca43
--- /dev/null
+++ b/doc/guides/prog_guide/img/soring-pic1.svg
@@ -0,0 +1,635 @@
+
+
+
+
+
+
+
+http://www.inkscape.org/namespaces/inkscape";
+   xmlns:sodipodi="http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd";
+   xmlns="http://www.w3.org/2000/svg";
+   xmlns:svg="http://www.w3.org/2000/svg";
+   xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#";
+   xmlns:cc="http://creativecommons.org/ns#";
+   xmlns:dc="http://purl.org/dc/elements/1.1/";>
+  
+
+  
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+
+
+
+
+
+
+
+  
+
+
+  
+
+
+
+  
+
+
+
+
+
+
+
+  
+
+
+  
+  
+
+  
+  
+
+  
+image/svg+xml
+http://purl.org/dc/dcmitype/StillImage"; />
+  
+
+  
+  
+
+
+  
+  
+  
+  
+  
+  
+  
+  
+
+obj5
+obj1
+obj2
+obj3
+
+cons_head
+cons_tail
+prod_head
+prod_tail
+
+lstages states
+
+producer and consumer 
states
+
+

[PATCH v10 6/7] app/test: add unit tests for soring API

Add both functional and stess test-cases for soring API.
Stress test serves as both functional and performance test of soring
enqueue/dequeue/acquire/release operations under high contention
(for both over committed and non-over committed scenarios).

Signed-off-by: Eimear Morrissey 
Signed-off-by: Konstantin Ananyev 
Acked-by: Stephen Hemminger 
---
 app/test/meson.build   |   3 +
 app/test/test_soring.c | 442 +++
 app/test/test_soring_mt_stress.c   |  40 ++
 app/test/test_soring_stress.c  |  48 ++
 app/test/test_soring_stress.h  |  35 ++
 app/test/test_soring_stress_impl.h | 838 +
 6 files changed, 1406 insertions(+)
 create mode 100644 app/test/test_soring.c
 create mode 100644 app/test/test_soring_mt_stress.c
 create mode 100644 app/test/test_soring_stress.c
 create mode 100644 app/test/test_soring_stress.h
 create mode 100644 app/test/test_soring_stress_impl.h

diff --git a/app/test/meson.build b/app/test/meson.build
index 40f22a54d5..b11da1bbc7 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -179,6 +179,9 @@ source_file_deps = {
 'test_security_proto.c' : ['cryptodev', 'security'],
 'test_seqlock.c': [],
 'test_service_cores.c': [],
+'test_soring.c': [],
+'test_soring_mt_stress.c': [],
+'test_soring_stress.c': [],
 'test_spinlock.c': [],
 'test_stack.c': ['stack'],
 'test_stack_perf.c': ['stack'],
diff --git a/app/test/test_soring.c b/app/test/test_soring.c
new file mode 100644
index 00..6773a399e6
--- /dev/null
+++ b/app/test/test_soring.c
@@ -0,0 +1,442 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2024 Huawei Technologies Co., Ltd
+ */
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include 
+
+#include "test.h"
+
+#define MAX_ACQUIRED 20
+
+#define SORING_TEST_ASSERT(val, expected) do { \
+   RTE_TEST_ASSERT(expected == val, \
+   "%s: expected %u got %u\n", #val, expected, val); \
+} while (0)
+
+static void
+set_soring_init_param(struct rte_soring_param *prm,
+   const char *name, uint32_t esize, uint32_t elems,
+   uint32_t stages, uint32_t stsize,
+   enum rte_ring_sync_type rst_prod,
+   enum rte_ring_sync_type rst_cons)
+{
+   prm->name = name;
+   prm->elem_size = esize;
+   prm->elems = elems;
+   prm->stages = stages;
+   prm->meta_size = stsize;
+   prm->prod_synt = rst_prod;
+   prm->cons_synt = rst_cons;
+}
+
+static int
+move_forward_stage(struct rte_soring *sor,
+   uint32_t num_packets, uint32_t stage)
+{
+   uint32_t acquired;
+   uint32_t ftoken;
+   uint32_t *acquired_objs[MAX_ACQUIRED];
+
+   acquired = rte_soring_acquire_bulk(sor, acquired_objs, stage,
+   num_packets, &ftoken, NULL);
+   SORING_TEST_ASSERT(acquired, num_packets);
+   rte_soring_release(sor, NULL, stage, num_packets,
+   ftoken);
+
+   return 0;
+}
+
+/*
+ * struct rte_soring_param param checking.
+ */
+static int
+test_soring_init(void)
+{
+   struct rte_soring *sor = NULL;
+   struct rte_soring_param prm;
+   int rc;
+   size_t sz;
+   memset(&prm, 0, sizeof(prm));
+
+/*init memory*/
+   set_soring_init_param(&prm, "alloc_memory", sizeof(uintptr_t),
+   4, 1, 4, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   sz = rte_soring_get_memsize(&prm);
+   sor = rte_zmalloc(NULL, sz, RTE_CACHE_LINE_SIZE);
+   RTE_TEST_ASSERT_NOT_NULL(sor, "could not allocate memory for soring");
+
+   set_soring_init_param(&prm, "test_invalid_stages", sizeof(uintptr_t),
+   4, 0, 4, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   rc = rte_soring_init(sor, &prm);
+   RTE_TEST_ASSERT_FAIL(rc, "initted soring with invalid num stages");
+
+   set_soring_init_param(&prm, "test_invalid_esize", 0,
+   4, 1, 4, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   rc = rte_soring_init(sor, &prm);
+   RTE_TEST_ASSERT_FAIL(rc, "initted soring with 0 esize");
+
+   set_soring_init_param(&prm, "test_invalid_esize", 9,
+   4, 1, 4, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   rc = rte_soring_init(sor, &prm);
+   RTE_TEST_ASSERT_FAIL(rc, "initted soring with esize not multiple of 4");
+
+   set_soring_init_param(&prm, "test_invalid_rsize", sizeof(uintptr_t),
+   4, 1, 3, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   rc = rte_soring_init(sor, &prm);
+   RTE_TEST_ASSERT_FAIL(rc, "initted soring with rcsize not multiple of 
4");
+
+   set_soring_init_param(&prm, "test_invalid_elems", sizeof(uintptr_t),
+   RTE_SORING_ELEM_MAX + 1, 1, 4, RTE_RING_SYNC_MT,
+

答复: [PATCH v6 05/17] net/r8169: add support for hw config

2024-11-11 Thread 王颢

Dear Ferruh,

Before splitting the driver into multiple small patches, I had checked to 
ensure all functions were static, so there’s no need to worry about that.

Best Regards,
Howard Wang

-邮件原件-
发件人: Ferruh Yigit  
发送时间: 2024年11月11日 8:15
收件人: 王颢 ; dev@dpdk.org
抄送: pro_nic_d...@realtek.com
主题: Re: [PATCH v6 05/17] net/r8169: add support for hw config


External mail.



On 11/8/2024 12:11 PM, Howard Wang wrote:
> Implement the rtl_hw_config function to configure the hardware.
>
> Signed-off-by: Howard Wang 
>

<...>

> +void
> +rtl_nic_reset(struct rtl_hw *hw)
> +{
> + int i;
> +
> + rtl_disable_rx_packet_filter(hw);
> +
> + rtl_enable_rxdvgate(hw);
> +
> + rtl_stop_all_request(hw);
> +
> + rtl_wait_txrx_fifo_empty(hw);
> +
> + rte_delay_ms(2);
> +
> + /* Soft reset the chip. */
> + RTL_W8(hw, ChipCmd, CmdReset);
> +
> + /* Check that the chip has finished the reset. */
> + for (i = 100; i > 0; i--) {
> + rte_delay_us(100);
> + if ((RTL_R8(hw, ChipCmd) & CmdReset) == 0)
> + break;
> + }
> +}
>

Can you please make functions static as much as possible? Like above 
'rtl_nic_reset()' seems can be static.
That is OK keep as it is if functions will be called later in the set.

Re: [PATCH v6 11/20] net/ice: remove use of VLAs

2024-11-11 Thread Burakov, Anatoly


On 09/11/2024 02:48, Andre Muezerie wrote:

From: Konstantin Ananyev 

../drivers/net/ice/ice_rxtx.c:1871:29: warning:
 variable length array used [-Wvla]

Here VLA is used as a temp array for mbufs that will be used as a split
RX data buffers.
As at any given time only one thread can do RX from particular queue,
at rx_queue_setup() we can allocate extra space for that array, and then
safely use it at RX fast-path.

Signed-off-by: Konstantin Ananyev 
---

Acked-by: Anatoly Burakov 

--
Thanks,
Anatoly

Re: 答复: [PATCH v6 08/17] net/r8169: add support for phy configuration

On 11/11/2024 8:18 AM, 王颢 wrote:
> Dear Ferruh,
> 
> Regarding the commit message, I will think about how to write it.
> 
> I’m sorry, but for the registers in the PMD that do not use macro 
> definitions, most of them are company confidential and their meanings will 
> not be disclosed to the public. This part will be maintained by us.
> 

I understand, but this is conflicting with open sourcing the driver,
after patch is merged it becomes responsibility of the community, of
course primarily we will rely on you to maintain the driver, but for the
cases like refactoring, or users fixing defects etc, community members
may needs to update the driver code, and this requires for us to
understand the code.

Overall, I think there is a value to have the driver in DPDK, so please
continue upstreaming, but please try to explain and clarify the code as
much as possible.


> Best Regards,
> Howard Wang
> 
> -邮件原件-
> 发件人: Ferruh Yigit  
> 发送时间: 2024年11月11日 8:17
> 收件人: 王颢 ; dev@dpdk.org
> 抄送: pro_nic_d...@realtek.com
> 主题: Re: [PATCH v6 08/17] net/r8169: add support for phy configuration
> 
> 
> External mail.
> 
> 
> 
> On 11/8/2024 12:11 PM, Howard Wang wrote:
>> This patch contains phy config, ephy config and so on.
>>
>> Signed-off-by: Howard Wang 
>>
> 
> <...>
> 
>> @@ -62,6 +62,12 @@ rtl_dev_start(struct rte_eth_dev *dev)
>>   struct rtl_hw *hw = &adapter->hw;
>>   int err;
>>
>> + rtl_powerup_pll(hw);
>> +
>> + rtl_hw_ephy_config(hw);
>> +
>> + rtl_hw_phy_config(hw);
>> +
>>   rtl_hw_config(hw);
>>
>>   /* Initialize transmission unit */ @@ -74,6 +80,8 @@ 
>> rtl_dev_start(struct rte_eth_dev *dev)
>>   goto error;
>>   }
>>
>> + rtl_mdio_write(hw, 0x1F, 0x);
>> +
>>
> 
> It is very hard know what code like above does, which makes harder for anyone 
> to read and understand your code, can you please either create macros for 
> registers or comment as much as possible, so others can improve/fix your code 
> when required.
>

Re: [PATCH v2 2/3] graph: add support for node free API

2024-11-11 Thread Huichao Cai

By the way, I don't know where to consult about this issue at the moment.
Please forgive me for discussing this issue here. Can we set up a chat box on 
this
page instead of having to comment through email or git send email? Many email 
clients
cannot set the In Reply To header or do not currently have a git send email 
environment,
which is not conducive to the timeliness and convenience of comments.

Re: [PATCH v6 10/20] net/ixgbe: remove use of VLAs

2024-11-11 Thread Burakov, Anatoly


On 09/11/2024 02:48, Andre Muezerie wrote:

From: Konstantin Ananyev 

1) ../drivers/net/ixgbe/ixgbe_ethdev.c:3556:46: warning:
 variable length array used [-Wvla]
2) ../drivers/net/ixgbe/ixgbe_ethdev.c:3739:23: warning:
 variable length array used [-Wvla]
3) ../drivers/net/ixgbe/ixgbe_rxtx_vec_common.h:17:24: warning:
 variable length array used [-Wvla]

For first two cases: in fact ixgbe_xstats_calc_num() always returns
constant value, so it should be safe to replace that function invocation
just with a macro that performs same calculations.

For case #3: reassemble_packets() is invoked only by
ixgbe_recv_scattered_burst_vec().
And in turn, ixgbe_recv_scattered_burst_vec() operates only on fixed
max amount of packets per call: RTE_IXGBE_MAX_RX_BURST.
So, it should be safe to replace VLA with fixed size array.

Signed-off-by: Konstantin Ananyev 
---

Acked-by: Anatoly Burakov 

--
Thanks,
Anatoly

答复: 答复: [PATCH v6 01/17] net/r8169: add PMD driver skeleton

2024-11-11 Thread 王颢

Dear 

Ahh, I see, matching is simple for 8125, device id and product number is same 
:), is it same for all supported devices?

===> Yes, up until now, the Linux inbox driver has been doing it this way too.

I am not sure if there is a misunderstanding but change requests are mostly 
simple ones, I expect they can be fixed in one or two days, and I am hopping to 
merge the driver in -rc2 or -rc3 at worst.
===> I will do it as soon as possible. I have had a fever recently, so it might 
be a bit slow.

Best Regards,
Howard Wang

-邮件原件-
发件人: Ferruh Yigit  
发送时间: 2024年11月11日 17:09
收件人: 王颢 ; dev@dpdk.org
抄送: pro_nic_d...@realtek.com
主题: Re: 答复: [PATCH v6 01/17] net/r8169: add PMD driver skeleton


External mail.



On 11/11/2024 3:46 AM, 王颢 wrote:
> Dear Ferruh,
>
> I apologize for the issues that have arisen during the process of splitting a 
> complete driver into several small patches.
>
> On 11/8/2024 12:11 PM, Howard Wang wrote:
>> Meson build infrastructure, r8169_ethdev minimal skeleton, header 
>> with Realtek NIC device and vendor IDs.
>>
>> Signed-off-by: Howard Wang 
>
> <...>
>
> ===> I referred to Atlantic’s approach for this, and as a matter of fact, his 
> first commit message was done in this manner.
> I will think again about how it should be modified.
>

No problem with the commit splits and orders, thanks for the update there.

Only comment in this patch is 'RTL_DEV_PRIVATE' macro is not used in the patch 
and it can be added when you use the macro, so this is a simple issue.

Also I have a comment that build fails after some patchs, this also does not 
require changing how patch split done, just fix build after each patch.

I am not sure if there is a misunderstanding but change requests are mostly 
simple ones, I expect they can be fixed in one or two days, and I am hopping to 
merge the driver in -rc2 or -rc3 at worst.

> net/atlantic: add PMD driver skeleton
> Makefile/meson build infrastructure, atl_ethdev minimal skeleton, 
> header with aquantia aQtion NIC device and vendor IDs.
>
> Signed-off-by: Igor Russkikh 
> Signed-off-by: Pavel Belous 
>
>
>
> What do you think to add macros for the PCI device IDs, naming them makes it 
> easier to know which devices are supported.
>
> ===> Regarding the device ID, for 0x8125, the supported device is indeed 
> called 8125. So, should we name a macro like DEVICE_ID_RTL8125 to replace it?
>


Ahh, I see, matching is simple for 8125, device id and product number is same 
:), is it same for all supported devices?
{ RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8125) },
{ RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8162) },
{ RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8126) },
{ RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x5000) },

>
> Best regards,
> Howard Wang
>
> -邮件原件-
> 发件人: Ferruh Yigit 
> 发送时间: 2024年11月11日 8:15
> 收件人: 王颢 ; dev@dpdk.org
> 抄送: pro_nic_d...@realtek.com
> 主题: Re: [PATCH v6 01/17] net/r8169: add PMD driver skeleton
>
>
> External mail.
>
>
>
> On 11/8/2024 12:11 PM, Howard Wang wrote:
>> Meson build infrastructure, r8169_ethdev minimal skeleton, header 
>> with Realtek NIC device and vendor IDs.
>>
>> Signed-off-by: Howard Wang 
>
> <...>
>
>> +/*
>> + * The set of PCI devices this driver supports  */ static const 
>> +struct rte_pci_id pci_id_r8169_map[] = {
>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8125) },
>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8162) },
>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8126) },
>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x5000) },
>>
>
> What do you think to add macros for the PCI device IDs, naming them makes it 
> easier to know which devices are supported.
>
> <...>
>
>> +
>> +#define RTL_DEV_PRIVATE(eth_dev) \
>> + ((struct rtl_adapter *)((eth_dev)->data->dev_private))
>> +
>>
>
> This macro is only used a few patches later, what do you think to add this 
> macro when used?
>
>

[RESEND PATCH v15 1/3] power: introduce PM QoS API on CPU wide

The deeper the idle state, the lower the power consumption, but the longer
the resume time. Some service are delay sensitive and very except the low
resume time, like interrupt packet receiving mode.

And the "/sys/devices/system/cpu/cpuX/power/pm_qos_resume_latency_us" sysfs
interface is used to set and get the resume latency limit on the cpuX for
userspace. Each cpuidle governor in Linux select which idle state to enter
based on this CPU resume latency in their idle task.

The per-CPU PM QoS API can be used to control this CPU's idle state
selection and limit just enter the shallowest idle state to low the delay
when wake up from by setting strict resume latency (zero value).

Signed-off-by: Huisong Li 
Acked-by: Morten Brørup 
Acked-by: Chengwen Feng 
Acked-by: Konstantin Ananyev 
Acked-by: Sivaprasad Tummala 
---
 doc/guides/prog_guide/power_man.rst|  19 
 doc/guides/rel_notes/release_24_11.rst |   5 +
 lib/power/meson.build  |   2 +
 lib/power/rte_power_qos.c  | 123 +
 lib/power/rte_power_qos.h  |  73 +++
 lib/power/version.map  |   4 +
 6 files changed, 226 insertions(+)
 create mode 100644 lib/power/rte_power_qos.c
 create mode 100644 lib/power/rte_power_qos.h

diff --git a/doc/guides/prog_guide/power_man.rst 
b/doc/guides/prog_guide/power_man.rst
index 1ebab77ee9..ecae6b46ef 100644
--- a/doc/guides/prog_guide/power_man.rst
+++ b/doc/guides/prog_guide/power_man.rst
@@ -107,6 +107,25 @@ User Cases
 The power management mechanism is used to save power when performing L3 
forwarding.
 
 
+PM QoS
+--
+
+The "/sys/devices/system/cpu/cpuX/power/pm_qos_resume_latency_us" sysfs
+interface is used to set and get the resume latency limit on the cpuX for
+userspace. Each cpuidle governor in Linux select which idle state to enter
+based on this CPU resume latency in their idle task.
+
+The deeper the idle state, the lower the power consumption, but the longer
+the resume time. Some service are latency sensitive and very except the low
+resume time, like interrupt packet receiving mode.
+
+Applications can set and get the CPU resume latency by the
+``rte_power_qos_set_cpu_resume_latency()`` and 
``rte_power_qos_get_cpu_resume_latency()``
+respectively. Applications can set a strict resume latency (zero value) by
+the ``rte_power_qos_set_cpu_resume_latency()`` to low the resume latency and
+get better performance (instead, the power consumption of platform may 
increase).
+
+
 Ethernet PMD Power Management API
 -
 
diff --git a/doc/guides/rel_notes/release_24_11.rst 
b/doc/guides/rel_notes/release_24_11.rst
index 543becba28..187e6823d7 100644
--- a/doc/guides/rel_notes/release_24_11.rst
+++ b/doc/guides/rel_notes/release_24_11.rst
@@ -276,6 +276,11 @@ New Features
   This field is used to pass an extra configuration settings such as ability
   to lookup IPv4 addresses in network byte order.
 
+* **Introduce per-CPU PM QoS interface.**
+
+  * Add per-CPU PM QoS interface to low the resume latency when wake up from
+idle state.
+
 * **Added new API to register telemetry endpoint callbacks with private 
arguments.**
 
   A new ``rte_telemetry_register_cmd_arg`` function is available to pass an 
opaque value to
diff --git a/lib/power/meson.build b/lib/power/meson.build
index 4f4dc19687..313aaa6701 100644
--- a/lib/power/meson.build
+++ b/lib/power/meson.build
@@ -16,6 +16,7 @@ sources = files(
 'rte_power_cpufreq.c',
 'rte_power_uncore.c',
 'rte_power_pmd_mgmt.c',
+'rte_power_qos.c',
 )
 headers = files(
 'power_cpufreq.h',
@@ -24,6 +25,7 @@ headers = files(
 'rte_power_guest_channel.h',
 'rte_power_pmd_mgmt.h',
 'rte_power_uncore.h',
+'rte_power_qos.h',
 )
 
 deps += ['timer', 'ethdev']
diff --git a/lib/power/rte_power_qos.c b/lib/power/rte_power_qos.c
new file mode 100644
index 00..4dd0532b36
--- /dev/null
+++ b/lib/power/rte_power_qos.c
@@ -0,0 +1,123 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2024 HiSilicon Limited
+ */
+
+#include 
+#include 
+#include 
+
+#include 
+#include 
+
+#include "power_common.h"
+#include "rte_power_qos.h"
+
+#define PM_QOS_SYSFILE_RESUME_LATENCY_US   \
+   "/sys/devices/system/cpu/cpu%u/power/pm_qos_resume_latency_us"
+
+#define PM_QOS_CPU_RESUME_LATENCY_BUF_LEN  32
+
+int
+rte_power_qos_set_cpu_resume_latency(uint16_t lcore_id, int latency)
+{
+   char buf[PM_QOS_CPU_RESUME_LATENCY_BUF_LEN];
+   uint32_t cpu_id;
+   FILE *f;
+   int ret;
+
+   if (!rte_lcore_is_enabled(lcore_id)) {
+   POWER_LOG(ERR, "lcore id %u is not enabled", lcore_id);
+   return -EINVAL;
+   }
+   ret = power_get_lcore_mapped_cpu_id(lcore_id, &cpu_id);
+   if (ret != 0)
+   return ret;
+
+   if (latency < 0) {
+   POWER_LOG(ERR, "latency should be greater than and equal to 0");
+

[RESEND PATCH v15 0/3] power: introduce PM QoS interface

The deeper the idle state, the lower the power consumption, but the longer
the resume time. Some service are delay sensitive and very except the low
resume time, like interrupt packet receiving mode.

And the "/sys/devices/system/cpu/cpuX/power/pm_qos_resume_latency_us" sysfs
interface is used to set and get the resume latency limit on the cpuX for
userspace. Please see the description in kernel document[1].
Each cpuidle governor in Linux select which idle state to enter based on
this CPU resume latency in their idle task.

The per-CPU PM QoS API can be used to control this CPU's idle state
selection and limit just enter the shallowest idle state to low the delay
when wake up from idle state by setting strict resume latency (zero value).

[1]
https://www.kernel.org/doc/html/latest/admin-guide/abi-testing.html?highlight=pm_qos_resume_latency_us#abi-sys-devices-power-pm-qos-resume-latency-us

---
v15:
- fix conflicts due to the new merged patches that rework power lib.
- add Acked-by: Konstantin Ananyev for patch 3/3.
v14:
- use parse_uint to parse --cpu-resume-latency instead of adding a new
parse_int()
v13:
- not allow negative value for --cpu-resume-latency.
- restore to the original value as Konstantin suggested.
v12:
- add Acked-by Chengwen and Konstantin
- fix overflow issue in l3fwd-power when parse command line
- add a command parameter to set CPU resume latency
v11:
- operate the cpu id the lcore mapped by the new function
power_get_lcore_mapped_cpu_id().
v10:
- replace LINE_MAX with a custom macro and fix two typos.
v9:
- move new feature description from release_24_07.rst to release_24_11.rst.
v8:
- update the latest code to resolve CI warning
v7:
- remove a dead code rte_lcore_is_enabled in patch[2/2]
v6:
- update release_24_07.rst based on dpdk repo to resolve CI warning.
v5:
- use LINE_MAX to replace BUFSIZ, and use snprintf to replace sprintf.
v4:
- fix some comments basd on Stephen
- add stdint.h include
- add Acked-by Morten Brørup
v3:
- add RTE_POWER_xxx prefix for some macro in header
- add the check for lcore_id with rte_lcore_is_enabled
v2:
- use PM QoS on CPU wide to replace the one on system wide

Huisong Li (3):
power: introduce PM QoS API on CPU wide
examples/l3fwd-power: fix data overflow when parse command line
examples/l3fwd-power: add PM QoS configuration

doc/guides/prog_guide/power_man.rst | 19 +++
doc/guides/rel_notes/release_24_11.rst| 5 +
.../sample_app_ug/l3_forward_power_man.rst| 5 +-
examples/l3fwd-power/main.c | 96 +++---
lib/power/meson.build | 2 +
lib/power/rte_power_qos.c | 123 ++
lib/power/rte_power_qos.h | 73 +++
lib/power/version.map | 4 +
8 files changed, 306 insertions(+), 21 deletions(-)
create mode 100644 lib/power/rte_power_qos.c
create mode 100644 lib/power/rte_power_qos.h

--
2.22.0

[RESEND PATCH v15 3/3] examples/l3fwd-power: add PM QoS configuration

The '--cpu-resume-latency' can use to control C-state selection.
Setting the CPU resume latency to 0 can limit the CPU just to enter
C0-state to improve performance, which also may increase the power
consumption of platform.

Signed-off-by: Huisong Li 
Acked-by: Morten Brørup 
Acked-by: Chengwen Feng 
Acked-by: Konstantin Ananyev 
---
 .../sample_app_ug/l3_forward_power_man.rst|  5 +-
 examples/l3fwd-power/main.c   | 55 +++
 2 files changed, 59 insertions(+), 1 deletion(-)

diff --git a/doc/guides/sample_app_ug/l3_forward_power_man.rst 
b/doc/guides/sample_app_ug/l3_forward_power_man.rst
index 9c9684fea7..70fa83669a 100644
--- a/doc/guides/sample_app_ug/l3_forward_power_man.rst
+++ b/doc/guides/sample_app_ug/l3_forward_power_man.rst
@@ -67,7 +67,8 @@ based on the speculative sleep duration of the core.
 In this application, we introduce a heuristic algorithm that allows packet 
processing cores to sleep for a short period
 if there is no Rx packet received on recent polls.
 In this way, CPUIdle automatically forces the corresponding cores to enter 
deeper C-states
-instead of always running to the C0 state waiting for packets.
+instead of always running to the C0 state waiting for packets. But user can 
set the CPU resume latency to control C-state selection.
+Setting the CPU resume latency to 0 can limit the CPU just to enter C0-state 
to improve performance, which may increase power consumption of platform.
 
 .. note::
 
@@ -105,6 +106,8 @@ where,
 
 *   --config (port,queue,lcore)[,(port,queue,lcore)]: determines which queues 
from which ports are mapped to which cores.
 
+*   --cpu-resume-latency LATENCY: set CPU resume latency to control C-state 
selection, 0 : just allow to enter C0-state.
+
 *   --max-pkt-len: optional, maximum packet length in decimal (64-9600)
 
 *   --no-numa: optional, disables numa awareness
diff --git a/examples/l3fwd-power/main.c b/examples/l3fwd-power/main.c
index 7bc524aa16..ae8b55924e 100644
--- a/examples/l3fwd-power/main.c
+++ b/examples/l3fwd-power/main.c
@@ -47,6 +47,7 @@
 #include 
 #include 
 #include 
+#include 
 
 #include "perf_core.h"
 #include "main.h"
@@ -265,6 +266,9 @@ static uint32_t pause_duration = 1;
 static uint32_t scale_freq_min;
 static uint32_t scale_freq_max;
 
+static int cpu_resume_latency = -1;
+static int resume_latency_bk[RTE_MAX_LCORE];
+
 static struct rte_mempool * pktmbuf_pool[NB_SOCKETS];
 
 
@@ -1497,6 +1501,8 @@ print_usage(const char *prgname)
"  -U: set min/max frequency for uncore to maximum value\n"
"  -i (frequency index): set min/max frequency for uncore to 
specified frequency index\n"
"  --config (port,queue,lcore): rx queues configuration\n"
+   "  --cpu-resume-latency LATENCY: set CPU resume latency to 
control C-state selection,"
+   " 0 : just allow to enter C0-state\n"
"  --high-perf-cores CORELIST: list of high performance cores\n"
"  --perf-config: similar as config, cores specified as indices"
" for bins containing high or regular performance cores\n"
@@ -1735,6 +1741,7 @@ parse_pmd_mgmt_config(const char *name)
 #define CMD_LINE_OPT_PAUSE_DURATION "pause-duration"
 #define CMD_LINE_OPT_SCALE_FREQ_MIN "scale-freq-min"
 #define CMD_LINE_OPT_SCALE_FREQ_MAX "scale-freq-max"
+#define CMD_LINE_OPT_CPU_RESUME_LATENCY "cpu-resume-latency"
 
 /* Parse the argument given in the command line of the application */
 static int
@@ -1749,6 +1756,7 @@ parse_args(int argc, char **argv)
{"perf-config", 1, 0, 0},
{"high-perf-cores", 1, 0, 0},
{"no-numa", 0, 0, 0},
+   {CMD_LINE_OPT_CPU_RESUME_LATENCY, 1, 0, 0},
{CMD_LINE_OPT_MAX_PKT_LEN, 1, 0, 0},
{CMD_LINE_OPT_PARSE_PTYPE, 0, 0, 0},
{CMD_LINE_OPT_LEGACY, 0, 0, 0},
@@ -1934,6 +1942,15 @@ parse_args(int argc, char **argv)
printf("Scaling frequency maximum 
configured\n");
}
 
+   if (!strncmp(lgopts[option_index].name,
+   CMD_LINE_OPT_CPU_RESUME_LATENCY,
+   
sizeof(CMD_LINE_OPT_CPU_RESUME_LATENCY))) {
+   if (parse_uint(optarg, INT_MAX,
+   (uint32_t 
*)&cpu_resume_latency) != 0)
+   return -1;
+   printf("PM QoS configured\n");
+   }
+
break;
 
default:
@@ -2257,6 +2274,35 @@ init_power_library(void)
return -1;
}
}
+
+   if (cpu_resume_latency != -1) {
+   RTE_LCORE_FOREACH(lcore_id) {
+   /* Back old CPU resume latency. */
+   ret = rte_power_qos_get_cpu_resume_latency(lcore_id);
+

Re: 答复: 答复: [PATCH v6 01/17] net/r8169: add PMD driver skeleton

On 11/11/2024 9:24 AM, 王颢 wrote:
> Dear 
> 
> Ahh, I see, matching is simple for 8125, device id and product number is same 
> :), is it same for all supported devices?
> 
> ===> Yes, up until now, the Linux inbox driver has been doing it this way too.
> 

Got it, so OK to keep the code as it is, thanks for the clarification.

> I am not sure if there is a misunderstanding but change requests are mostly 
> simple ones, I expect they can be fixed in one or two days, and I am hopping 
> to merge the driver in -rc2 or -rc3 at worst.
> ===> I will do it as soon as possible. I have had a fever recently, so it 
> might be a bit slow.
>

Great, thanks. (And I hope you get better soon ;)

> 
> Best Regards,
> Howard Wang
> 
> -邮件原件-
> 发件人: Ferruh Yigit  
> 发送时间: 2024年11月11日 17:09
> 收件人: 王颢 ; dev@dpdk.org
> 抄送: pro_nic_d...@realtek.com
> 主题: Re: 答复: [PATCH v6 01/17] net/r8169: add PMD driver skeleton
> 
> 
> External mail.
> 
> 
> 
> On 11/11/2024 3:46 AM, 王颢 wrote:
>> Dear Ferruh,
>>
>> I apologize for the issues that have arisen during the process of splitting 
>> a complete driver into several small patches.
>>
>> On 11/8/2024 12:11 PM, Howard Wang wrote:
>>> Meson build infrastructure, r8169_ethdev minimal skeleton, header 
>>> with Realtek NIC device and vendor IDs.
>>>
>>> Signed-off-by: Howard Wang 
>>
>> <...>
>>
>> ===> I referred to Atlantic’s approach for this, and as a matter of fact, 
>> his first commit message was done in this manner.
>> I will think again about how it should be modified.
>>
> 
> No problem with the commit splits and orders, thanks for the update there.
> 
> Only comment in this patch is 'RTL_DEV_PRIVATE' macro is not used in the 
> patch and it can be added when you use the macro, so this is a simple issue.
> 
> Also I have a comment that build fails after some patchs, this also does not 
> require changing how patch split done, just fix build after each patch.
> 
> I am not sure if there is a misunderstanding but change requests are mostly 
> simple ones, I expect they can be fixed in one or two days, and I am hopping 
> to merge the driver in -rc2 or -rc3 at worst.
> 
>> net/atlantic: add PMD driver skeleton
>> Makefile/meson build infrastructure, atl_ethdev minimal skeleton, 
>> header with aquantia aQtion NIC device and vendor IDs.
>>
>> Signed-off-by: Igor Russkikh 
>> Signed-off-by: Pavel Belous 
>>
>>
>>
>> What do you think to add macros for the PCI device IDs, naming them makes it 
>> easier to know which devices are supported.
>>
>> ===> Regarding the device ID, for 0x8125, the supported device is indeed 
>> called 8125. So, should we name a macro like DEVICE_ID_RTL8125 to replace it?
>>
> 
> 
> Ahh, I see, matching is simple for 8125, device id and product number is same 
> :), is it same for all supported devices?
> { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8125) },
> { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8162) },
> { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8126) },
> { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x5000) },
> 
>>
>> Best regards,
>> Howard Wang
>>
>> -邮件原件-
>> 发件人: Ferruh Yigit 
>> 发送时间: 2024年11月11日 8:15
>> 收件人: 王颢 ; dev@dpdk.org
>> 抄送: pro_nic_d...@realtek.com
>> 主题: Re: [PATCH v6 01/17] net/r8169: add PMD driver skeleton
>>
>>
>> External mail.
>>
>>
>>
>> On 11/8/2024 12:11 PM, Howard Wang wrote:
>>> Meson build infrastructure, r8169_ethdev minimal skeleton, header 
>>> with Realtek NIC device and vendor IDs.
>>>
>>> Signed-off-by: Howard Wang 
>>
>> <...>
>>
>>> +/*
>>> + * The set of PCI devices this driver supports  */ static const 
>>> +struct rte_pci_id pci_id_r8169_map[] = {
>>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8125) },
>>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8162) },
>>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8126) },
>>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x5000) },
>>>
>>
>> What do you think to add macros for the PCI device IDs, naming them makes it 
>> easier to know which devices are supported.
>>
>> <...>
>>
>>> +
>>> +#define RTL_DEV_PRIVATE(eth_dev) \
>>> + ((struct rtl_adapter *)((eth_dev)->data->dev_private))
>>> +
>>>
>>
>> This macro is only used a few patches later, what do you think to add this 
>> macro when used?
>>
>>
>

Re: Bus scan is too chatty

10/11/2024 20:17, Stephen Hemminger:
> DPDK EAL needs to be less verbose for non-existent buses.
> Most users won't have embedded devices, and messages like:
> 
> EAL: Scan for (fslmc) bus failed.
> 
> will be confusing.

When merging the recent fslmc patches, I've reduced the verbosity
of some fslmc logs appearing at the cleanup stage.

This log is from EAL because fslmc scan fails.
It should not return an error if there is no bus or no environment variable.

Sachin, Hemant, please could you fix it?

[RESEND PATCH v15 2/3] examples/l3fwd-power: fix data overflow when parse command line

Many variables are 'uint32_t', like, 'pause_duration', 'scale_freq_min'
and so on. They use parse_int() to parse it from command line.
But overflow problem occurs when this function return.

Fixes: 59f2853c4cae ("examples/l3fwd_power: add configuration options")
Cc: sta...@dpdk.org

Signed-off-by: Huisong Li 
Acked-by: Konstantin Ananyev 
Acked-by: Chengwen Feng 
Acked-by: Sivaprasad Tummala 
---
 examples/l3fwd-power/main.c | 41 +++--
 1 file changed, 21 insertions(+), 20 deletions(-)

diff --git a/examples/l3fwd-power/main.c b/examples/l3fwd-power/main.c
index 272e069207..7bc524aa16 100644
--- a/examples/l3fwd-power/main.c
+++ b/examples/l3fwd-power/main.c
@@ -1520,8 +1520,12 @@ print_usage(const char *prgname)
prgname);
 }
 
+/*
+ * Caller must give the right upper limit so as to ensure receiver variable
+ * doesn't overflow.
+ */
 static int
-parse_int(const char *opt)
+parse_uint(const char *opt, uint32_t max, uint32_t *res)
 {
char *end = NULL;
unsigned long val;
@@ -1531,23 +1535,15 @@ parse_int(const char *opt)
if ((opt[0] == '\0') || (end == NULL) || (*end != '\0'))
return -1;
 
-   return val;
-}
-
-static int parse_max_pkt_len(const char *pktlen)
-{
-   char *end = NULL;
-   unsigned long len;
-
-   /* parse decimal string */
-   len = strtoul(pktlen, &end, 10);
-   if ((pktlen[0] == '\0') || (end == NULL) || (*end != '\0'))
+   if (val > max) {
+   RTE_LOG(ERR, L3FWD_POWER, "%s parameter shouldn't exceed %u.\n",
+   opt, max);
return -1;
+   }
 
-   if (len == 0)
-   return -1;
+   *res = val;
 
-   return len;
+   return 0;
 }
 
 static int
@@ -1894,8 +1890,9 @@ parse_args(int argc, char **argv)
if (!strncmp(lgopts[option_index].name,
CMD_LINE_OPT_MAX_PKT_LEN,
sizeof(CMD_LINE_OPT_MAX_PKT_LEN))) {
+   if (parse_uint(optarg, UINT32_MAX, 
&max_pkt_len) != 0)
+   return -1;
printf("Custom frame size is configured\n");
-   max_pkt_len = parse_max_pkt_len(optarg);
}
 
if (!strncmp(lgopts[option_index].name,
@@ -1908,29 +1905,33 @@ parse_args(int argc, char **argv)
if (!strncmp(lgopts[option_index].name,
CMD_LINE_OPT_MAX_EMPTY_POLLS,
sizeof(CMD_LINE_OPT_MAX_EMPTY_POLLS))) {
+   if (parse_uint(optarg, UINT32_MAX, 
&max_empty_polls) != 0)
+   return -1;
printf("Maximum empty polls configured\n");
-   max_empty_polls = parse_int(optarg);
}
 
if (!strncmp(lgopts[option_index].name,
CMD_LINE_OPT_PAUSE_DURATION,
sizeof(CMD_LINE_OPT_PAUSE_DURATION))) {
+   if (parse_uint(optarg, UINT32_MAX, 
&pause_duration) != 0)
+   return -1;
printf("Pause duration configured\n");
-   pause_duration = parse_int(optarg);
}
 
if (!strncmp(lgopts[option_index].name,
CMD_LINE_OPT_SCALE_FREQ_MIN,
sizeof(CMD_LINE_OPT_SCALE_FREQ_MIN))) {
+   if (parse_uint(optarg, UINT32_MAX, 
&scale_freq_min) != 0)
+   return -1;
printf("Scaling frequency minimum 
configured\n");
-   scale_freq_min = parse_int(optarg);
}
 
if (!strncmp(lgopts[option_index].name,
CMD_LINE_OPT_SCALE_FREQ_MAX,
sizeof(CMD_LINE_OPT_SCALE_FREQ_MAX))) {
+   if (parse_uint(optarg, UINT32_MAX, 
&scale_freq_max) != 0)
+   return -1;
printf("Scaling frequency maximum 
configured\n");
-   scale_freq_max = parse_int(optarg);
}
 
break;
-- 
2.22.0

RE: Bus scan is too chatty

2024-11-11 Thread Hemant Agrawal


> -Original Message-
> From: Thomas Monjalon 
> Sent: Monday, November 11, 2024 3:02 PM
> To: Stephen Hemminger ; Gagandeep Singh
> ; Sachin Saxena (OSS) ;
> Hemant Agrawal 
> Cc: dev@dpdk.org
> Subject: Re: Bus scan is too chatty
> Importance: High
> 
> 10/11/2024 20:17, Stephen Hemminger:
> > DPDK EAL needs to be less verbose for non-existent buses.
> > Most users won't have embedded devices, and messages like:
> >
> > EAL: Scan for (fslmc) bus failed.
> >
> > will be confusing.
> 
> When merging the recent fslmc patches, I've reduced the verbosity of some
> fslmc logs appearing at the cleanup stage.
> 
> This log is from EAL because fslmc scan fails.
> It should not return an error if there is no bus or no environment variable.
> 
> Sachin, Hemant, please could you fix it?
> 
[Hemant] yes, let me have a look and provide a fix.

Re: [PATCH v2 2/3] graph: add support for node free API

2024-11-11 Thread Huichao Cai

Do these newly added functions need to protect the linked list with spin lock 
graph_stpinlock_lock?

Re: 答复: [PATCH v6 01/17] net/r8169: add PMD driver skeleton

On 11/11/2024 3:46 AM, 王颢 wrote:
> Dear Ferruh,
> 
> I apologize for the issues that have arisen during the process of splitting a 
> complete driver into several small patches.
> 
> On 11/8/2024 12:11 PM, Howard Wang wrote:
>> Meson build infrastructure, r8169_ethdev minimal skeleton, header with 
>> Realtek NIC device and vendor IDs.
>>
>> Signed-off-by: Howard Wang 
> 
> <...>
> 
> ===> I referred to Atlantic’s approach for this, and as a matter of fact, his 
> first commit message was done in this manner. 
> I will think again about how it should be modified.
> 

No problem with the commit splits and orders, thanks for the update there.

Only comment in this patch is 'RTL_DEV_PRIVATE' macro is not used in the
patch and it can be added when you use the macro, so this is a simple issue.

Also I have a comment that build fails after some patchs, this also does
not require changing how patch split done, just fix build after each patch.

I am not sure if there is a misunderstanding but change requests are
mostly simple ones, I expect they can be fixed in one or two days, and I
am hopping to merge the driver in -rc2 or -rc3 at worst.

> net/atlantic: add PMD driver skeleton
> Makefile/meson build infrastructure, atl_ethdev minimal skeleton,
> header with aquantia aQtion NIC device and vendor IDs.
> 
> Signed-off-by: Igor Russkikh 
> Signed-off-by: Pavel Belous 
> 
> 
> 
> What do you think to add macros for the PCI device IDs, naming them makes it 
> easier to know which devices are supported.
> 
> ===> Regarding the device ID, for 0x8125, the supported device is indeed 
> called 8125. So, should we name a macro like DEVICE_ID_RTL8125 to replace it?
>


Ahh, I see, matching is simple for 8125, device id and product number is
same :), is it same for all supported devices?
{ RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8125) },
{ RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8162) },
{ RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8126) },
{ RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x5000) },

> 
> Best regards,
> Howard Wang
> 
> -邮件原件-
> 发件人: Ferruh Yigit  
> 发送时间: 2024年11月11日 8:15
> 收件人: 王颢 ; dev@dpdk.org
> 抄送: pro_nic_d...@realtek.com
> 主题: Re: [PATCH v6 01/17] net/r8169: add PMD driver skeleton
> 
> 
> External mail.
> 
> 
> 
> On 11/8/2024 12:11 PM, Howard Wang wrote:
>> Meson build infrastructure, r8169_ethdev minimal skeleton, header with 
>> Realtek NIC device and vendor IDs.
>>
>> Signed-off-by: Howard Wang 
> 
> <...>
> 
>> +/*
>> + * The set of PCI devices this driver supports  */ static const 
>> +struct rte_pci_id pci_id_r8169_map[] = {
>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8125) },
>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8162) },
>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x8126) },
>> + { RTE_PCI_DEVICE(PCI_VENDOR_ID_REALTEK, 0x5000) },
>>
> 
> What do you think to add macros for the PCI device IDs, naming them makes it 
> easier to know which devices are supported.
> 
> <...>
> 
>> +
>> +#define RTL_DEV_PRIVATE(eth_dev) \
>> + ((struct rtl_adapter *)((eth_dev)->data->dev_private))
>> +
>>
> 
> This macro is only used a few patches later, what do you think to add this 
> macro when used?
> 
>

Re: [PATCH v15 0/3] power: introduce PM QoS interface

11/11/2024 03:25, Huisong Li:
> The deeper the idle state, the lower the power consumption, but the longer
> the resume time. Some service are delay sensitive and very except the low
> resume time, like interrupt packet receiving mode.
> 
> And the "/sys/devices/system/cpu/cpuX/power/pm_qos_resume_latency_us" sysfs
> interface is used to set and get the resume latency limit on the cpuX for
> userspace. Please see the description in kernel document[1].
> Each cpuidle governor in Linux select which idle state to enter based on
> this CPU resume latency in their idle task.
> 
> The per-CPU PM QoS API can be used to control this CPU's idle state
> selection and limit just enter the shallowest idle state to low the delay
> when wake up from idle state by setting strict resume latency (zero value).
> 
> [1] 
> https://www.kernel.org/doc/html/latest/admin-guide/abi-testing.html?highlight=pm_qos_resume_latency_us#abi-sys-devices-power-pm-qos-resume-latency-us

Applied, thanks.

[PATCH v9 2/7] ring: common functions for 'move head' ops

Note upfront: that change doesn't introduce any functional or
performance changes.
It is just a code-reordering for:
 - code deduplication
 - ability in future to re-use the same code to introduce new functionality

For each sync mode corresponding move_prod_head() and
move_cons_head() are nearly identical to each other,
the only differences are:
 - do we need to use a @capacity to calculate number of entries or not.
 - what we need to update (prod/cons) and what is used as
   read-only counterpart.
So instead of having 2 copies of nearly identical functions,
introduce a new common one that could be used by both functions:
move_prod_head() and move_cons_head().

As another positive thing - we can get rid of referencing whole rte_ring
structure in that new common sub-function.

Signed-off-by: Konstantin Ananyev 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 lib/ring/rte_ring_c11_pvt.h  | 156 ++-
 lib/ring/rte_ring_elem_pvt.h |  66 +
 lib/ring/rte_ring_generic_pvt.h  | 143 +---
 lib/ring/rte_ring_hts_elem_pvt.h | 107 ++---
 lib/ring/rte_ring_rts_elem_pvt.h | 107 ++---
 5 files changed, 255 insertions(+), 324 deletions(-)

diff --git a/lib/ring/rte_ring_c11_pvt.h b/lib/ring/rte_ring_c11_pvt.h
index 629b2d9288..b9388af0da 100644
--- a/lib/ring/rte_ring_c11_pvt.h
+++ b/lib/ring/rte_ring_c11_pvt.h
@@ -11,6 +11,17 @@
 #ifndef _RTE_RING_C11_PVT_H_
 #define _RTE_RING_C11_PVT_H_
 
+/**
+ * @file rte_ring_c11_pvt.h
+ * It is not recommended to include this file directly,
+ * include  instead.
+ * Contains internal helper functions for MP/SP and MC/SC ring modes.
+ * For more information please refer to .
+ */
+
+/**
+ * @internal This function updates tail values.
+ */
 static __rte_always_inline void
 __rte_ring_update_tail(struct rte_ring_headtail *ht, uint32_t old_val,
uint32_t new_val, uint32_t single, uint32_t enqueue)
@@ -29,40 +40,45 @@ __rte_ring_update_tail(struct rte_ring_headtail *ht, 
uint32_t old_val,
 }
 
 /**
- * @internal This function updates the producer head for enqueue
+ * @internal This is a helper function that moves the producer/consumer head
  *
- * @param r
- *   A pointer to the ring structure
- * @param is_sp
- *   Indicates whether multi-producer path is needed or not
+ * @param d
+ *   A pointer to the headtail structure with head value to be moved
+ * @param s
+ *   A pointer to the counter-part headtail structure. Note that this
+ *   function only reads tail value from it
+ * @param capacity
+ *   Either ring capacity value (for producer), or zero (for consumer)
+ * @param is_st
+ *   Indicates whether multi-thread safe path is needed or not
  * @param n
- *   The number of elements we will want to enqueue, i.e. how far should the
- *   head be moved
+ *   The number of elements we want to move head value on
  * @param behavior
- *   RTE_RING_QUEUE_FIXED:Enqueue a fixed number of items from a ring
- *   RTE_RING_QUEUE_VARIABLE: Enqueue as many items as possible from ring
+ *   RTE_RING_QUEUE_FIXED:Move on a fixed number of items
+ *   RTE_RING_QUEUE_VARIABLE: Move on as many items as possible
  * @param old_head
- *   Returns head value as it was before the move, i.e. where enqueue starts
+ *   Returns head value as it was before the move
  * @param new_head
- *   Returns the current/new head value i.e. where enqueue finishes
- * @param free_entries
- *   Returns the amount of free space in the ring BEFORE head was moved
+ *   Returns the new head value
+ * @param entries
+ *   Returns the number of ring entries available BEFORE head was moved
  * @return
- *   Actual number of objects enqueued.
- *   If behavior == RTE_RING_QUEUE_FIXED, this will be 0 or n only.
+ *   Actual number of objects the head was moved on
+ *   If behavior == RTE_RING_QUEUE_FIXED, this will be 0 or n only
  */
 static __rte_always_inline unsigned int
-__rte_ring_move_prod_head(struct rte_ring *r, unsigned int is_sp,
-   unsigned int n, enum rte_ring_queue_behavior behavior,
-   uint32_t *old_head, uint32_t *new_head,
-   uint32_t *free_entries)
+__rte_ring_headtail_move_head(struct rte_ring_headtail *d,
+   const struct rte_ring_headtail *s, uint32_t capacity,
+   unsigned int is_st, unsigned int n,
+   enum rte_ring_queue_behavior behavior,
+   uint32_t *old_head, uint32_t *new_head, uint32_t *entries)
 {
-   const uint32_t capacity = r->capacity;
-   uint32_t cons_tail;
-   unsigned int max = n;
+   uint32_t stail;
int success;
+   unsigned int max = n;
 
-   *old_head = rte_atomic_load_explicit(&r->prod.head, 
rte_memory_order_relaxed);
+   *old_head = rte_atomic_load_explicit(&d->head,
+   rte_memory_order_relaxed);
do {
/* Reset n to the initial burst count */
n = max;
@@ -73,112 +89,36 @@

[PATCH v9 4/7] ring: make dump function more verbose

From: Eimear Morrissey 

The current rte_ring_dump function uses the generic rte_ring_headtail
structure to access head/tail positions. This is incorrect for the RTS
case where the head is stored in a different offset in the union of
structs. Switching to a separate function for each sync type allows
to dump correct head/tail values and extra metadata.

Signed-off-by: Eimear Morrissey 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 .mailmap |  1 +
 app/test/test_ring_stress_impl.h |  1 +
 lib/ring/rte_ring.c  | 87 ++--
 lib/ring/rte_ring.h  | 15 ++
 lib/ring/version.map |  7 +++
 5 files changed, 107 insertions(+), 4 deletions(-)

diff --git a/.mailmap b/.mailmap
index 49734b899f..a22d28989c 100644
--- a/.mailmap
+++ b/.mailmap
@@ -388,6 +388,7 @@ Eduard Serra 
 Edward Makarov 
 Edwin Brossette 
 Eelco Chaudron 
+Eimear Morrissey 
 Elad Nachman 
 Elad Persiko 
 Elena Agostini 
diff --git a/app/test/test_ring_stress_impl.h b/app/test/test_ring_stress_impl.h
index ee5274aeef..f99a7ff675 100644
--- a/app/test/test_ring_stress_impl.h
+++ b/app/test/test_ring_stress_impl.h
@@ -380,6 +380,7 @@ test_mt1(int (*test)(void *))
}
 
lcore_stat_dump(stdout, UINT32_MAX, &arg[mc].stats);
+   rte_ring_dump(stdout, r);
mt1_fini(r, data);
return rc;
 }
diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c
index aebb6d6728..a7d9d6b037 100644
--- a/lib/ring/rte_ring.c
+++ b/lib/ring/rte_ring.c
@@ -364,20 +364,99 @@ rte_ring_free(struct rte_ring *r)
rte_free(te);
 }
 
+static const char *
+ring_get_sync_type(const enum rte_ring_sync_type st)
+{
+   switch (st) {
+   case RTE_RING_SYNC_ST:
+   return "ST";
+   case RTE_RING_SYNC_MT:
+   return "MT";
+   case RTE_RING_SYNC_MT_RTS:
+   return "MT_RTS";
+   case RTE_RING_SYNC_MT_HTS:
+   return "MT_HTS";
+   default:
+   return "unknown";
+   }
+}
+
+static void
+ring_dump_ht_headtail(FILE *f, const char *prefix,
+   const struct rte_ring_headtail *ht)
+{
+   fprintf(f, "%ssync_type=%s\n", prefix,
+   ring_get_sync_type(ht->sync_type));
+   fprintf(f, "%shead=%"PRIu32"\n", prefix, ht->head);
+   fprintf(f, "%stail=%"PRIu32"\n", prefix, ht->tail);
+}
+
+static void
+ring_dump_rts_headtail(FILE *f, const char *prefix,
+   const struct rte_ring_rts_headtail *rts)
+{
+   fprintf(f, "%ssync_type=%s\n", prefix,
+   ring_get_sync_type(rts->sync_type));
+   fprintf(f, "%shead.pos=%"PRIu32"\n", prefix, rts->head.val.pos);
+   fprintf(f, "%shead.cnt=%"PRIu32"\n", prefix, rts->head.val.cnt);
+   fprintf(f, "%stail.pos=%"PRIu32"\n", prefix, rts->tail.val.pos);
+   fprintf(f, "%stail.cnt=%"PRIu32"\n", prefix, rts->tail.val.cnt);
+   fprintf(f, "%shtd_max=%"PRIu32"\n", prefix, rts->htd_max);
+}
+
+static void
+ring_dump_hts_headtail(FILE *f, const char *prefix,
+   const struct rte_ring_hts_headtail *hts)
+{
+   fprintf(f, "%ssync_type=%s\n", prefix,
+   ring_get_sync_type(hts->sync_type));
+   fprintf(f, "%shead=%"PRIu32"\n", prefix, hts->ht.pos.head);
+   fprintf(f, "%stail=%"PRIu32"\n", prefix, hts->ht.pos.tail);
+}
+
+void
+rte_ring_headtail_dump(FILE *f, const char *prefix,
+   const struct rte_ring_headtail *r)
+{
+   if (f == NULL || r == NULL)
+   return;
+
+   prefix = (prefix != NULL) ? prefix : "";
+
+   switch (r->sync_type) {
+   case RTE_RING_SYNC_ST:
+   case RTE_RING_SYNC_MT:
+   ring_dump_ht_headtail(f, prefix, r);
+   break;
+   case RTE_RING_SYNC_MT_RTS:
+   ring_dump_rts_headtail(f, prefix,
+   (const struct rte_ring_rts_headtail *)r);
+   break;
+   case RTE_RING_SYNC_MT_HTS:
+   ring_dump_hts_headtail(f, prefix,
+   (const struct rte_ring_hts_headtail *)r);
+   break;
+   default:
+   RING_LOG(ERR, "Invalid ring sync type detected");
+   }
+}
+
 /* dump the status of the ring on the console */
 void
 rte_ring_dump(FILE *f, const struct rte_ring *r)
 {
+   if (f == NULL || r == NULL)
+   return;
+
fprintf(f, "ring <%s>@%p\n", r->name, r);
fprintf(f, "  flags=%x\n", r->flags);
fprintf(f, "  size=%"PRIu32"\n", r->size);
fprintf(f, "  capacity=%"PRIu32"\n", r->capacity);
-   fprintf(f, "  ct=%"PRIu32"\n", r->cons.tail);
-   fprintf(f, "  ch=%"PRIu32"\n", r->cons.head);
-   fprintf(f, "  pt=%"PRIu32"\n", r->prod.tail);
-   fprintf(f, "  ph=%"PRIu32"\n", r->prod.head);
fprintf(f, "  used=%u\n", rte_ring_count(r));
fprintf(f, "  avail=%u\n", rte_ring_free_count(r));
+
+   rte_ring_headtail_dump(f, "  cons.", &(r->cons));
+

[PATCH v9 0/7] Stage-Ordered API and other extensions for ring library

Testing coverage (passed):
x86_64, i686, PPC, ARM   

Would like to express my gratitude to all community members who helped
with testing it on different platforms, in particular:
David Christensen 
Cody Cheng 
Patrick Robb 
Phanendra Vukkisala 
Chengwen Feng 

v8 -> v9
- fix grammar (Thomas)
- test: use stdatomic API (Thomas)

v7 -> v8
- fixed several nits (spotted by Morten):
  indentation, variable names, comments, etc.

v6 -> v7
- updated Programmer Guide (Jerin, Morten, Stephen)
- fix some functions in public headers without comments (Morten)
- update debug checks, added new macro for that: RTE_SORING_DEBUG
  (disabled by default).

v5 -> v6
- fix problem with ring_stress_autotest (Phanendra)
- added more checks and debug output

v4 -> v5
- fix public API/doc comments from Jerin
- update devtools/build-dict.sh (Stephen)
- fix MSVC warnings
- introduce new test-suite for meson (stress) with
  ring_stress_autotest and soring_stress_autotest in it
- enhance error report in tests
- reorder some sync code in soring and add extra checks
  (for better debuggability)

v3 -> v4:
- fix compilation/doxygen complains (attempt #2)
- updated release notes

v2 -> v3:
- fix compilation/doxygen complains
- dropped patch:
  "examples/l3fwd: make ACL work in pipeline and eventdev modes": [2]
  As was mentioned in the patch desctiption it was way too big,
  controversial and incomplete. If the community is ok to introduce
  pipeline model into the l3fwd, then it is propbably worth to be
  a separate patch series.

v1 -> v2:
- rename 'elmst/objst' to 'meta' (Morten)
- introduce new data-path APIs set: one with both meta{} and objs[],
  second with just objs[] (Morten)
- split data-path APIs into burst/bulk flavours (same as rte_ring)
- added dump function for te_soring and improved dump() for rte_ring.
- dropped patch:
  " ring: minimize reads of the counterpart cache-line"
  - no performance gain observed
  - actually it does change behavior of conventional rte_ring
enqueue/dequeue APIs -
it could return available/free less then actually exist in the ring.
As in some other libs we reliy on that information - it will
introduce problems.

The main aim of these series is to extend ring library with
new API that allows user to create/use Staged-Ordered-Ring (SORING)
abstraction. In addition to that there are few other patches that serve
different purposes:
- first two patches are just code reordering to de-duplicate
  and generalize existing rte_ring code.
- patch #3 extends rte_ring_dump() to correctly print head/tail metadata
  for different sync modes.
- next two patches introduce SORING API into the ring library and
  provide UT for it.

SORING overview
===
Staged-Ordered-Ring (SORING) provides a SW abstraction for 'ordered' queues
with multiple processing 'stages'. It is based on conventional DPDK
rte_ring, re-uses many of its concepts, and even substantial part of
its code.
It can be viewed as an 'extension' of rte_ring functionality.
In particular, main SORING properties:
- circular ring buffer with fixed size objects
- producer, consumer plus multiple processing stages in between.
- allows to split objects processing into multiple stages.
- objects remain in the same ring while moving from one stage to the other,
  initial order is preserved, no extra copying needed.
- preserves the ingress order of objects within the queue across multiple
  stages
- each stage (and producer/consumer) can be served by single and/or
  multiple threads.

- number of stages, size and number of objects in the ring are
 configurable at ring initialization time.

Data-path API provides four main operations:
- enqueue/dequeue works in the same manner as for conventional rte_ring,
  all rte_ring synchronization types are supported.
- acquire/release - for each stage there is an acquire (start) and
  release (finish) operation. After some objects are 'acquired' -
  given thread can safely assume that it has exclusive ownership of
  these objects till it will invoke 'release' for them.
  After 'release', objects can be 'acquired' by next stage and/or dequeued
  by the consumer (in case of last stage).

Expected use-case: applications that uses pipeline model
(probably with multiple stages) for packet processing, when preserving
incoming packet order is important.

The concept of ‘ring with stages’ is similar to DPDK OPDL eventdev PMD [1],
but the internals are different.
In particular, SORING maintains internal array of 'states' for each element
in the ring that is  shared by all threads/processes that access the ring.
That allows 'release' to avoid excessive waits on the tail value and helps
to improve performancei and scalability.
In terms of performance, with our measurements rte_soring and
conventional rte_ring provide nearly identical numbers.
As an example, on our SUT: Intel ICX CPU @ 2.00GHz,
l3fwd (--lookup=acl) in pipeline mode [2] both
rte_ring and rte_soring reach ~20Mpps for single I/O lcore and same
num

[PATCH v9 1/7] test/ring: fix failure with custom number of lcores

ring_stress_autotest fails to initialize the ring when RTE_MAX_LCORE value
is not a number of 2.
There is a flaw in calculation required number of elements in the ring.
Fix it by aligning number of elements to next power of 2.

Fixes: bf28df24e915 ("test/ring: add contention stress test")
Cc: sta...@dpdk.org

Reported-by: Phanendra Vukkisala 
Signed-off-by: Konstantin Ananyev 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 app/test/test_ring_stress_impl.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/app/test/test_ring_stress_impl.h b/app/test/test_ring_stress_impl.h
index 8b0bfb11fe..ee5274aeef 100644
--- a/app/test/test_ring_stress_impl.h
+++ b/app/test/test_ring_stress_impl.h
@@ -297,7 +297,7 @@ mt1_init(struct rte_ring **rng, void **data, uint32_t num)
*data = elm;
 
/* alloc ring */
-   nr = 2 * num;
+   nr = rte_align32pow2(2 * num);
sz = rte_ring_get_memsize(nr);
r = rte_zmalloc(NULL, sz, alignof(typeof(*r)));
if (r == NULL) {
-- 
2.35.3

[PATCH v9 5/7] ring/soring: introduce Staged Ordered Ring

Staged-Ordered-Ring (SORING) provides a SW abstraction for 'ordered' queues
with multiple processing 'stages'.
It is based on conventional DPDK rte_ring, re-uses many of its concepts,
and even substantial part of its code.
It can be viewed as an 'extension' of rte_ring functionality.
In particular, main SORING properties:
- circular ring buffer with fixed size objects
- producer, consumer plus multiple processing stages in the middle.
- allows to split objects processing into multiple stages.
- objects remain in the same ring while moving from one stage to the other,
  initial order is preserved, no extra copying needed.
- preserves the ingress order of objects within the queue across multiple
  stages, i.e.:
  at the same stage multiple threads can process objects from the ring in
  any order, but for the next stage objects will always appear in the
  original order.
- each stage (and producer/consumer) can be served by single and/or
  multiple threads.
- number of stages, size and number of objects in the ring are
  configurable at ring initialization time.

Data-path API provides four main operations:
- enqueue/dequeue works in the same manner as for conventional rte_ring,
  all rte_ring synchronization types are supported.
- acquire/release - for each stage there is an acquire (start) and
  release (finish) operation.
  after some objects are 'acquired' - given thread can safely assume that
  it has exclusive possession of these objects till 'release' for them is
  invoked.
  Note that right now user has to release exactly the same number of
  objects that was acquired before.
  After 'release', objects can be 'acquired' by next stage and/or dequeued
  by the consumer (in case of last stage).

Expected use-case: applications that uses pipeline model
(probably with multiple stages) for packet processing, when preserving
incoming packet order is important. I.E.: IPsec processing, etc.

Signed-off-by: Eimear Morrissey 
Signed-off-by: Konstantin Ananyev 
Acked-by: Morten Brørup 
Acked-by: Stephen Hemminger 
---
 devtools/build-dict.sh|   1 +
 doc/api/doxy-api-index.md |   1 +
 doc/guides/prog_guide/img/soring-pic1.svg | 635 ++
 doc/guides/prog_guide/ring_lib.rst| 202 +++
 doc/guides/rel_notes/release_24_11.rst|   8 +
 lib/ring/meson.build  |   4 +-
 lib/ring/rte_soring.c | 198 +++
 lib/ring/rte_soring.h | 557 +++
 lib/ring/soring.c | 613 +
 lib/ring/soring.h | 138 +
 lib/ring/version.map  |  19 +
 11 files changed, 2374 insertions(+), 2 deletions(-)
 create mode 100644 doc/guides/prog_guide/img/soring-pic1.svg
 create mode 100644 lib/ring/rte_soring.c
 create mode 100644 lib/ring/rte_soring.h
 create mode 100644 lib/ring/soring.c
 create mode 100644 lib/ring/soring.h

diff --git a/devtools/build-dict.sh b/devtools/build-dict.sh
index a8cac49029..5eb592abfa 100755
--- a/devtools/build-dict.sh
+++ b/devtools/build-dict.sh
@@ -17,6 +17,7 @@ sed '/^..->/d' |
 sed '/^uint->/d' |
 sed "/^doesn'->/d" |
 sed '/^wasn->/d' |
+sed '/^soring->/d' |
 
 # print to stdout
 cat
diff --git a/doc/api/doxy-api-index.md b/doc/api/doxy-api-index.md
index 266c8b90dc..ac7f3cd010 100644
--- a/doc/api/doxy-api-index.md
+++ b/doc/api/doxy-api-index.md
@@ -173,6 +173,7 @@ The public API headers are grouped by topics:
   [mbuf](@ref rte_mbuf.h),
   [mbuf pool ops](@ref rte_mbuf_pool_ops.h),
   [ring](@ref rte_ring.h),
+  [soring](@ref rte_soring.h),
   [stack](@ref rte_stack.h),
   [tailq](@ref rte_tailq.h),
   [bitset](@ref rte_bitset.h),
diff --git a/doc/guides/prog_guide/img/soring-pic1.svg 
b/doc/guides/prog_guide/img/soring-pic1.svg
new file mode 100644
index 00..c97e66ca43
--- /dev/null
+++ b/doc/guides/prog_guide/img/soring-pic1.svg
@@ -0,0 +1,635 @@
+
+
+
+
+
+
+
+http://www.inkscape.org/namespaces/inkscape";
+   xmlns:sodipodi="http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd";
+   xmlns="http://www.w3.org/2000/svg";
+   xmlns:svg="http://www.w3.org/2000/svg";
+   xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#";
+   xmlns:cc="http://creativecommons.org/ns#";
+   xmlns:dc="http://purl.org/dc/elements/1.1/";>
+  
+
+  
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+
+
+
+
+
+
+
+  
+
+
+  
+
+
+
+  
+
+
+
+
+
+
+
+  
+
+
+  
+  
+
+  
+  
+
+  
+image/svg+xml
+http://purl.org/dc/dcmitype/StillImage"; />
+  
+
+  
+  
+
+
+  
+  
+  
+  
+  
+  
+  
+  
+
+obj5
+obj1
+obj2
+obj3
+
+cons_head
+cons_tail
+prod_head
+prod_tail
+
+lstages states
+
+producer and consumer 
states
+
+

[PATCH v9 6/7] app/test: add unit tests for soring API

Add both functional and stess test-cases for soring API.
Stress test serves as both functional and performance test of soring
enqueue/dequeue/acquire/release operations under high contention
(for both over committed and non-over committed scenarios).

Signed-off-by: Eimear Morrissey 
Signed-off-by: Konstantin Ananyev 
Acked-by: Stephen Hemminger 
---
 app/test/meson.build   |   3 +
 app/test/test_soring.c | 442 +++
 app/test/test_soring_mt_stress.c   |  40 ++
 app/test/test_soring_stress.c  |  48 ++
 app/test/test_soring_stress.h  |  35 ++
 app/test/test_soring_stress_impl.h | 838 +
 6 files changed, 1406 insertions(+)
 create mode 100644 app/test/test_soring.c
 create mode 100644 app/test/test_soring_mt_stress.c
 create mode 100644 app/test/test_soring_stress.c
 create mode 100644 app/test/test_soring_stress.h
 create mode 100644 app/test/test_soring_stress_impl.h

diff --git a/app/test/meson.build b/app/test/meson.build
index 0f7e11969a..e59f27fd51 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -177,6 +177,9 @@ source_file_deps = {
 'test_security_proto.c' : ['cryptodev', 'security'],
 'test_seqlock.c': [],
 'test_service_cores.c': [],
+'test_soring.c': [],
+'test_soring_mt_stress.c': [],
+'test_soring_stress.c': [],
 'test_spinlock.c': [],
 'test_stack.c': ['stack'],
 'test_stack_perf.c': ['stack'],
diff --git a/app/test/test_soring.c b/app/test/test_soring.c
new file mode 100644
index 00..6773a399e6
--- /dev/null
+++ b/app/test/test_soring.c
@@ -0,0 +1,442 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2024 Huawei Technologies Co., Ltd
+ */
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include 
+
+#include "test.h"
+
+#define MAX_ACQUIRED 20
+
+#define SORING_TEST_ASSERT(val, expected) do { \
+   RTE_TEST_ASSERT(expected == val, \
+   "%s: expected %u got %u\n", #val, expected, val); \
+} while (0)
+
+static void
+set_soring_init_param(struct rte_soring_param *prm,
+   const char *name, uint32_t esize, uint32_t elems,
+   uint32_t stages, uint32_t stsize,
+   enum rte_ring_sync_type rst_prod,
+   enum rte_ring_sync_type rst_cons)
+{
+   prm->name = name;
+   prm->elem_size = esize;
+   prm->elems = elems;
+   prm->stages = stages;
+   prm->meta_size = stsize;
+   prm->prod_synt = rst_prod;
+   prm->cons_synt = rst_cons;
+}
+
+static int
+move_forward_stage(struct rte_soring *sor,
+   uint32_t num_packets, uint32_t stage)
+{
+   uint32_t acquired;
+   uint32_t ftoken;
+   uint32_t *acquired_objs[MAX_ACQUIRED];
+
+   acquired = rte_soring_acquire_bulk(sor, acquired_objs, stage,
+   num_packets, &ftoken, NULL);
+   SORING_TEST_ASSERT(acquired, num_packets);
+   rte_soring_release(sor, NULL, stage, num_packets,
+   ftoken);
+
+   return 0;
+}
+
+/*
+ * struct rte_soring_param param checking.
+ */
+static int
+test_soring_init(void)
+{
+   struct rte_soring *sor = NULL;
+   struct rte_soring_param prm;
+   int rc;
+   size_t sz;
+   memset(&prm, 0, sizeof(prm));
+
+/*init memory*/
+   set_soring_init_param(&prm, "alloc_memory", sizeof(uintptr_t),
+   4, 1, 4, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   sz = rte_soring_get_memsize(&prm);
+   sor = rte_zmalloc(NULL, sz, RTE_CACHE_LINE_SIZE);
+   RTE_TEST_ASSERT_NOT_NULL(sor, "could not allocate memory for soring");
+
+   set_soring_init_param(&prm, "test_invalid_stages", sizeof(uintptr_t),
+   4, 0, 4, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   rc = rte_soring_init(sor, &prm);
+   RTE_TEST_ASSERT_FAIL(rc, "initted soring with invalid num stages");
+
+   set_soring_init_param(&prm, "test_invalid_esize", 0,
+   4, 1, 4, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   rc = rte_soring_init(sor, &prm);
+   RTE_TEST_ASSERT_FAIL(rc, "initted soring with 0 esize");
+
+   set_soring_init_param(&prm, "test_invalid_esize", 9,
+   4, 1, 4, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   rc = rte_soring_init(sor, &prm);
+   RTE_TEST_ASSERT_FAIL(rc, "initted soring with esize not multiple of 4");
+
+   set_soring_init_param(&prm, "test_invalid_rsize", sizeof(uintptr_t),
+   4, 1, 3, RTE_RING_SYNC_MT, RTE_RING_SYNC_MT);
+   rc = rte_soring_init(sor, &prm);
+   RTE_TEST_ASSERT_FAIL(rc, "initted soring with rcsize not multiple of 
4");
+
+   set_soring_init_param(&prm, "test_invalid_elems", sizeof(uintptr_t),
+   RTE_SORING_ELEM_MAX + 1, 1, 4, RTE_RING_SYNC_MT,
+

Re: [PATCH v15 1/3] power: introduce PM QoS API on CPU wide

2024-11-11 Thread lihuisong (C)




在 2024/11/11 18:18, Thomas Monjalon 写道:

11/11/2024 03:25, Huisong Li:

Some service are delay sensitive and very except the low
resume time, like interrupt packet receiving mode.

I don't understand this sentence.
Are you sure about "very except"?
It is also in the guide.
If something needs a reword, please send a patch.


Sorry, it's a typo. except-->expect.
I will send a patch to fix it.


.

Re: [PATCH v4 0/4] RSS hash key generation

24/10/2024 20:46, Vladimir Medvedkin:
> Vladimir Medvedkin (4):
>   thash: add RSS hash key generation API
>   hash: add dynamic polynomial calculation
>   hash: implement RSS hash key generation API
>   test/thash: add tests for RSS key generation API

Squash stub patch and applied with minor improvements for spacing and 
punctuation.

RE: [RFC] ip_frag: support IPv6 reassembly with extensions


> 14/02/2024 04:51, Stephen Hemminger:
> > On Tue, 13 Feb 2024 12:47:27 +0100
> >  wrote:
> >
> > > +/*
> > > + * Function to crawl through the extension header stack.
> > > + * This function breaks as soon a the fragment header is
> > > + * found and returns the total length the traversed exts
> > > + * and the last extension before the fragment header
> > > + */
> > > +static inline uint32_t
> > > +ip_frag_get_last_exthdr(struct rte_ipv6_hdr *ip_hdr, uint8_t **last_ext)
> > > +{
> > > + uint32_t total_len = 0;
> > > + size_t ext_len = 0;
> > > + *last_ext = (uint8_t *)(ip_hdr + 1);
> > > + int next_proto = ip_hdr->proto;
> > > +
> > > + while (next_proto != IPPROTO_FRAGMENT &&
> > > + (next_proto = rte_ipv6_get_next_ext(
> > > + *last_ext, next_proto, &ext_len)) >= 0) {
> > > +
> > > + total_len += ext_len;
> > > +
> > > + if (next_proto == IPPROTO_FRAGMENT)
> > > + return total_len;
> > > +
> > > + *last_ext += ext_len;
> > > + }
> > > +
> > > + return total_len;
> > > +}
> >
> > Doing endless loop like this opens up DoS attacks.
> > Better to use rte_next_skip_ip6_ext() or do similar limited loop.
> 
> There was no reply to this interesting comment?

I think there is a limit is the latest version for that patch:
https://patchwork.dpdk.org/project/dpdk/patch/20241015082133.3910533-1-vignesh.purushotham.srini...@ericsson.com/
It is also has an ACK from me...
Though looking at it once again - we'd better have an extra check here to
make sure that total_len would not exceed mbuf->data_len.

Re: [PATCH] bus/fslmc: fix bus scan return failure value

11/11/2024 11:26, Hemant Agrawal:
> fslmc bus will not be available on non DPAA2 platform.
> Do not return error on scan failure.
> 
> Signed-off-by: Hemant Agrawal 

Fixes: 57cb02edf122 ("bus/fslmc: enhance MC VFIO multi-process support")

Applied, thanks.

[DPDK/ethdev Bug 1575] The 82599ES NIC encounters a core dump when using dpdk-proc-info to retrieve firmware information.

2024-11-11 Thread bugzilla

https://bugs.dpdk.org/show_bug.cgi?id=1575

Bug ID: 1575
   Summary: The 82599ES NIC encounters a core dump when using
dpdk-proc-info to retrieve firmware information.
   Product: DPDK
   Version: 23.11
  Hardware: All
OS: All
Status: UNCONFIRMED
  Severity: normal
  Priority: Normal
 Component: ethdev
  Assignee: dev@dpdk.org
  Reporter: junwan...@cestc.cn
  Target Milestone: ---

Hi, I am using an 82599ES NIC and encountered a core dump when executing
dpdk-proc-info to retrieve firmware information. I added some debug print
statements in the ixgbe driver and found that the core dump occurred in
ixgbe_get_oem_prod_version. Additionally, my x710 NIC is able to retrieve
firmware information without any issues.


# /dpdk/app/dpdk-proc-info -- -p 1 --firmware-version
== show - firmware version ==
ixgbe_fw_version_get(): ixgbe_fw_version_get hw:0x17ff11440
ixgbe_get_oem_prod_version(): ixgbe_get_oem_prod_version 111 read:0x63e52f
Segmentation fault (core dumped)

# /dpdk/app/dpdk-proc-info -- -p 4 --firmware-version
== show - firmware version ==
Ethdev port 2 firmware version: 8.15 0x8000a4e8 1.2879.0

# lspci |grep Eth
3b:00.0 Ethernet controller: Intel Corporation 82599ES 10-Gigabit SFI/SFP+
Network Connection (rev 01)
3b:00.1 Ethernet controller: Intel Corporation 82599ES 10-Gigabit SFI/SFP+
Network Connection (rev 01)
3c:00.0 Ethernet controller: Intel Corporation 82599ES 10-Gigabit SFI/SFP+
Network Connection (rev 01)
3c:00.1 Ethernet controller: Intel Corporation 82599ES 10-Gigabit SFI/SFP+
Network Connection (rev 01)
af:00.0 Ethernet controller: Intel Corporation Ethernet Controller X710 for
10GbE SFP+ (rev 02)
af:00.1 Ethernet controller: Intel Corporation Ethernet Controller X710 for
10GbE SFP+ (rev 02)
# /dpdk/app/dpdk-dumpcap -D
0. :3c:00.0
1. :3c:00.1
2. :af:00.0
3. :af:00.1

-- 
You are receiving this mail because:
You are the assignee for the bug.

Re: [v1 3/3] meson: skip symlinks to PMDs script on no sub directory

06/08/2024 15:57, Bruce Richardson:
> On Tue, Aug 06, 2024 at 07:12:18PM +0530, Gagandeep Singh wrote:
> > if user gives option -Ddrivers_install_subdir= or
> > -Ddrivers_install_subdir=. to install all the PMDs in
> > parent directory, The post installation script
> > "symlink-drivers-solibs.sh" can create invalid library file
> > named 'librte_*.so*' or meson installation errors:
> > 
> > ln: './librte_*' and './librte_*' are the same file
> > FAILED: install script '/bin/sh /home/nxa12342/upstream/dpdk/config/.
> > /buildtools/symlink-drivers-solibs.sh lib/x86_64-linux-gnu .' exit code 1,
> > stopped
> > FAILED: meson-install
> > 
> > This patch removes this post-install script execution to symlink
> > the PMDs libraries when no sub directory is given.
> > 
> > Signed-off-by: Gagandeep Singh 
> 
> Acked-by: Bruce Richardson 

Applied only this patch of the series.
Let's continue the discussion for the others, thanks.

[PATCH] bus/fslmc: fix bus scan return failure value

2024-11-11 Thread Hemant Agrawal

fslmc bus will not be available on non DPAA2 platform.
Do not return error on scan failure.

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/fslmc_bus.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index 17b3b6156c..68ad2b801e 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -243,8 +243,6 @@ rte_fslmc_parse(const char *name, void *addr)
uint8_t sep_exists = 0;
int ret = -1;
 
-   DPAA2_BUS_DEBUG("Parsing dev=(%s)", name);
-
/* There are multiple ways this can be called, with bus:dev, name=dev
 * or just dev. In all cases, the 'addr' is actually a string.
 */
@@ -329,7 +327,8 @@ rte_fslmc_scan(void)
group_name = getenv("DPRC");
if (!group_name) {
DPAA2_BUS_DEBUG("DPAA2: DPRC not available");
-   return -EINVAL;
+   ret = -EINVAL;
+   goto scan_fail;
}
 
ret = fslmc_get_container_group(group_name, &groupid);
-- 
2.25.1

Re: [PATCH] git: show hidden tracked files

17/10/2024 18:26, Stephen Hemminger:
> On Thu, 17 Oct 2024 17:15:03 +0200
> Thomas Monjalon  wrote:
> 
> > The hidden files are not considered because of the first rule in
> > .gitignore file.
> > 
> > In order to see changes in the hidden tracked files,
> > they are added in .gitignore as explicit include rules.
> > 
> > The derivatives of the first level files are also shown.
> > Example: .mailmap.rej
> > 
> > Signed-off-by: Thomas Monjalon 
> Acked-by: Stephen Hemminger 

Applied

[PATCH v9 7/7] test: add stress test suite

Add a new test suite which purpose is to run 'stress' tests:
main purpose is put a pressure to dpdk sync algorithms
to flag their misbehaving/slowdown/etc.
Right now it consists from just 2 test-cases:
meson test --suite stress-tests --list
DPDK:stress-tests / ring_stress_autotest
DPDK:stress-tests / soring_stress_autotest

These tests are quite time consuming (~15 mins each),
that's another reason to put them into a separate test-suite.

Signed-off-by: Konstantin Ananyev 
Acked-by: Stephen Hemminger 
---
 app/test/suites/meson.build   | 10 ++
 app/test/test.h   |  1 +
 app/test/test_ring_stress.c   |  2 +-
 app/test/test_soring_stress.c |  2 +-
 4 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/app/test/suites/meson.build b/app/test/suites/meson.build
index 191702cf76..e482373330 100644
--- a/app/test/suites/meson.build
+++ b/app/test/suites/meson.build
@@ -5,6 +5,7 @@
 # to complete, so timeout to 10 minutes
 timeout_seconds = 600
 timeout_seconds_fast = 10
+timeout_seconds_stress = 900
 
 test_no_huge_args = ['--no-huge', '-m', '2048']
 has_hugepage = run_command(has_hugepages_cmd, check: true).stdout().strip() != 
'0'
@@ -21,6 +22,7 @@ endif
 # - fast_tests
 # - perf_tests
 # - driver_tests
+# - stress_tests
 test_suites = run_command(get_test_suites_cmd, autotest_sources,
  check: true).stdout().strip().split()
 foreach suite:test_suites
@@ -39,6 +41,14 @@ foreach suite:test_suites
 timeout: timeout_seconds,
 is_parallel: false)
 endforeach
+elif suite_name == 'stress-tests'
+foreach t: suite_tests
+test(t, dpdk_test,
+env: ['DPDK_TEST=' + t],
+timeout: timeout_seconds_stress,
+is_parallel: false,
+suite: suite_name)
+endforeach
 elif suite_name != 'fast-tests'
 # simple cases - tests without parameters or special handling
 foreach t: suite_tests
diff --git a/app/test/test.h b/app/test/test.h
index 15e23d297f..ebc4864bf8 100644
--- a/app/test/test.h
+++ b/app/test/test.h
@@ -208,5 +208,6 @@ void add_test_command(struct test_command *t);
 #define REGISTER_FAST_TEST(cmd, no_huge, ASan, func)  
REGISTER_TEST_COMMAND(cmd, func)
 #define REGISTER_PERF_TEST REGISTER_TEST_COMMAND
 #define REGISTER_DRIVER_TEST REGISTER_TEST_COMMAND
+#define REGISTER_STRESS_TEST REGISTER_TEST_COMMAND
 
 #endif
diff --git a/app/test/test_ring_stress.c b/app/test/test_ring_stress.c
index 1af45e0fc8..82e19b02c3 100644
--- a/app/test/test_ring_stress.c
+++ b/app/test/test_ring_stress.c
@@ -63,4 +63,4 @@ test_ring_stress(void)
return (k != n);
 }
 
-REGISTER_TEST_COMMAND(ring_stress_autotest, test_ring_stress);
+REGISTER_STRESS_TEST(ring_stress_autotest, test_ring_stress);
diff --git a/app/test/test_soring_stress.c b/app/test/test_soring_stress.c
index 334af6a29c..e5655d49cb 100644
--- a/app/test/test_soring_stress.c
+++ b/app/test/test_soring_stress.c
@@ -45,4 +45,4 @@ test_ring_stress(void)
return (k != n);
 }
 
-REGISTER_TEST_COMMAND(soring_stress_autotest, test_ring_stress);
+REGISTER_STRESS_TEST(soring_stress_autotest, test_ring_stress);
-- 
2.35.3

Re: [PATCH v1 0/2] Testpmd flow update/destroy fixes

On 10/31/2024 3:00 PM, Danylo Vodopianov wrote:
> These patches provide next fixes:
>   1. The testpmd command “flow update…“ provides a nullptr as the
>  context variable.
>   2. Avoid removal of additional flows after requested number of flows has
>  been already removed.
> 
> Danylo Vodopianov (2):
>   app/testpmd: fix flow update
>   app/testpmd: fix flow destroy
>

Hi Aman, Ori,

Can you please help reviewing this patch series?

RE: [PATCH 0/2] Build with Gcc 15 fixes

2024-11-11 Thread Morten Brørup

For the series,
Reviewed-by: Morten Brørup

RE: [PATCH] doc: update match with compare result item limitation

2024-11-11 Thread Bing Zhao

@Suanming Mou, could you please reword the description as Stephen suggested and 
send a new version?
Hi, @Stephen Hemminger, PSB. Rule should be item, that would be more accurate.

> -Original Message-
> From: Stephen Hemminger 
> Sent: Thursday, November 7, 2024 2:29 AM
> To: Suanming Mou 
> Cc: Dariusz Sosnowski ; Slava Ovsiienko
> ; Bing Zhao ; Ori Kam
> ; Matan Azrad ; dev@dpdk.org;
> sta...@dpdk.org
> Subject: Re: [PATCH] doc: update match with compare result item limitation
> 
> External email: Use caution opening links or attachments
> 
> 
> On Tue, 5 Nov 2024 09:47:36 +0800
> Suanming Mou  wrote:
> 
> > +  - In switch mode, when ``repr_matching_en`` flag is enabled in the
> devarg,
> > +the match with compare result item is not supported to the
> ``ingress``
> > +rule as an implicit REPRESENTED_PORT need to be added to the
> matcher.
> > +That REPRESENTED_PORT item conflicts with the single item
> limitation for
> > +match with compare result item.
> 
> These sentences are hard to read, the wording is very awkward.
> 
> Have no detailed insight into the mlx5 matching but maybe this wording:
> 
> In switch mode, when the ``repr_matching_en`` flag is used then matching
is used (by default)
> with ingress comparison is not supported because an implicit rule is
Implicit rule -> implicit item
> required to match the REPRESENTED_PORT. And that implicit rule would
Match the REP with the same rule
Rule-> item 
> conflict with the ingress compare rule.
Comparison item.

答复: [PATCH v6 08/17] net/r8169: add support for phy configuration

2024-11-11 Thread 王颢

Dear Ferruh,

Regarding the commit message, I will think about how to write it.

I’m sorry, but for the registers in the PMD that do not use macro definitions, 
most of them are company confidential and their meanings will not be disclosed 
to the public. This part will be maintained by us.

Best Regards,
Howard Wang

-邮件原件-
发件人: Ferruh Yigit  
发送时间: 2024年11月11日 8:17
收件人: 王颢 ; dev@dpdk.org
抄送: pro_nic_d...@realtek.com
主题: Re: [PATCH v6 08/17] net/r8169: add support for phy configuration

External mail.

On 11/8/2024 12:11 PM, Howard Wang wrote:
> This patch contains phy config, ephy config and so on.
>
> Signed-off-by: Howard Wang 
>

<...>

> @@ -62,6 +62,12 @@ rtl_dev_start(struct rte_eth_dev *dev)
>   struct rtl_hw *hw = &adapter->hw;
>   int err;
>
> + rtl_powerup_pll(hw);
> +
> + rtl_hw_ephy_config(hw);
> +
> + rtl_hw_phy_config(hw);
> +
>   rtl_hw_config(hw);
>
>   /* Initialize transmission unit */ @@ -74,6 +80,8 @@ 
> rtl_dev_start(struct rte_eth_dev *dev)
>   goto error;
>   }
>
> + rtl_mdio_write(hw, 0x1F, 0x);
> +
>

It is very hard know what code like above does, which makes harder for anyone 
to read and understand your code, can you please either create macros for 
registers or comment as much as possible, so others can improve/fix your code 
when required.

Re: [PATCH v15 1/3] power: introduce PM QoS API on CPU wide

11/11/2024 03:25, Huisong Li:
> Some service are delay sensitive and very except the low
> resume time, like interrupt packet receiving mode.

I don't understand this sentence.
Are you sure about "very except"?
It is also in the guide.
If something needs a reword, please send a patch.

[PATCH v6] doc: modernize build instructions on Windows

Meson 0.57 was an unstable version and is now outdated.
The referenced bug in Meson 0.58 is fixed in stable releases.
Recommend the first version containing the fix: 1.5.2.

Building DPDK applications that run on 32-bit Windows is
currently not supported. However, some Visual Studio environments
default to producing 32-bit binaries.
Recommend instructing the developer prompt to produce 64-bit binaries
when that is the case.

Signed-off-by: Andre Muezerie 
---
 .mailmap  |  1 +
 doc/guides/windows_gsg/build_dpdk.rst | 20 +++-
 2 files changed, 16 insertions(+), 5 deletions(-)

diff --git a/.mailmap b/.mailmap
index 5290420258..60db768b98 100644
--- a/.mailmap
+++ b/.mailmap
@@ -98,6 +98,7 @@ Anatolii Gerasymenko 
 Anatoly Burakov 
 Anbarasan Murugesan 
 Anders Roxell 
+Andre Muezerie  
 Andrea Arcangeli 
 Andrea Grandi 
 Andre Richter 
diff --git a/doc/guides/windows_gsg/build_dpdk.rst 
b/doc/guides/windows_gsg/build_dpdk.rst
index 708875c16f..2c7bf1a309 100644
--- a/doc/guides/windows_gsg/build_dpdk.rst
+++ b/doc/guides/windows_gsg/build_dpdk.rst
@@ -72,10 +72,7 @@ A good option to choose is the MSI installer for both meson 
and ninja together::
 

http://mesonbuild.com/Getting-meson.html#installing-meson-and-ninja-with-the-msi-installer%22
 
-Required version is Meson 0.57.
-
-Versions starting from 0.58 are unusable with LLVM toolchain
-because of a `Meson issue `_.
+The minimal Meson supported version is 1.5.2.
 
 
 Install the Backend
@@ -129,11 +126,24 @@ Depending on the distribution, paths in this file may 
need adjustments.
 Option 3. Native Build on Windows using MSVC
 
 
-Open a 'Developer PowerShell for VS 2022' prompt from the start menu.
+Open a 'Visual Studio Developer Command Prompt'.
 The developer prompt will configure the environment
 to select the appropriate compiler, linker and SDK paths
 required to build with Visual Studio 2022.
 
+Building DPDK applications that run on 32-bit Windows is currently not
+supported. If your Visual Studio environment defaults to producing
+32-bit binaries you can instruct the toolset to produce 64-bit binaries using 
"-arch" parameter.
+For more details about the Developer Prompt options look at the `Visual Studio 
Developer
+Command Prompt and Developer PowerShell
+`_.
+
+.. code-block:: console
+
+"C:\Program Files\Microsoft Visual 
Studio\2022\Enterprise\Common7\Tools\VsDevCmd.bat" -arch=amd64
+
+Compile the code from the developer prompt.
+
 .. code-block:: console
 
cd C:\Users\me\dpdk
-- 
2.34.1

RE: [PATCH v2 1/2] config/arm: strict use of -mcpu for supported CPUs

2024-11-11 Thread Wathsala Wathawana Vithanage


> Subject: Re: [PATCH v2 1/2] config/arm: strict use of -mcpu for supported
> CPUs
> 
> Please check the CI results, there are some failures.
> 

Hi Thomas,

CI is failing because gcc version 11 does not enable crypto extension for 
thunderx2,
with -mcpu=thunderx2t99 unless +crypto feature modifier is used.
Goal of this patch is to drop use of such feature modifiers when an -mcpu is 
available
for a given platform.
This issue is fixed in gcc version 13 for thunderx2t99. 

Thanks.

--wathsala

Re: [PATCH 00/10] require checking ethdev get return value

> Stephen Hemminger (10):
>   app/test: remove redundant call
>   net/memif: check return value from rte_eth_dev_info_get
>   graph: check return value from rte_eth_dev_info_get
>   examples/ethtool: handle devices without registers
>   examples/l3fwd: check return value from ethdev info
>   examples/ntb: always check return value
>   examples/pipeline: check return value of ethdev functions
>   examples/qos_sched: check return value from rte_eth_link_get
>   ethdev: check return value from rte_eth_dev_info_get
>   ethdev: require checking results of info_get functions

Applied, thanks.

Re: [PATCH v1] net/cpfl: fix IO port forwarding issue

2024-11-11 Thread Bruce Richardson

On Fri, Nov 08, 2024 at 11:09:23AM +, Praveen Shetty wrote:
> Issue was reported in dpdk internal regression testing that
> forwarding the packets from IDPF VF to IO ports using the cpfl
> represented_port action was not working.
> This patch will fix this issue.
> 
> Fixes: 861261957684 ("net/cpfl: add checks for flow action types")
> Cc: sta...@dpdk.org
> 
> Signed-off-by: Praveen Shetty 
> ---
>  drivers/net/cpfl/cpfl_flow_engine_fxp.c | 5 -
>  1 file changed, 5 deletions(-)
> 

Following discussion with Praveen to clarify the issue and the fix, applied
with expanded commit-log message (below) to next-net-intel.

"""
net/cpfl: fix forwarding to physical port

CPFL PMD should be able to support below traffic forwarding capabilities
based on the rte flow action types.

 1. Forwarding the traffic to the local CPFL vports using
port_representor action.

 2. Forwarding the traffic to the physical IO ports using
represented_port action.

 3. Forwarding the traffic to an IDPF VF using represented_port action.

The 2nd use case, forwarding the traffic to IO ports using
represented_port action, is not working due to the additional check
added in the previous patch (861261957684 ("net/cpfl: add checks for
flow action types"))

This patch removes the incorrect check to fix the issue.
"""

Acked-by: Bruce Richardson 

Applied,
Thanks.

Re: [PATCH] config: limit lcore variable maximum size to 4k

On Mon, 11 Nov 2024 08:22:46 +0100
Mattias Rönnblom  wrote:

> On 2024-11-09 00:52, Morten Brørup wrote:
> >> From: Mattias Rönnblom [mailto:hof...@lysator.liu.se]
> >> Sent: Friday, 8 November 2024 23.23
> >>
> >> On 2024-11-08 20:53, Morten Brørup wrote:  
>  From: Morten Brørup [mailto:m...@smartsharesystems.com]
>  Sent: Friday, 8 November 2024 19.35
>   
> > From: David Marchand [mailto:david.march...@redhat.com]
> > Sent: Friday, 8 November 2024 19.18
> >
> > OVS locks all pages to avoid page faults while processing packets.  
> >>>
> >>> It sounds smart, so I just took a look at how it does this. I'm not  
> >> sure, but it seems like it only locks pages that are actually mapped
> >> (current and future).  
> >>>  
> >>
> >> mlockall(MLOCK_CURRENT) will bring in the whole BSS, it seems. Plus all
> >> the rest like unused parts of the execution stacks, the data section
> >> and
> >> unused code (text) in the binary and all libraries it has linked to.
> >>
> >> It makes a simple (e.g., a unit test) DPDK 24.07 program use ~33x more
> >> residential memory. After lcore variables, the same MLOCK_CURRENT-ed
> >> program is ~30% larger than before. So, a relatively modest increase.  
> > 
> > Thank you for testing this, Mattias.
> > What are the absolute numbers, i.e. in KB, to get an idea of the numbers I 
> > should be looking for?
> >   
> 
> Hello world type program with static linking. Default DPDK config. x86_64.
> 
> DPDK version  MAX_LCORE_VAR   EAL params mlock  RSS [MB]
> 22.11 -   --no-huge -m 1000  no 22
> 24.11 1048576 --no-huge -m 1000  no 22
> 24.11 1048576 --no-huge -m 1000  yes1576
> 24.11 4096--no-huge -m 1000  yes1445
> 22.11 -   -  yes333*
> 24.11 1048576 -  yes542*
> 24.11 4096-  yes411*
> 
> * Excluding huge pages
> 
> If you are more selective what libraries you bring in, the footprint 
> will be lower. How large a fraction is effectively unavoidable, I don't 
> know. The relative increase will depends on how much memory the 
> application uses, obviously. The hello world app doesn't have any 
> app-level state.
> 
> > I wonder why the footprint grows at all... Intuitively the same variables 
> > should consume approximately the same amount of RAM, regardless how they 
> > are allocated.
> > Speculating...  
> 
> lcore variables use malloc(), which in turn does not bring in memory 
> pages unless they are needed. Much of the lcore buffer will be unused, 
> and not resident. I covered this, including some example calculation of 
> the space savings, in an earlier thread. It may be in the programmer's 
> guide as well, I don't remember.

I suspect that glibc malloc assumes a virtual memory backed model.
It is lazy about reclaiming memory and grows in large chunks.
This is one of the reasons malloc() is faster than rte_malloc()
for allocation.

https://sourceware.org/glibc/wiki/MallocInternals

RE: [PATCH v2 1/2] config/arm: strict use of -mcpu for supported CPUs

2024-11-11 Thread Wathsala Wathawana Vithanage

> > Hi Thomas,
> >
> > CI is failing because gcc version 11 does not enable crypto extension
> > for thunderx2, with -mcpu=thunderx2t99 unless +crypto feature modifier is
> used.
> > Goal of this patch is to drop use of such feature modifiers when an
> > -mcpu is available for a given platform.
> > This issue is fixed in gcc version 13 for thunderx2t99.
> 
> We cannot afford dropping the support of GCC 11.
> Which solution do you propose?
> 
Adding a pseudo mcpu for thunderx2t99 as described in the patch. I will send a 
v3
with that change.

RE: [PATCH v2] doc: update match with compare result item limitation

2024-11-11 Thread Bing Zhao

Thanks.

> -Original Message-
> From: Suanming Mou 
> Sent: Wednesday, November 6, 2024 5:34 PM
> To: Dariusz Sosnowski ; Slava Ovsiienko
> ; Bing Zhao ; Ori Kam
> ; Matan Azrad 
> Cc: dev@dpdk.org; sta...@dpdk.org
> Subject: [PATCH v2] doc: update match with compare result item limitation
> 
> In switch mode, when ``repr_matching_en`` flag is enabled in the devarg,
> the match with compare result item is not supported to the ``ingress``
> rule as an implicit REPRESENTED_PORT need to be added to the matcher.
> That REPRESENTED_PORT item conflicts with the single item limitation for
> match with compare result item.
> 
> Fixes: cb25df7ce9d6 ("net/mlx5: support comparison matching")
> Cc: sta...@dpdk.org
> 
> Signed-off-by: Suanming Mou 

Acked-by: Bing Zhao 

> ---
> 
> v2: Add `repr_matching_en`` enabled by default explanation.
> 
> ---
>  doc/guides/nics/mlx5.rst | 5 +
>  1 file changed, 5 insertions(+)
> 
> diff --git a/doc/guides/nics/mlx5.rst b/doc/guides/nics/mlx5.rst index
> 7ced6ca8f9..64933a0823 100644
> --- a/doc/guides/nics/mlx5.rst
> +++ b/doc/guides/nics/mlx5.rst
> @@ -460,6 +460,11 @@ Limitations
>- The field type ``RTE_FLOW_FIELD_VALUE`` must be the base (``b``)
> field.
>- The field type ``RTE_FLOW_FIELD_RANDOM`` can only be compared with
>  ``RTE_FLOW_FIELD_VALUE``.
> +  - In switch mode, when ``repr_matching_en`` flag is enabled (by
> default)
> +in the devarg, the match with compare result item is not supported to
> +the ``ingress`` rule as an implicit REPRESENTED_PORT need to be added
> to
> +the matcher. That REPRESENTED_PORT item conflicts with the single
> item
> +limitation for match with compare result item.
> 
>  - No Tx metadata go to the E-Switch steering domain for the Flow group 0.
>The flows within group 0 and set metadata action are rejected by
> hardware.
> --
> 2.34.1

Re: [v4,2/5] raw/gdtc: add support for queue setup operation

On Tue, 29 Oct 2024 21:45:24 +0800
Yong Zhang  wrote:

> + /* Init sw_ring */
> + memset(name, 0, sizeof(name));
> + snprintf(name, RTE_RAWDEV_NAME_MAX_LEN, "gdma_vq%d_sw_ring", queue_id);
> + size = queue->queue_size * sizeof(struct zxdh_gdma_job *);
> + queue->sw_ring.job = rte_zmalloc(name, size, 0);

This works as is, but:
1. Do not need memset() before snprintf since snprintf will create null 
terminated string
2. rte_zmalloc() doesn't really care about the name arg
3. You should be using rte_calloc() if it is an array, because that is 
more common
   and may help some checkers in future when DPDK can tell compiler 
that rte_calloc is functional
   equivalent of calloc().
4. The name is being used as a memzone name, and not a rawdev name.
RTE_RAWDEV_NAME_MAX_LEN = 64
RTE_MEMZONE_NAMESIZE = 32
   so you are at some risk of future bug where memzone names becomes 
too long.

> +
> + if (queue->sw_ring.job != NULL)
> + rte_free(queue->sw_ring.job);
> +
> + if (queue->ring.ring_mz != NULL)
> + rte_memzone_free(queue->ring.ring_mz);

Both these function handle NULL as a no-op, so the check for NULL is 
unnecessary.
There is a coccinelle script that finds these.

Re: [v4,1/5] raw/gdtc: introduce gdtc raw device driver

On Tue, 29 Oct 2024 21:45:22 +0800
Yong Zhang  wrote:

> +char zxdh_gdma_driver_name[] = "rawdev_zxdh_gdma";
> +char dev_name[] = "zxdh_gdma";
> +
> +uint32_t
> +zxdh_gdma_read_reg(struct rte_rawdev *dev, uint16_t queue_id, uint32_t 
> offset)
> +{
> + struct zxdh_gdma_rawdev *gdmadev = zxdh_gdma_rawdev_get_priv(dev);
> + uint32_t addr = 0;
> + uint32_t val = 0;
> +
> + addr = offset + queue_id * ZXDH_GDMA_CHAN_SHIFT;
> + val = *(uint32_t *)(gdmadev->base_addr + addr);
> +
> + return val;
> +}
> +
> +void
> +zxdh_gdma_write_reg(struct rte_rawdev *dev, uint16_t queue_id, uint32_t 
> offset, uint32_t val)
> +{
> + struct zxdh_gdma_rawdev *gdmadev = zxdh_gdma_rawdev_get_priv(dev);
> + uint32_t addr = 0;
> +
> + addr = offset + queue_id * ZXDH_GDMA_CHAN_SHIFT;
> + *(uint32_t *)(gdmadev->base_addr + addr) = val;
> +}

This driver is made up one C file. Please make all these helper functions
and the names static. If static checkers and optimizers can work better.

Re: [PATCH 0/2] ethdev: support single queue per port

25/10/2024 13:52, Morten Brørup:
> When configuring DPDK for one queue per port
> (#define RTE_MAX_QUEUES_PER_PORT 1), compilation of some network drivers
> fails with e.g.:
> 
> ../drivers/net/bnxt/bnxt_rxq.c: In function 'bnxt_rx_queue_stop':
> ../drivers/net/bnxt/bnxt_rxq.c:587:34: error: array subscript 1 is above 
> array bounds of 'uint8_t[1]' {aka 'unsigned char[1]'} [-Werror=array-bounds=]
>   587 | dev->data->rx_queue_state[q_id] = RTE_ETH_QUEUE_STATE_STOPPED;
>   | ~^~
> In file included from ../drivers/net/bnxt/bnxt.h:16,
>  from ../drivers/net/bnxt/bnxt_rxq.c:10:
> ../lib/ethdev/ethdev_driver.h:168:17: note: while referencing 'rx_queue_state'
>   168 | uint8_t rx_queue_state[RTE_MAX_QUEUES_PER_PORT];
>   | ^~
> 
> This series fixes those compiler emitted errros as follows:
> 1. A precondition hint macro is introduced, which can be used to prevent
>the compiler/optimizer from considering scenarios that cannot occur.
> 2. The hint is added to the network drivers where a compiler in the CI has
>been seen to emit the above error when DPDK is configured for one queue
>per port, but we know that the error cannot occur.
> 
> Morten Brørup (2):
>   eal: add unreachable and precondition hints
>   drivers/net: support single queue per port

Applied, thanks.

Re: [PATCH] raw/gdtc: introduce gdtc raw device driver

On Tue, 12 Nov 2024 05:12:32 +0100
Thomas Monjalon  wrote:

> Hello,
> 
> 04/11/2024 03:15, Yong Zhang:
> > Hi thomas,
> > 
> > I submitted a patch on October 29 and haven't received response.
> > Wondering if any modifications are needed? Could you provide an update on 
> > the review?
> > Thanks. Looking forward to hearing from you.  
> 
> I cannot apply your patches because the format looks wrong.
> It looks there an issue with line ending. Windows?

I made a bundle in patchwork, then downloaded it and did:

 $ git am ~/Downloads/bundle-1307-gdtc.mbox

And it worked for me, no whitespace or new line issues.
I wonder if there is some hidden setting in patchwork or git cleaning it for me?

Re: RFC - Tap io_uring PMD

On Wed, 6 Nov 2024 08:46:55 +0100
Maxime Coquelin  wrote:

> Hi Stephen,
> 
> On 10/30/24 22:56, Stephen Hemminger wrote:
> > The current tap device is slow both due to architectural choices and the
> > overhead of Linux system calls. I am exploring a how to fix that but some
> > of the choices require some tradeoffs. Which leads to some open questions:
> > 
> > 1. DPDK tap also support tunnel (TUN) mode where there is no Ethernet header
> > only L3. Does anyone actually use this? It is different than what every 
> > other
> > PMD expects.
> > 
> > 2. The fastest way to use kernel TAP device would be to use io_uring.
> > But this was added in 5.1 kernel (2019). Rather than having conditional 
> > or
> > dual mode in DPDK tap device, perhaps there should just be a new PMD 
> > tap_uring?
> > 
> > 3. Current TAP device provides hooks for several rte_flow types by playing
> > games with kernel qdisc. Does anyone really use this? Propose just not 
> > doing
> > this in new tap_uring.
> > 
> > 4. What other features of TAP device beyond basic send/receive make sense?
> > It looks like new device could support better statistics.
> > 
> > 5. What about Rx interrupt support?
> > 
> > Probably the hardest part of using io_uring is figuring out how to collect
> > completions. The simplest way would be to handle all completions rx and tx
> > in the rx_burst function.
> >   
> 
> Why not just use Virtio-user PMD with Vhost-kernel backend [0]?
> Are there any missing features that io_uring can address?
> 
> Regards,
> Maxime
> 
> [0]: http://doc.dpdk.org/guides/howto/virtio_user_as_exception_path.html
> 

Yes, I looked at that but:
  - virtio user ends up with a busy kernel thread which is not acceptable
in SOC environment where all resources are locked down. In the SOC I was 
working
on DPDK was limited to 4 polling isolated CPU's and 1 sleeping main thread.
The rest of the CPU resources were hard constrained by cgroups. The virtio 
user
thread was a problem.

  - virtio user device is not persistent. If DPDK is being used a dataplane, 
need to
be able to quickly restart and not disturb applications and routing in the 
kernel
while the tap device is unavailable.  I.e having device present but in no 
carrier
state is better than having to teach applications about hot plug or play 
around
with multiple addresses on loopback device (which is what Cisco routers do).

Re: [v4,1/5] raw/gdtc: introduce gdtc raw device driver

On Tue, 29 Oct 2024 21:45:22 +0800
Yong Zhang  wrote:

> diff --git a/MAINTAINERS b/MAINTAINERS
> index c5a703b5c0..32fc4c801e 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -1511,6 +1511,11 @@ M: Gagandeep Singh 
>  F: drivers/raw/dpaa2_cmdif/
>  F: doc/guides/rawdevs/dpaa2_cmdif.rst
>  
> +ZTE GDTC
> +M: Yong Zhang 
> +F: drivers/raw/gdtc/
> +F: doc/guides/rawdevs/gdtc.rst
> +

Please add your mail address to .mailmap as well.

$ ./devtools/check-git-log.sh -n5
Contributor name/email mismatch with .mailmap: 
Yong Zhang  is unknown in .mailmap

Invalid patch(es) found - checked 5 patches

Re: [PATCH] raw/gdtc: introduce gdtc raw device driver