[PATCH v5 13/16] app/testpmd: pack structures when building with MSVC
From: Tyler Retzlaff
Remove __rte_packed where it appears natural alignment without
packing would produce the same layout or the struct is internal.
Signed-off-by: Tyler Retzlaff
---
app/test-pmd/csumonly.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/app/test-pmd/csumonly.c b/app/test-pmd/csumonly.c
index 2246c22e8e..d77a140641 100644
--- a/app/test-pmd/csumonly.c
+++ b/app/test-pmd/csumonly.c
@@ -92,7 +92,7 @@ struct testpmd_offload_info {
struct simple_gre_hdr {
uint16_t flags;
uint16_t proto;
-} __rte_packed;
+};
static uint16_t
get_udptcp_checksum(struct rte_mbuf *m, void *l3_hdr, uint16_t l4_off,
--
2.47.0.vfs.0.3
Re: 答复: [PATCH v8 12/17] net/r8169: implement Tx path
The change below is replacing 1 rte_smp_rmb with 2 calls,
so no it is not what I am asking for.
Please could you check how to not calling this function at all?
This series is already merged in the main branch,
so any new change should be submitted as a new patch.
Thank you
19/11/2024 07:29, 王颢:
> Dear Thomas,
>
> After our discussion, we concluded that we can make the following changes.
> What do you think? Additionally, I may not have been very clear in my last
> email. Should I submit the entire series of changes as PATCH v9, or should I
> respond with [PATCH v9 12/17] net/r8169: implement Tx path?
>
> static u32
> rtl_get_opts1(struct rtl_tx_desc *txd)
> {
> - rte_smp_rmb();
>
> return rte_le_to_cpu_32(txd->opts1);
> }
>
> static void
> rtl_tx_clean(struct rtl_hw *hw, struct rtl_tx_queue *txq)
> {
> ...
> tx_left = (rte_atomic_load_explicit(&txq->tx_tail,
> rte_memory_order_relaxed) %
> nb_tx_desc) - head;
> }
>
> + rte_smp_rmb();
>
> while (tx_left > 0) {
> txd = &txq->hw_ring[head];
>
> if (!enable_tx_no_close && (rtl_get_opts1(txd) & DescOwn))
> break;
> ...
> }
>
>
> int
> rtl_tx_done_cleanup(void *tx_queue, uint32_t free_cnt)
> {
> ...
> tx_left = (rte_atomic_load_explicit(&txq->tx_tail,
> rte_memory_order_relaxed) %
> nb_tx_desc) - head;
> }
>
> + rte_smp_rmb();
>
> while (tx_left > 0) {
> txd = &txq->hw_ring[head];
>
> status = rtl_get_opts1(txd);
>
> if (!enable_tx_no_close && (status & DescOwn))
> break;
>
> ...
> }
>
> Best Regards,
> Howard Wang
>
> -邮件原件-
> 发件人: Thomas Monjalon
> 发送时间: 2024年11月18日 22:59
> 收件人: pro_nic_d...@realtek.com; 王颢
> 抄送: dev@dpdk.org; Ferruh Yigit
> 主题: Re: [PATCH v8 12/17] net/r8169: implement Tx path
>
>
> External mail.
>
>
>
> Hello,
>
> 13/11/2024 10:28, Howard Wang:
> > +static u32
> > +rtl_get_opts1(struct rtl_tx_desc *txd) {
> > + rte_smp_rmb();
> > +
> > + return rte_le_to_cpu_32(txd->opts1); }
>
> We should avoid using such heavy memory barrier.
> Is it possible to use a lighter fence or atomic API?
>
>
>
Re: [PATCH v2] doc: add security document
Hi Nandini, One minor comment.. - Hemant On 18-11-2024 22:43, Nandini Persad wrote: This is a new document covering security protocols implemented in DPDK. +PDCP + + +Packet Data Convergence Protocol (PDCP) is a sublayer in the LTE radio protocol stack +that provides security and integrity protections to Protocol Data Units (PDU) in both +the control and data planes. PDCP is located between the Radio Link Control (RLC) layer +and the upper layers of the network, such as the IP layer. + + +**Wikipedia Link** +* https://en.wikipedia.org/wiki/Packet_Data_Convergence_Protocol + +**Standard Link** +* https://portal.3gpp.org/desktopmodules/Specifications/SpecificationDetails.aspx?specificationId=1177 + +**Level of Support in DPDK** +* Supported. High-level library: :doc:`PDCP Library <../prog_guide/pdcp_lib>` +* rte_security based PDCP sessions are also supported + +**Supported Algorithms** +* Encryption algo - NULL, AES-CTR, SNOW, ZUC +* Authentication algo - NULL, AES-CMAC, SNOW, ZUC + + +**Supported Drivers** +* Drivers supporting rte_security PDCP: +* XPdpaa2_sec, dpaa_sec Change it to NXP dpaa2_sec, dpaa_sec
答复: [PATCH v8 12/17] net/r8169: implement Tx path
Dear Thomas,
After our discussion, we concluded that we can make the following changes. What
do you think? Additionally, I may not have been very clear in my last email.
Should I submit the entire series of changes as PATCH v9, or should I respond
with [PATCH v9 12/17] net/r8169: implement Tx path?
static u32
rtl_get_opts1(struct rtl_tx_desc *txd)
{
- rte_smp_rmb();
return rte_le_to_cpu_32(txd->opts1);
}
static void
rtl_tx_clean(struct rtl_hw *hw, struct rtl_tx_queue *txq)
{
...
tx_left = (rte_atomic_load_explicit(&txq->tx_tail,
rte_memory_order_relaxed) %
nb_tx_desc) - head;
}
+ rte_smp_rmb();
while (tx_left > 0) {
txd = &txq->hw_ring[head];
if (!enable_tx_no_close && (rtl_get_opts1(txd) & DescOwn))
break;
...
}
int
rtl_tx_done_cleanup(void *tx_queue, uint32_t free_cnt)
{
...
tx_left = (rte_atomic_load_explicit(&txq->tx_tail,
rte_memory_order_relaxed) %
nb_tx_desc) - head;
}
+ rte_smp_rmb();
while (tx_left > 0) {
txd = &txq->hw_ring[head];
status = rtl_get_opts1(txd);
if (!enable_tx_no_close && (status & DescOwn))
break;
...
}
Best Regards,
Howard Wang
-邮件原件-
发件人: Thomas Monjalon
发送时间: 2024年11月18日 22:59
收件人: pro_nic_d...@realtek.com; 王颢
抄送: dev@dpdk.org; Ferruh Yigit
主题: Re: [PATCH v8 12/17] net/r8169: implement Tx path
External mail.
Hello,
13/11/2024 10:28, Howard Wang:
> +static u32
> +rtl_get_opts1(struct rtl_tx_desc *txd) {
> + rte_smp_rmb();
> +
> + return rte_le_to_cpu_32(txd->opts1); }
We should avoid using such heavy memory barrier.
Is it possible to use a lighter fence or atomic API?
[PATCH v5 14/16] app/test: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
app/test/test_efd.c| 1 +
app/test/test_hash.c | 1 +
app/test/test_member.c | 1 +
3 files changed, 3 insertions(+)
diff --git a/app/test/test_efd.c b/app/test/test_efd.c
index 1c0986b9bc..df176a0fbb 100644
--- a/app/test/test_efd.c
+++ b/app/test/test_efd.c
@@ -33,6 +33,7 @@ test_efd(void)
static unsigned int test_socket_id;
/* 5-tuple key type */
+__rte_msvc_pack
struct flow_key {
uint32_t ip_src;
uint32_t ip_dst;
diff --git a/app/test/test_hash.c b/app/test/test_hash.c
index 65b9cad93c..9b621ec6c3 100644
--- a/app/test/test_hash.c
+++ b/app/test/test_hash.c
@@ -79,6 +79,7 @@ static uint32_t hashtest_key_lens[] = {0, 2, 4, 5, 6, 7, 8,
10, 11, 15, 16, 21,
* Should be packed to avoid holes with potentially
* undefined content in the middle.
*/
+__rte_msvc_pack
struct flow_key {
uint32_t ip_src;
uint32_t ip_dst;
diff --git a/app/test/test_member.c b/app/test/test_member.c
index 5a4d2750db..ccd481a3a1 100644
--- a/app/test/test_member.c
+++ b/app/test/test_member.c
@@ -32,6 +32,7 @@ struct rte_member_setsum *setsum_vbf;
struct rte_member_setsum *setsum_sketch;
/* 5-tuple key type */
+__rte_msvc_pack
struct flow_key {
uint32_t ip_src;
uint32_t ip_dst;
--
2.47.0.vfs.0.3
[PATCH v5 15/16] examples: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Remove __rte_packed where it appears natural alignment without
packing would produce the same layout or the struct is internal.
Signed-off-by: Tyler Retzlaff
---
examples/common/neon/port_group.h | 1 +
examples/ip_pipeline/cli.c| 10 +-
examples/ipsec-secgw/ipsec.h | 1 +
examples/l3fwd-power/main.c | 4 +++-
examples/l3fwd/l3fwd_route.h | 2 ++
examples/ptpclient/ptpclient.c| 8
examples/vhost_blk/blk_spec.h | 1 +
7 files changed, 21 insertions(+), 6 deletions(-)
diff --git a/examples/common/neon/port_group.h
b/examples/common/neon/port_group.h
index 421e2e8613..bd40590418 100644
--- a/examples/common/neon/port_group.h
+++ b/examples/common/neon/port_group.h
@@ -21,6 +21,7 @@ static inline uint16_t *
port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp, uint16x8_t dp1,
uint16x8_t dp2)
{
+ __rte_msvc_pack
union {
uint16_t u16[FWDSTEP + 1];
uint64_t u64;
diff --git a/examples/ip_pipeline/cli.c b/examples/ip_pipeline/cli.c
index 92dfacdeb0..766fc8e46e 100644
--- a/examples/ip_pipeline/cli.c
+++ b/examples/ip_pipeline/cli.c
@@ -2571,7 +2571,7 @@ struct pkt_key_qinq {
uint16_t svlan;
uint16_t ethertype_cvlan;
uint16_t cvlan;
-} __rte_packed;
+};
struct pkt_key_ipv4_5tuple {
uint8_t time_to_live;
@@ -2581,7 +2581,7 @@ struct pkt_key_ipv4_5tuple {
uint32_t da;
uint16_t sp;
uint16_t dp;
-} __rte_packed;
+};
struct pkt_key_ipv6_5tuple {
uint16_t payload_length;
@@ -2591,15 +2591,15 @@ struct pkt_key_ipv6_5tuple {
struct rte_ipv6_addr da;
uint16_t sp;
uint16_t dp;
-} __rte_packed;
+};
struct pkt_key_ipv4_addr {
uint32_t addr;
-} __rte_packed;
+};
struct pkt_key_ipv6_addr {
struct rte_ipv6_addr addr;
-} __rte_packed;
+};
static uint32_t
parse_match(char **tokens,
diff --git a/examples/ipsec-secgw/ipsec.h b/examples/ipsec-secgw/ipsec.h
index f12f57e2d5..dc146bccb8 100644
--- a/examples/ipsec-secgw/ipsec.h
+++ b/examples/ipsec-secgw/ipsec.h
@@ -274,6 +274,7 @@ struct socket_ctx {
struct rte_mempool *session_pool;
};
+__rte_msvc_pack
struct cnt_blk {
uint32_t salt;
uint64_t iv;
diff --git a/examples/l3fwd-power/main.c b/examples/l3fwd-power/main.c
index ae8b55924e..ce3c901b45 100644
--- a/examples/l3fwd-power/main.c
+++ b/examples/l3fwd-power/main.c
@@ -282,7 +282,8 @@ static struct rte_mempool * pktmbuf_pool[NB_SOCKETS];
#define DEFAULT_HASH_FUNC rte_jhash
#endif
-struct ipv4_5tuple {
+__rte_msvc_pack
+__rte_msvc_packstruct ipv4_5tuple {
uint32_t ip_dst;
uint32_t ip_src;
uint16_t port_dst;
@@ -290,6 +291,7 @@ struct ipv4_5tuple {
uint8_t proto;
} __rte_packed;
+__rte_msvc_pack
struct ipv6_5tuple {
uint8_t ip_dst[IPV6_ADDR_LEN];
uint8_t ip_src[IPV6_ADDR_LEN];
diff --git a/examples/l3fwd/l3fwd_route.h b/examples/l3fwd/l3fwd_route.h
index 62263c3540..8c7be4b9d6 100644
--- a/examples/l3fwd/l3fwd_route.h
+++ b/examples/l3fwd/l3fwd_route.h
@@ -36,6 +36,7 @@ struct ipv6_l3fwd_route {
uint8_t if_out;
};
+__rte_msvc_pack
struct ipv4_5tuple {
uint32_t ip_dst;
uint32_t ip_src;
@@ -44,6 +45,7 @@ struct ipv4_5tuple {
uint8_t proto;
} __rte_packed;
+__rte_msvc_pack
struct ipv6_5tuple {
uint8_t ip_dst[IPV6_ADDR_LEN];
uint8_t ip_src[IPV6_ADDR_LEN];
diff --git a/examples/ptpclient/ptpclient.c b/examples/ptpclient/ptpclient.c
index 23fa487081..d7caad41d3 100644
--- a/examples/ptpclient/ptpclient.c
+++ b/examples/ptpclient/ptpclient.c
@@ -85,6 +85,7 @@ static const struct rte_ether_addr ether_multicast = {
};
/* Structs used for PTP handling. */
+__rte_msvc_pack
struct tstamp {
uint16_t sec_msb;
uint32_t sec_lsb;
@@ -95,11 +96,13 @@ struct clock_id {
uint8_t id[8];
};
+__rte_msvc_pack
struct port_id {
struct clock_idclock_id;
uint16_t port_number;
} __rte_packed;
+__rte_msvc_pack
struct ptp_header {
uint8_t msg_type;
uint8_t ver;
@@ -115,22 +118,26 @@ struct ptp_header {
int8_t log_message_interval;
} __rte_packed;
+__rte_msvc_pack
struct sync_msg {
struct ptp_header hdr;
struct tstamp origin_tstamp;
} __rte_packed;
+__rte_msvc_pack
struct follow_up_msg {
struct ptp_header hdr;
struct tstamp precise_origin_tstamp;
uint8_t suffix[];
} __rte_packed;
+__rte_msvc_pack
struct delay_req_msg {
struct ptp_header hdr;
struct tstamp origin_tstamp;
} __rte_packed;
+__rte_msvc_pack
struct delay_resp_msg {
struct ptp_headerhdr;
struct tstamprx_tstamp;
@@ -139,6 +146,7 @@
[PATCH v5 12/16] net/octeon_ep: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
drivers/net/octeon_ep/otx_ep_mbox.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/net/octeon_ep/otx_ep_mbox.h
b/drivers/net/octeon_ep/otx_ep_mbox.h
index ec96e4edc5..44da0fa84f 100644
--- a/drivers/net/octeon_ep/otx_ep_mbox.h
+++ b/drivers/net/octeon_ep/otx_ep_mbox.h
@@ -76,6 +76,7 @@ enum otx_ep_link_autoneg {
#define OTX_EP_MBOX_MORE_FRAG_FLAG 1
#define OTX_EP_MBOX_WRITE_WAIT_TIME msecs_to_jiffies(1)
+__rte_msvc_pack
union otx_ep_mbox_word {
uint64_t u64;
struct {
--
2.47.0.vfs.0.3
[PATCH v2 1/5] net/bnxt/tf_ulp: fix F1F2 vxlan counter acccumulation for Thor2
From: Shuanglin Wang
Add code for Thor2 to support the counter accumulation for
F1F2 vxlan parent-child flows.
Also, add a check for device state in the TF tunnel free API,
if it is zero then TF will skip the resource free as it
was already done by bnxt_free_all_hwrm_resources.
Fixes: 0513f0af034d ("net/bnxt/tf_ulp: add stats cache for Thor2")
Reviewed-by: Kishore Padmanabha
Signed-off-by: Shuanglin Wang
Signed-off-by: Sangtani Parag Satishbhai
Signed-off-by: Sriharsha Basavapatna
---
drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c | 12 ++-
drivers/net/bnxt/tf_ulp/ulp_flow_db.c | 4 +-
drivers/net/bnxt/tf_ulp/ulp_flow_db.h | 13 +++
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c | 110 +++--
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.h | 18 +++-
5 files changed, 142 insertions(+), 15 deletions(-)
diff --git a/drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c
b/drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c
index c8f3a5abc2..5854f13447 100644
--- a/drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c
+++ b/drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c
@@ -587,10 +587,16 @@ bnxt_pmd_global_tunnel_set(struct bnxt_ulp_context
*ulp_ctx,
return -EINVAL;
}
- if (udp_port)
+ if (udp_port) {
rc = bnxt_udp_tunnel_port_add_op(eth_dev, &udp_tunnel);
- else
- rc = bnxt_udp_tunnel_port_del_op(eth_dev, &udp_tunnel);
+ } else {
+ /* TODO: Make the counters shareable so the resource
+* free can be synced up between core dpdk path and
+* the tf path.
+*/
+ if (eth_dev->data->dev_started != 0)
+ rc = bnxt_udp_tunnel_port_del_op(eth_dev,
&udp_tunnel);
+ }
} else {
bp = bnxt_pmd_get_bp(port_id);
if (!bp) {
diff --git a/drivers/net/bnxt/tf_ulp/ulp_flow_db.c
b/drivers/net/bnxt/tf_ulp/ulp_flow_db.c
index d6fd653b4a..943c9b799c 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_flow_db.c
+++ b/drivers/net/bnxt/tf_ulp/ulp_flow_db.c
@@ -1176,7 +1176,7 @@ ulp_default_flow_db_cfa_action_get(struct
bnxt_ulp_context *ulp_ctx,
}
/* internal validation function for parent flow tbl */
-static struct ulp_fdb_parent_info *
+struct ulp_fdb_parent_info *
ulp_flow_db_pc_db_entry_get(struct bnxt_ulp_context *ulp_ctxt,
uint32_t pc_idx)
{
@@ -1634,6 +1634,8 @@ ulp_flow_db_parent_flow_create(struct
bnxt_ulp_mapper_parms *parms)
}
}
+ /* Set parent flow entry idx in stats cache entry */
+ ulp_sc_mgr_set_pc_idx(parms->ulp_ctx, parms->flow_id, pc_idx);
return 0;
}
diff --git a/drivers/net/bnxt/tf_ulp/ulp_flow_db.h
b/drivers/net/bnxt/tf_ulp/ulp_flow_db.h
index 762ebb039d..4301094cfe 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_flow_db.h
+++ b/drivers/net/bnxt/tf_ulp/ulp_flow_db.h
@@ -417,4 +417,17 @@ void ulp_flow_db_shared_session_set(struct
ulp_flow_db_res_params *res,
*/
enum bnxt_ulp_session_type
ulp_flow_db_shared_session_get(struct ulp_flow_db_res_params *res);
+
+/*
+ * Get the parent flow table info
+ *
+ * ulp_ctxt [in] Ptr to ulp_context
+ * pc_idx [in] The index to parent child db
+ *
+ * returns Pointer of parent flow tbl
+ */
+struct ulp_fdb_parent_info *
+ulp_flow_db_pc_db_entry_get(struct bnxt_ulp_context *ulp_ctxt,
+ uint32_t pc_idx);
+
#endif /* _ULP_FLOW_DB_H_ */
diff --git a/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
b/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
index 13069126f0..1317668555 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
+++ b/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
@@ -441,8 +441,17 @@ int ulp_sc_mgr_query_count_get(struct bnxt_ulp_context
*ctxt,
{
struct ulp_sc_tfc_stats_cache_entry *sce;
struct bnxt_ulp_sc_info *ulp_sc_info;
+ struct ulp_fdb_parent_info *pc_entry;
+ struct bnxt_ulp_flow_db *flow_db;
+ uint32_t max_array;
+ uint32_t child_fid;
+ uint32_t a_idx;
+ uint32_t f2_cnt;
+ uint64_t *t;
+ uint64_t bs;
int rc = 0;
+ /* Get stats cache info */
ulp_sc_info = bnxt_ulp_cntxt_ptr2_sc_info_get(ctxt);
if (!ulp_sc_info)
return -ENODEV;
@@ -450,18 +459,66 @@ int ulp_sc_mgr_query_count_get(struct bnxt_ulp_context
*ctxt,
sce = ulp_sc_info->stats_cache_tbl;
sce += flow_id;
- /* If entry is not valid return an error */
- if (!(sce->flags & ULP_SC_ENTRY_FLAG_VALID))
- return -EBUSY;
-
- count->hits = sce->packet_count;
- count->hits_set = 1;
- count->bytes = sce->byte_count;
- count->bytes_set = 1;
+ /* To handle the parent flow */
+ if (sce->flags & ULP_SC_ENTRY_FLAG_PARENT) {
+ flow_db = bnxt_ulp_cntxt_ptr2_flow_db_get(ctxt);
+ if (!flow_db) {
+
[PATCH v2 0/5] TruFlow fixes for Thor2
This patch series fixes a few issues in TruFlow for Broadcom Thor2 NIC. ** v1->v2: - Added patch-5 to fix coverity errors ** Kishore Padmanabha (2): net/bnxt/tf_ulp: fix vfr clean up and stats lockup net/bnxt/tf_ulp: update template files Peter Spreadborough (2): net/bnxt/tf_ulp: performance and tuning changes for thor2 stats cache net/bnxt/tf_ulp: fix coverity errors Shuanglin Wang (1): net/bnxt/tf_ulp: fix F1F2 vxlan counter acccumulation for Thor2 drivers/net/bnxt/hcapi/cfa_v3/mm/cfa_mm.c | 6 +- drivers/net/bnxt/tf_core/v3/tfc.h |12 +- drivers/net/bnxt/tf_core/v3/tfc_act.c |26 +- drivers/net/bnxt/tf_core/v3/tfc_em.c | 1 + drivers/net/bnxt/tf_core/v3/tfc_tbl_scope.c | 2 +- drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c|12 +- drivers/net/bnxt/tf_ulp/bnxt_ulp.c|17 + drivers/net/bnxt/tf_ulp/bnxt_ulp.h| 3 + drivers/net/bnxt/tf_ulp/bnxt_ulp_utils.h | 3 +- .../generic_templates/ulp_template_db_enum.h | 8 +- .../ulp_template_db_thor2_class.c | 10126 .../ulp_template_db_thor_class.c | 8736 +++-- drivers/net/bnxt/tf_ulp/ulp_fc_mgr_tfc.c | 5 +- drivers/net/bnxt/tf_ulp/ulp_flow_db.c | 4 +- drivers/net/bnxt/tf_ulp/ulp_flow_db.h |13 + drivers/net/bnxt/tf_ulp/ulp_mapper.c |10 +- drivers/net/bnxt/tf_ulp/ulp_rte_parser.c | 7 +- drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c | 309 +- drivers/net/bnxt/tf_ulp/ulp_sc_mgr.h |33 +- drivers/net/bnxt/tf_ulp/ulp_sc_mgr_tfc.c | 4 +- 20 files changed, 9650 insertions(+), 9687 deletions(-) -- 2.39.3
[PATCH v2 2/5] net/bnxt/tf_ulp: fix vfr clean up and stats lockup
From: Kishore Padmanabha
The representor flows were not being deleted as part of the
vfr clean up. Added code to delete flows related to vfr interface.
Also fixed the stats counter thread lockup.
Fixes: 0513f0af034d ("net/bnxt/tf_ulp: add stats cache for Thor2")
Reviewed-by: Peter Spreadborough
Reviewed-by: Shuanglin Wang
Signed-off-by: Kishore Padmanabha
Signed-off-by: Sriharsha Basavapatna
---
drivers/net/bnxt/tf_ulp/bnxt_ulp.c | 17 +++
drivers/net/bnxt/tf_ulp/bnxt_ulp.h | 3 +
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c | 164 ---
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.h | 12 +-
4 files changed, 74 insertions(+), 122 deletions(-)
diff --git a/drivers/net/bnxt/tf_ulp/bnxt_ulp.c
b/drivers/net/bnxt/tf_ulp/bnxt_ulp.c
index e28a481f5e..1bfc88cf79 100644
--- a/drivers/net/bnxt/tf_ulp/bnxt_ulp.c
+++ b/drivers/net/bnxt/tf_ulp/bnxt_ulp.c
@@ -596,6 +596,9 @@ bnxt_ulp_port_deinit(struct bnxt *bp)
/* close the session associated with this port */
bp->ulp_ctx->ops->ulp_ctx_detach(bp, session);
} else {
+ /* Free the ulp context in the context entry list */
+ bnxt_ulp_cntxt_list_del(bp->ulp_ctx);
+
/* clean up default flows */
bnxt_ulp_destroy_df_rules(bp, true);
@@ -662,6 +665,20 @@ bnxt_ulp_cntxt_list_del(struct bnxt_ulp_context *ulp_ctx)
rte_spinlock_unlock(&bnxt_ulp_ctxt_lock);
}
+int
+bnxt_ulp_cntxt_list_count(void)
+{
+ struct ulp_context_list_entry *entry, *temp;
+ int count_1 = 0;
+
+ rte_spinlock_lock(&bnxt_ulp_ctxt_lock);
+ RTE_TAILQ_FOREACH_SAFE(entry, &ulp_cntx_list, next, temp) {
+ count_1++;
+ }
+ rte_spinlock_unlock(&bnxt_ulp_ctxt_lock);
+ return count_1;
+}
+
struct bnxt_ulp_context *
bnxt_ulp_cntxt_entry_acquire(void *arg)
{
diff --git a/drivers/net/bnxt/tf_ulp/bnxt_ulp.h
b/drivers/net/bnxt/tf_ulp/bnxt_ulp.h
index 83fb205f68..e0e31532fd 100644
--- a/drivers/net/bnxt/tf_ulp/bnxt_ulp.h
+++ b/drivers/net/bnxt/tf_ulp/bnxt_ulp.h
@@ -297,6 +297,9 @@ bnxt_ulp_cntxt_list_add(struct bnxt_ulp_context *ulp_ctx);
void
bnxt_ulp_cntxt_list_del(struct bnxt_ulp_context *ulp_ctx);
+int
+bnxt_ulp_cntxt_list_count(void);
+
struct bnxt_ulp_context *
bnxt_ulp_cntxt_entry_acquire(void *arg);
diff --git a/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
b/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
index 1317668555..c82fdaf6dd 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
+++ b/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
@@ -89,12 +89,6 @@ int32_t ulp_sc_mgr_init(struct bnxt_ulp_context *ctxt)
ulp_sc_info->sc_ops = sc_ops;
ulp_sc_info->flags = 0;
- rc = pthread_mutex_init(&ulp_sc_info->sc_lock, NULL);
- if (rc) {
- BNXT_DRV_DBG(ERR, "Failed to initialize sc mutex\n");
- goto error;
- }
-
/* Add the SC info tbl to the ulp context. */
bnxt_ulp_cntxt_ptr2_sc_info_set(ctxt, ulp_sc_info);
@@ -109,9 +103,10 @@ int32_t ulp_sc_mgr_init(struct bnxt_ulp_context *ctxt)
* Size is determined by the number of flows + 10% to cover IDs
* used for resources.
*/
+ ulp_sc_info->cache_tbl_size = ulp_sc_info->num_counters +
+ (ulp_sc_info->num_counters / 10);
stats_cache_tbl_sz = sizeof(struct ulp_sc_tfc_stats_cache_entry) *
- (ulp_sc_info->num_counters +
-(ulp_sc_info->num_counters / 10));
+ ulp_sc_info->cache_tbl_size;
ulp_sc_info->stats_cache_tbl = rte_zmalloc("ulp_stats_cache_tbl",
stats_cache_tbl_sz, 0);
@@ -153,12 +148,6 @@ ulp_sc_mgr_deinit(struct bnxt_ulp_context *ctxt)
if (!ulp_sc_info)
return -EINVAL;
- pthread_mutex_lock(&ulp_sc_info->sc_lock);
-
- ulp_sc_mgr_thread_cancel(ctxt);
-
- pthread_mutex_destroy(&ulp_sc_info->sc_lock);
-
if (ulp_sc_info->stats_cache_tbl)
rte_free(ulp_sc_info->stats_cache_tbl);
@@ -173,13 +162,13 @@ ulp_sc_mgr_deinit(struct bnxt_ulp_context *ctxt)
return 0;
}
-#define ULP_SC_PERIOD_S 1
-#define ULP_SC_PERIOD_MS (ULP_SC_PERIOD_S * 1000)
+#define ULP_SC_PERIOD_US 256
+#define ULP_SC_CTX_DELAY 1
static uint32_t ulp_stats_cache_main_loop(void *arg)
{
struct ulp_sc_tfc_stats_cache_entry *count;
- const struct bnxt_ulp_sc_core_ops *sc_ops;
+ const struct bnxt_ulp_sc_core_ops *sc_ops = NULL;
struct ulp_sc_tfc_stats_cache_entry *sce;
struct ulp_sc_tfc_stats_cache_entry *sce_end;
struct tfc_mpc_batch_info_t batch_info;
@@ -188,95 +177,68 @@ static uint32_t ulp_stats_cache_main_loop(void *arg)
uint16_t words = (ULP_TFC_CNTR_READ_BYTES + ULP_TFC_ACT_WORD_SZ - 1) /
ULP_TFC_ACT_WORD_SZ;
uint32_t batch_size;
struct tfc *tfcp = NULL;
- uint32_t batch;
- uint32_t delay = ULP
[PATCH v2 3/5] net/bnxt/tf_ulp: performance and tuning changes for thor2 stats cache
From: Peter Spreadborough
This change adds changes the act get API to expect a physical
address for the host memory rather than a pointer to a local
virtual address. The change was made because the virt->phys
API call has a very high overhead. Also included in this change
to the mutex locks and delays within the main loop.
Fixes: 0513f0af034d ("net/bnxt/tf_ulp: add stats cache for Thor2")
Reviewed-by: Farah Smith
Reviewed-by: Jay Ding
Signed-off-by: Peter Spreadborough
Signed-off-by: Sriharsha Basavapatna
---
drivers/net/bnxt/tf_core/v3/tfc.h| 12 ---
drivers/net/bnxt/tf_core/v3/tfc_act.c| 26 ++
drivers/net/bnxt/tf_ulp/ulp_fc_mgr_tfc.c | 5 ++-
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c | 45
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.h | 3 +-
drivers/net/bnxt/tf_ulp/ulp_sc_mgr_tfc.c | 4 +--
6 files changed, 65 insertions(+), 30 deletions(-)
diff --git a/drivers/net/bnxt/tf_core/v3/tfc.h
b/drivers/net/bnxt/tf_core/v3/tfc.h
index 98725b32c2..4176aa01bd 100644
--- a/drivers/net/bnxt/tf_core/v3/tfc.h
+++ b/drivers/net/bnxt/tf_core/v3/tfc.h
@@ -705,7 +705,9 @@ enum tfc_tbl_scope_bucket_factor {
TFC_TBL_SCOPE_BUCKET_FACTOR_4 = 4,
TFC_TBL_SCOPE_BUCKET_FACTOR_8 = 8,
TFC_TBL_SCOPE_BUCKET_FACTOR_16 = 16,
- TFC_TBL_SCOPE_BUCKET_FACTOR_MAX = TFC_TBL_SCOPE_BUCKET_FACTOR_16
+ TFC_TBL_SCOPE_BUCKET_FACTOR_32 = 32,
+ TFC_TBL_SCOPE_BUCKET_FACTOR_64 = 64,
+ TFC_TBL_SCOPE_BUCKET_FACTOR_MAX = TFC_TBL_SCOPE_BUCKET_FACTOR_64
};
/**
@@ -1420,8 +1422,9 @@ int tfc_act_set(struct tfc *tfcp,
* @param[in] cmm_clr
* Pointer to cmm clr
*
- * @param[in,out] data
- * Data read. Must be word aligned, i.e. [1:0] must be 0.
+ * @param[in,out] host_address
+ * Data read. Must be word aligned, i.e. [1:0] must be 0. The address
+ * must be the ret_mem_virt2iova() version of the virt address.
*
* @param[in,out] data_sz_words
* Data buffer size in words.Size could be 8/16/24/32/64B
@@ -1437,7 +1440,8 @@ int tfc_act_get(struct tfc *tfcp,
struct tfc_mpc_batch_info_t *batch_info,
const struct tfc_cmm_info *cmm_info,
struct tfc_cmm_clr *clr,
- uint8_t *data, uint16_t *data_sz_words);
+ uint64_t *host_address,
+ uint16_t *data_sz_words);
/**
* Free a CMM Resource
diff --git a/drivers/net/bnxt/tf_core/v3/tfc_act.c
b/drivers/net/bnxt/tf_core/v3/tfc_act.c
index 3e215f4881..0e98bd30d7 100644
--- a/drivers/net/bnxt/tf_core/v3/tfc_act.c
+++ b/drivers/net/bnxt/tf_core/v3/tfc_act.c
@@ -367,7 +367,7 @@ int tfc_act_get_only_response(struct cfa_bld_mpcinfo
*mpc_info,
static int tfc_act_get_only(struct tfc *tfcp,
struct tfc_mpc_batch_info_t *batch_info,
const struct tfc_cmm_info *cmm_info,
- uint8_t *data,
+ uint64_t *host_address,
uint16_t *data_sz_words)
{
int rc = 0;
@@ -378,7 +378,6 @@ static int tfc_act_get_only(struct tfc *tfcp,
uint32_t buff_len;
struct cfa_mpc_data_obj fields_cmd[CFA_BLD_MPC_READ_CMD_MAX_FLD] = {
{0} };
uint32_t entry_offset;
- uint64_t host_address;
struct bnxt_mpc_mbuf mpc_msg_in;
struct bnxt_mpc_mbuf mpc_msg_out;
uint32_t record_size;
@@ -405,13 +404,11 @@ static int tfc_act_get_only(struct tfc *tfcp,
}
/* Check that data pointer is word aligned */
- if (unlikely(((uint64_t)data) & 0x3ULL)) {
+ if (unlikely(*host_address & 0x3ULL)) {
PMD_DRV_LOG_LINE(ERR, "data pointer not word aligned");
return -EINVAL;
}
- host_address = (uint64_t)rte_mem_virt2iova(data);
-
/* Check that MPC APIs are bound */
if (unlikely(mpc_info->mpcops == NULL)) {
PMD_DRV_LOG_LINE(ERR, "MPC not initialized");
@@ -450,7 +447,7 @@ static int tfc_act_get_only(struct tfc *tfcp,
fields_cmd[CFA_BLD_MPC_READ_CMD_HOST_ADDRESS_FLD].field_id =
CFA_BLD_MPC_READ_CMD_HOST_ADDRESS_FLD;
- fields_cmd[CFA_BLD_MPC_READ_CMD_HOST_ADDRESS_FLD].val = host_address;
+ fields_cmd[CFA_BLD_MPC_READ_CMD_HOST_ADDRESS_FLD].val = *host_address;
buff_len = TFC_MPC_MAX_TX_BYTES;
@@ -547,7 +544,7 @@ int tfc_act_get_clear_response(struct cfa_bld_mpcinfo
*mpc_info,
static int tfc_act_get_clear(struct tfc *tfcp,
struct tfc_mpc_batch_info_t *batch_info,
const struct tfc_cmm_info *cmm_info,
-uint8_t *data,
+uint64_t *host_address,
uint16_t *data_sz_words,
uint8_t clr_offset,
uint8_t clr_size)
@@ -560,7 +557,6 @@ static int tfc_act_get_clear(struct tfc *tfcp,
uint32_t buff_len;
struct cfa_mpc_d
[PATCH v2 5/5] net/bnxt/tf_ulp: fix coverity errors
From: Peter Spreadborough
This patch fixes a few coverity errors reported in the feature
patchset (merged) to support TruFlow on Thor2.
Fixes: 0513f0af034d ("net/bnxt/tf_ulp: add stats cache for Thor2")
Signed-off-by: Peter Spreadborough
Signed-off-by: Sriharsha Basavapatna
---
drivers/net/bnxt/hcapi/cfa_v3/mm/cfa_mm.c | 6 +-
drivers/net/bnxt/tf_core/v3/tfc_em.c| 1 +
drivers/net/bnxt/tf_core/v3/tfc_tbl_scope.c | 2 +-
drivers/net/bnxt/tf_ulp/bnxt_ulp_utils.h| 3 ++-
drivers/net/bnxt/tf_ulp/ulp_mapper.c| 10 +++---
drivers/net/bnxt/tf_ulp/ulp_rte_parser.c| 7 ++-
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c| 2 ++
7 files changed, 24 insertions(+), 7 deletions(-)
diff --git a/drivers/net/bnxt/hcapi/cfa_v3/mm/cfa_mm.c
b/drivers/net/bnxt/hcapi/cfa_v3/mm/cfa_mm.c
index 61fafadb20..05528dd3e4 100644
--- a/drivers/net/bnxt/hcapi/cfa_v3/mm/cfa_mm.c
+++ b/drivers/net/bnxt/hcapi/cfa_v3/mm/cfa_mm.c
@@ -123,7 +123,11 @@ int cfa_mm_open(void *cmm, struct cfa_mm_open_parms *parms)
}
for (i = 0; i < num_blocks; i++) {
- context->blk_tbl[i].prev_blk_idx = i - 1;
+ if (i == 0)
+ context->blk_tbl[i].prev_blk_idx = CFA_MM_INVALID32;
+ else
+ context->blk_tbl[i].prev_blk_idx = i - 1;
+
context->blk_tbl[i].next_blk_idx = i + 1;
context->blk_tbl[i].num_free_records = records_per_block;
context->blk_tbl[i].first_free_record = 0;
diff --git a/drivers/net/bnxt/tf_core/v3/tfc_em.c
b/drivers/net/bnxt/tf_core/v3/tfc_em.c
index a70e35b6b1..d460ff2ee0 100644
--- a/drivers/net/bnxt/tf_core/v3/tfc_em.c
+++ b/drivers/net/bnxt/tf_core/v3/tfc_em.c
@@ -560,6 +560,7 @@ int tfc_em_delete_raw(struct tfc *tfcp,
mpc_msg_out.cmp_type = CMPL_BASE_TYPE_MID_PATH_LONG;
mpc_msg_out.msg_data = &rx_msg[TFC_MPC_HEADER_SIZE_BYTES];
mpc_msg_out.msg_size = TFC_MPC_MAX_RX_BYTES;
+ mpc_msg_out.chnl_id = 0;
rc = tfc_mpc_send(tfcp->bp,
&mpc_msg_in,
diff --git a/drivers/net/bnxt/tf_core/v3/tfc_tbl_scope.c
b/drivers/net/bnxt/tf_core/v3/tfc_tbl_scope.c
index 1770069295..c29933b803 100644
--- a/drivers/net/bnxt/tf_core/v3/tfc_tbl_scope.c
+++ b/drivers/net/bnxt/tf_core/v3/tfc_tbl_scope.c
@@ -468,7 +468,7 @@ static int alloc_link_pbl(struct tfc_ts_mem_cfg *mem_cfg,
uint32_t page_size,
* and page tables. The allocation will occur once only per backing
* store and will located by name and reused on subsequent runs.
*/
- total_size = page_size * total_pages;
+ total_size = (uint64_t)page_size * (uint64_t)total_pages;
if (total_size <= (1024 * 256))
mz_size = RTE_MEMZONE_256KB;
diff --git a/drivers/net/bnxt/tf_ulp/bnxt_ulp_utils.h
b/drivers/net/bnxt/tf_ulp/bnxt_ulp_utils.h
index 5e0d906fbd..e849df2713 100644
--- a/drivers/net/bnxt/tf_ulp/bnxt_ulp_utils.h
+++ b/drivers/net/bnxt/tf_ulp/bnxt_ulp_utils.h
@@ -1084,8 +1084,9 @@ bnxt_ulp_cap_feat_process(uint64_t feat_bits, uint64_t
*out_bits)
if (bit & BNXT_ULP_FEATURE_BIT_PARENT_DMAC)
BNXT_DRV_DBG(ERR, "Parent Mac Address Feature is enabled\n");
- if (bit & BNXT_ULP_FEATURE_BIT_PORT_DMAC)
+ else if (bit & BNXT_ULP_FEATURE_BIT_PORT_DMAC)
BNXT_DRV_DBG(ERR, "Port Mac Address Feature is enabled\n");
+
if (bit & BNXT_ULP_FEATURE_BIT_MULTI_TUNNEL_FLOW)
BNXT_DRV_DBG(ERR, "Multi Tunnel Flow Feature is enabled\n");
diff --git a/drivers/net/bnxt/tf_ulp/ulp_mapper.c
b/drivers/net/bnxt/tf_ulp/ulp_mapper.c
index 2429ac2f1a..1a68cf5dfd 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_mapper.c
+++ b/drivers/net/bnxt/tf_ulp/ulp_mapper.c
@@ -3612,9 +3612,13 @@ ulp_mapper_func_cond_list_process(struct
bnxt_ulp_mapper_parms *parms,
}
}
/* write the value into result */
- ulp_operand_read(val, res_local + res_size -
-ULP_BITS_2_BYTE_NR(oper_size),
-ULP_BITS_2_BYTE_NR(val_len));
+ if (unlikely(ulp_operand_read(val, res_local + res_size -
+ ULP_BITS_2_BYTE_NR(oper_size),
+ ULP_BITS_2_BYTE_NR(val_len {
+ BNXT_DRV_DBG(ERR,
+"field idx operand read failed\n");
+ return -EINVAL;
+ }
/* convert the data to cpu format */
*res = tfp_be_to_cpu_64(*res);
diff --git a/drivers/net/bnxt/tf_ulp/ulp_rte_parser.c
b/drivers/net/bnxt/tf_ulp/ulp_rte_parser.c
index dd5985cd7b..f75606ca81 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_rte_parser.c
+++ b/drivers/net/bnxt/tf_ulp/ulp_rte_parser.c
@@ -517,7 +517,12 @@ ulp_rte_parser_svif_set(struct ulp_rte_parser_params
*params,
else
svif_type = BNXT_ULP_DRV_FUNC_SVIF;
}
- ulp_port_db_svif_get(params->ulp_ctx, ifindex, s
Re: [PATCH 16/16] net/dpaa2: fix bitmask truncation
Acked-by: Hemant Agrawal
On 15-11-2024 11:35, Stephen Hemminger wrote:
The dqrr_held mask is 64 bit but updates were getting truncated
because 1 is of type int (32 bit) and the result shift of int is of
type int (32 bit); therefore any value >= 32 would get truncated.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: 2d3788631862 ("net/dpaa2: support atomic queues")
Cc: nipun.gu...@nxp.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
drivers/net/dpaa2/dpaa2_rxtx.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/dpaa2/dpaa2_rxtx.c b/drivers/net/dpaa2/dpaa2_rxtx.c
index e3b6c7e460..e253bccecd 100644
--- a/drivers/net/dpaa2/dpaa2_rxtx.c
+++ b/drivers/net/dpaa2/dpaa2_rxtx.c
@@ -933,7 +933,7 @@ dpaa2_dev_process_atomic_event(struct qbman_swp *swp
__rte_unused,
dqrr_index = qbman_get_dqrr_idx(dq);
*dpaa2_seqn(ev->mbuf) = dqrr_index + 1;
DPAA2_PER_LCORE_DQRR_SIZE++;
- DPAA2_PER_LCORE_DQRR_HELD |= 1 << dqrr_index;
+ DPAA2_PER_LCORE_DQRR_HELD |= UINT64_C(1) << dqrr_index;
DPAA2_PER_LCORE_DQRR_MBUF(dqrr_index) = ev->mbuf;
}
@@ -1317,7 +1317,7 @@ dpaa2_dev_tx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
flags[loop] = QBMAN_ENQUEUE_FLAG_DCA |
dqrr_index;
DPAA2_PER_LCORE_DQRR_SIZE--;
- DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dqrr_index);
+ DPAA2_PER_LCORE_DQRR_HELD &= ~(UINT64_C(1) <<
dqrr_index);
*dpaa2_seqn(*bufs) = DPAA2_INVALID_MBUF_SEQN;
}
@@ -1575,7 +1575,7 @@ dpaa2_set_enqueue_descriptor(struct dpaa2_queue *dpaa2_q,
dq_idx = *dpaa2_seqn(m) - 1;
qbman_eq_desc_set_dca(eqdesc, 1, dq_idx, 0);
DPAA2_PER_LCORE_DQRR_SIZE--;
- DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dq_idx);
+ DPAA2_PER_LCORE_DQRR_HELD &= ~(UINT64_C(1) << dq_idx);
}
*dpaa2_seqn(m) = DPAA2_INVALID_MBUF_SEQN;
}
[PATCH v3 1/2] net/bnxt: update HWRM API
Update HWRM API to select ring profile.
Signed-off-by: Ajit Khaparde
---
drivers/net/bnxt/hsi_struct_def_dpdk.h | 71 +-
1 file changed, 69 insertions(+), 2 deletions(-)
diff --git a/drivers/net/bnxt/hsi_struct_def_dpdk.h
b/drivers/net/bnxt/hsi_struct_def_dpdk.h
index 8f348c20fb..737bf2693b 100644
--- a/drivers/net/bnxt/hsi_struct_def_dpdk.h
+++ b/drivers/net/bnxt/hsi_struct_def_dpdk.h
@@ -15617,7 +15617,46 @@ struct hwrm_func_qcaps_output {
* (SR-IOV) disabled or on a VF.
*/
uint32_troce_vf_max_gid;
- uint8_t unused_3[3];
+ uint32_tflags_ext3;
+ /*
+* When this bit is '1', firmware supports the driver using
+* FUNC_CFG (or FUNC_VF_CFG) to decrease resource reservations
+* while some resources are still allocated. An error is returned
+* if the driver tries to set the reservation to be less than the
+* number of allocated resources.
+*/
+ #define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_RM_RSV_WHILE_ALLOC_CAP \
+ UINT32_C(0x1)
+ /*
+* When this bit is '1', the PF requires an L2 filter to be
+* allocated by the driver using HWRM_CFA_L2_FILTER_ALLOC after
+* bringing the interface up, before traffic is sent.
+*/
+ #define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_REQUIRE_L2_FILTER \
+ UINT32_C(0x2)
+ /*
+* When set to 1, indicates the field max_roce_vfs in the structure
+* is valid. If this bit is 0, the driver should not use the
+* 'max_roce_vfs' field.
+*/
+ #define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_MAX_ROCE_VFS_SUPPORTED \
+ UINT32_C(0x4)
+ /*
+* When set to 1, indicates the field 'rx_rate_profile_sel' in
+* RING_ALLOC can specify a valid RX rate profile when allocating
+* RX or RX aggregation rings. If this bit is 0, the driver
+* should not use the 'rx_rate_profile_sel' field.
+*/
+ #define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_RX_RATE_PROFILE_SEL_SUPPORTED
\
+ UINT32_C(0x8)
+ /*
+* The number of VFs that can be used for RoCE on the function. If less
+* than max_vfs, roce vfs will be assigned to the first VF of the
+* function and be contiguous.
+* This is valid only on the PF with SR-IOV and RDMA enabled.
+*/
+ uint16_tmax_roce_vfs;
+ uint8_t unused_3[5];
/*
* This field is used in Output records to indicate that the output
* is completely written to RAM. This field should be read as '1'
@@ -45026,6 +45065,14 @@ struct hwrm_ring_alloc_input {
*/
#define HWRM_RING_ALLOC_INPUT_ENABLES_STEERING_TAG_VALID \
UINT32_C(0x800)
+ /*
+* This bit must be '1' for the rx_rate_profile_sel field to
+* be configured. This should only be used when
+* 'rx_rate_profile_sel_supported' bit is set in flags_ext3
+* field of FUNC_QCAPS response.
+*/
+ #define HWRM_RING_ALLOC_INPUT_ENABLES_RX_RATE_PROFILE_VALID \
+ UINT32_C(0x1000)
/* Ring Type. */
uint8_t ring_type;
/* L2 Completion Ring (CR) */
@@ -45362,7 +45409,27 @@ struct hwrm_ring_alloc_input {
#define HWRM_RING_ALLOC_INPUT_MPC_CHNLS_TYPE_PRIMATE UINT32_C(0x4)
#define HWRM_RING_ALLOC_INPUT_MPC_CHNLS_TYPE_LAST \
HWRM_RING_ALLOC_INPUT_MPC_CHNLS_TYPE_PRIMATE
- uint8_t unused_4[2];
+ /* RX rate profile select */
+ uint8_t rx_rate_profile_sel;
+ /*
+* Indicate default RX rate profile when allocating
+* RX or RX aggregation rings. This should only be
+* used when 'rx_rate_profile_sel_supported' bit is
+* set in flags_ext3 field of FUNC_QCAPS response.
+*/
+ #define HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_DEFAULT \
+ UINT32_C(0x0)
+ /*
+* Indicate poll_mode RX rate profile when allocating
+* RX or RX aggregation rings. This should only be
+* used when 'rx_rate_profile_sel_supported' bit is
+* set in flags_ext3 field of FUNC_QCAPS response.
+*/
+ #define HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_POLL_MODE \
+ UINT32_C(0x1)
+ #define HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_LAST \
+ HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_POLL_MODE
+ uint8_t unused_4;
/*
* The cq_handle is specified when allocating a completion ring. For
* devices that support NQs, this cq_handle will be included in the
--
2.39.5 (Apple Git-154)
[PATCH v3 0/2] bnxt patchset
v1->v2: Rebasing patchset based on latest dpdk tree. v2->v3: Dropped an incomplete patch which ended up on mailing list. Please apply. Ajit Khaparde (2): net/bnxt: update HWRM API net/bnxt: add support for Rx profile selection drivers/net/bnxt/bnxt.h| 1 + drivers/net/bnxt/bnxt_hwrm.c | 15 +- drivers/net/bnxt/hsi_struct_def_dpdk.h | 71 +- 3 files changed, 84 insertions(+), 3 deletions(-) -- 2.39.5 (Apple Git-154)
[PATCH v2 0/2] Testpmd flow update/destroy fixes
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. v2: 1. Rephase commit messages. 2. Copy user_id to the flow list for flow_update command. 3. Enclose the case's body for flow_destroy command in braces. Danylo Vodopianov (2): app/testpmd: fix flow update app/testpmd: fix aged flow destroy app/test-pmd/config.c | 27 --- 1 file changed, 24 insertions(+), 3 deletions(-) -- 2.43.5
[PATCH v3 2/2] net/bnxt: add support for Rx profile selection
Some firmware versions can support the selection of Rx profile
during Rx and AGG ring allocation.
Check if the firmware sets the
HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_RX_RATE_PROFILE_SEL_SUPPORTED flag
and set the new Rx profile.
Signed-off-by: Ajit Khaparde
Reviewed-by: Andy Gospodarek
---
drivers/net/bnxt/bnxt.h | 1 +
drivers/net/bnxt/bnxt_hwrm.c | 15 ++-
2 files changed, 15 insertions(+), 1 deletion(-)
diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h
index 771349de6c..0402de3eb9 100644
--- a/drivers/net/bnxt/bnxt.h
+++ b/drivers/net/bnxt/bnxt.h
@@ -866,6 +866,7 @@ struct bnxt {
#define BNXT_FW_CAP_TX_COAL_CMPL BIT(10)
#define BNXT_FW_CAP_RX_ALL_PKT_TS BIT(11)
#define BNXT_FW_CAP_BACKING_STORE_V2 BIT(12)
+#define BNXT_FW_CAP_RX_RATE_PROFILEBIT(17)
#define BNXT_FW_BACKING_STORE_V2_EN(bp)\
((bp)->fw_cap & BNXT_FW_CAP_BACKING_STORE_V2)
#define BNXT_FW_BACKING_STORE_V1_EN(bp)\
diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c
index 351effb28f..d015ba2b9c 100644
--- a/drivers/net/bnxt/bnxt_hwrm.c
+++ b/drivers/net/bnxt/bnxt_hwrm.c
@@ -1139,8 +1139,8 @@ static int __bnxt_hwrm_func_qcaps(struct bnxt *bp)
int rc = 0;
struct hwrm_func_qcaps_input req = {.req_type = 0 };
struct hwrm_func_qcaps_output *resp = bp->hwrm_cmd_resp_addr;
+ uint32_t flags, flags_ext2, flags_ext3;
uint16_t new_max_vfs;
- uint32_t flags, flags_ext2;
HWRM_PREP(&req, HWRM_FUNC_QCAPS, BNXT_USE_CHIMP_MB);
@@ -1153,6 +1153,7 @@ static int __bnxt_hwrm_func_qcaps(struct bnxt *bp)
bp->max_ring_grps = rte_le_to_cpu_32(resp->max_hw_ring_grps);
flags = rte_le_to_cpu_32(resp->flags);
flags_ext2 = rte_le_to_cpu_32(resp->flags_ext2);
+ flags_ext3 = rte_le_to_cpu_32(resp->flags_ext3);
if (BNXT_PF(bp)) {
bp->pf->port_id = resp->port_id;
@@ -1259,6 +1260,8 @@ static int __bnxt_hwrm_func_qcaps(struct bnxt *bp)
bp->fw_cap |= BNXT_FW_CAP_RX_ALL_PKT_TS;
if (flags_ext2 & HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT2_UDP_GSO_SUPPORTED)
bp->fw_cap |= BNXT_FW_CAP_UDP_GSO;
+ if (flags_ext3 &
HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_RX_RATE_PROFILE_SEL_SUPPORTED)
+ bp->fw_cap |= BNXT_FW_CAP_RX_RATE_PROFILE;
unlock:
HWRM_UNLOCK();
@@ -2227,6 +2230,11 @@ int bnxt_hwrm_ring_alloc(struct bnxt *bp,
if (stats_ctx_id != INVALID_STATS_CTX_ID)
enables |=
HWRM_RING_ALLOC_INPUT_ENABLES_STAT_CTX_ID_VALID;
+ if (bp->fw_cap & BNXT_FW_CAP_RX_RATE_PROFILE) {
+ req.rx_rate_profile_sel =
+
HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_POLL_MODE;
+ enables |=
HWRM_RING_ALLOC_INPUT_ENABLES_RX_RATE_PROFILE_VALID;
+ }
break;
case HWRM_RING_ALLOC_INPUT_RING_TYPE_L2_CMPL:
req.ring_type = ring_type;
@@ -2257,6 +2265,11 @@ int bnxt_hwrm_ring_alloc(struct bnxt *bp,
enables |= HWRM_RING_ALLOC_INPUT_ENABLES_RX_RING_ID_VALID |
HWRM_RING_ALLOC_INPUT_ENABLES_RX_BUF_SIZE_VALID |
HWRM_RING_ALLOC_INPUT_ENABLES_STAT_CTX_ID_VALID;
+ if (bp->fw_cap & BNXT_FW_CAP_RX_RATE_PROFILE) {
+ req.rx_rate_profile_sel =
+
HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_POLL_MODE;
+ enables |=
HWRM_RING_ALLOC_INPUT_ENABLES_RX_RATE_PROFILE_VALID;
+ }
break;
default:
PMD_DRV_LOG_LINE(ERR, "hwrm alloc invalid ring type %d",
--
2.39.5 (Apple Git-154)
Re: [PATCH 13/16] crypto/dpaa_sec: fix bitmask truncation
On 15-11-2024 11:35, Stephen Hemminger wrote:
The dqrr_held mask is 64 bit but updates were getting truncated
because 1 is of type int (32 bit) and the result shift of int is of
type int (32 bit); therefore any value >= 32 would get truncated.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: fe3688ba7950 ("crypto/dpaa_sec: support event crypto adapter")
Cc: akhil.go...@nxp.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
drivers/crypto/dpaa_sec/dpaa_sec.c | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/crypto/dpaa_sec/dpaa_sec.c
b/drivers/crypto/dpaa_sec/dpaa_sec.c
index 3fa88ca968..e117cd77a6 100644
--- a/drivers/crypto/dpaa_sec/dpaa_sec.c
+++ b/drivers/crypto/dpaa_sec/dpaa_sec.c
@@ -1907,13 +1907,12 @@ dpaa_sec_enqueue_burst(void *qp, struct rte_crypto_op
**ops,
op = *(ops++);
if (*dpaa_seqn(op->sym->m_src) != 0) {
index = *dpaa_seqn(op->sym->m_src) - 1;
- if (DPAA_PER_LCORE_DQRR_HELD & (1 << index)) {
+ if (DPAA_PER_LCORE_DQRR_HELD & (UINT64_C(1) <<
index)) {
/* QM_EQCR_DCA_IDXMASK = 0x0f */
flags[loop] = ((index & 0x0f) << 8);
flags[loop] |= QMAN_ENQUEUE_FLAG_DCA;
DPAA_PER_LCORE_DQRR_SIZE--;
- DPAA_PER_LCORE_DQRR_HELD &=
- ~(1 << index);
+ DPAA_PER_LCORE_DQRR_HELD &= ~(UINT64_C(1)
<< index);
}
}
@@ -3500,7 +3499,7 @@ dpaa_sec_process_atomic_event(void *event,
/* Save active dqrr entries */
index = ((uintptr_t)dqrr >> 6) & (16/*QM_DQRR_SIZE*/ - 1);
DPAA_PER_LCORE_DQRR_SIZE++;
- DPAA_PER_LCORE_DQRR_HELD |= 1 << index;
+ DPAA_PER_LCORE_DQRR_HELD |= UINT64_C(1) << index;
DPAA_PER_LCORE_DQRR_MBUF(index) = ctx->op->sym->m_src;
ev->impl_opaque = index + 1;
*dpaa_seqn(ctx->op->sym->m_src) = (uint32_t)index + 1;
Acked-by: Hemant Agrawal
RE: [PATCH v2 2/2] app/testpmd: fix aged flow destroy
> -Original Message-
> From: Danylo Vodopianov
> Sent: Monday, November 18, 2024 12:26
> To: NBU-Contact-Thomas Monjalon (EXTERNAL) ;
> aman.deep.si...@intel.com; yuying.zh...@intel.com; Ori Kam
> ; mko-...@napatech.com; c...@napatech.com; Dariusz
> Sosnowski ; sil-...@napatech.com
> Cc: Gregory Etelson ; Alexander Kozyrev
> ; dev@dpdk.org; sta...@dpdk.org;
> ferruh.yi...@amd.com
> Subject: [PATCH v2 2/2] app/testpmd: fix aged flow destroy
>
> Avoid removal of additional flows after requested number of flows has been
> already removed.
>
> port_flow_destroy() function goes through all flows and compares given flow
> ‘id’
> with them. However in some cases it can advance pointer with “given ID” and
> thus remove additional flow.
Could you please remove the first two paragraphs in the commit message?
I don't think they are relevant.
>
> port_flow_destroy() function never assumed that rule array can be freed when
> it's executing, and port_flow_aged() just violated that assumption.
Also here, could you please describe what exactly is changed in port_flow_aged?
Current content explains the reason for the bug and the commit should describe
how it's fixed.
>
> Fixes: de956d5ecf08 ("app/testpmd: support age shared action context")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Danylo Vodopianov
> ---
> app/test-pmd/config.c | 6 --
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c index
> c831166431..04de2fe59d 100644
> --- a/app/test-pmd/config.c
> +++ b/app/test-pmd/config.c
> @@ -4160,8 +4160,9 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
> }
> type = (enum age_action_context_type *)contexts[idx];
> switch (*type) {
> - case ACTION_AGE_CONTEXT_TYPE_FLOW:
> + case ACTION_AGE_CONTEXT_TYPE_FLOW: {
> ctx.pf = container_of(type, struct port_flow,
> age_type);
> + uint64_t flow_id = ctx.pf->id;
Could you please refactor these lines as follows?
```
case ACTION_AGE_CONTEXT_TYPE_FLOW: {
uint64_t flow_id;
ctx.pf = container_of(type, struct port_flow, age_type);
flow_id = ctx.pf->id;
```
This is to make the style aligned - variables declared at the beginning of the
block.
> printf("%-20s\t%" PRIu64 "\t%" PRIu32 "\t%" PRIu32
>
> "\t%c%c%c\t\n",
>"Flow",
> @@ -4172,9 +4173,10 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
>ctx.pf->rule.attr->egress ? 'e' : '-',
>ctx.pf->rule.attr->transfer ? 't' : '-');
> if (destroy && !port_flow_destroy(port_id, 1,
> - &ctx.pf->id, false))
> + &flow_id,
> + false))
> total++;
> break;
> + }
> case ACTION_AGE_CONTEXT_TYPE_INDIRECT_ACTION:
> ctx.pia = container_of(type,
> struct port_indirect_action,
> age_type);
> --
> 2.43.5
[PATCH] vhost: fix read vs write lock mismatch
If lock is acquired for write, it must be released for write
or a deadlock is likely.
Bugzilla ID: 1582
Fixes: 9fc93a1e2320 ("vhost: fix virtqueue access check in datapath")
Cc: david.march...@redhat.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
lib/vhost/virtio_net.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
index 298a5dae74..d764d4bc6a 100644
--- a/lib/vhost/virtio_net.c
+++ b/lib/vhost/virtio_net.c
@@ -2538,7 +2538,7 @@ virtio_dev_rx_async_submit(struct virtio_net *dev, struct
vhost_virtqueue *vq,
if (unlikely(!vq->access_ok)) {
vhost_user_iotlb_rd_unlock(vq);
- rte_rwlock_read_unlock(&vq->access_lock);
+ rte_rwlock_write_unlock(&vq->access_lock);
virtio_dev_vring_translate(dev, vq);
goto out_no_unlock;
--
2.45.2
[PATCH v2 15/19] net/dpaa: fix bitmask truncation
The dqrr_held mask is 64 bit but updates were getting truncated
because 1 is of type int (32 bit) and the result shift of int is of
type int (32 bit); therefore any value >= 32 would get truncated.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: 5e7455931442 ("net/dpaa: support Rx queue configurations with eventdev")
Cc: sunil.k...@nxp.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
Acked-by: Hemant Agrawal
---
drivers/net/dpaa/dpaa_rxtx.c | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/net/dpaa/dpaa_rxtx.c b/drivers/net/dpaa/dpaa_rxtx.c
index 247e7b92ba..05bd73becf 100644
--- a/drivers/net/dpaa/dpaa_rxtx.c
+++ b/drivers/net/dpaa/dpaa_rxtx.c
@@ -842,7 +842,7 @@ dpaa_rx_cb_atomic(void *event,
/* Save active dqrr entries */
index = DQRR_PTR2IDX(dqrr);
DPAA_PER_LCORE_DQRR_SIZE++;
- DPAA_PER_LCORE_DQRR_HELD |= 1 << index;
+ DPAA_PER_LCORE_DQRR_HELD |= UINT64_C(1) << index;
DPAA_PER_LCORE_DQRR_MBUF(index) = mbuf;
ev->impl_opaque = index + 1;
*dpaa_seqn(mbuf) = (uint32_t)index + 1;
@@ -1338,13 +1338,12 @@ dpaa_eth_queue_tx(void *q, struct rte_mbuf **bufs,
uint16_t nb_bufs)
seqn = *dpaa_seqn(mbuf);
if (seqn != DPAA_INVALID_MBUF_SEQN) {
index = seqn - 1;
- if (DPAA_PER_LCORE_DQRR_HELD & (1 << index)) {
+ if (DPAA_PER_LCORE_DQRR_HELD & (UINT64_C(1) <<
index)) {
flags[loop] =
((index & QM_EQCR_DCA_IDXMASK) << 8);
flags[loop] |= QMAN_ENQUEUE_FLAG_DCA;
DPAA_PER_LCORE_DQRR_SIZE--;
- DPAA_PER_LCORE_DQRR_HELD &=
- ~(1 << index);
+ DPAA_PER_LCORE_DQRR_HELD &=
~(UINT64_C(1) << index);
}
}
--
2.45.2
[PATCH v2 12/19] crypto/dpaa2_sec: fix bitmask truncation
The dqrr_held mask is 64 bit but updates were getting truncated
because 1 is of type int (32 bit) and the result shift of int is of
type int (32 bit); therefore any value >= 32 would get truncated.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: a77db24643b7 ("crypto/dpaa2_sec: support atomic queues")
Cc: ashish.j...@nxp.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
Acked-by: Hemant Agrawal
---
drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
b/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
index ec6577f64c..7ad8fd47dd 100644
--- a/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
+++ b/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
@@ -1491,8 +1491,8 @@ dpaa2_sec_enqueue_burst(void *qp, struct rte_crypto_op
**ops,
if (*dpaa2_seqn((*ops)->sym->m_src)) {
if (*dpaa2_seqn((*ops)->sym->m_src) &
QBMAN_ENQUEUE_FLAG_DCA) {
DPAA2_PER_LCORE_DQRR_SIZE--;
- DPAA2_PER_LCORE_DQRR_HELD &= ~(1 <<
- *dpaa2_seqn((*ops)->sym->m_src) &
+ DPAA2_PER_LCORE_DQRR_HELD &=
~(UINT64_C(1) <<
+ *dpaa2_seqn((*ops)->sym->m_src)
&
QBMAN_EQCR_DCA_IDXMASK);
}
flags[loop] = *dpaa2_seqn((*ops)->sym->m_src);
@@ -1772,7 +1772,7 @@ dpaa2_sec_set_enqueue_descriptor(struct dpaa2_queue
*dpaa2_q,
dq_idx = *dpaa2_seqn(m) - 1;
qbman_eq_desc_set_dca(eqdesc, 1, dq_idx, 0);
DPAA2_PER_LCORE_DQRR_SIZE--;
- DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dq_idx);
+ DPAA2_PER_LCORE_DQRR_HELD &= ~(UINT64_C(1) << dq_idx);
}
*dpaa2_seqn(m) = DPAA2_INVALID_MBUF_SEQN;
}
@@ -4055,7 +4055,7 @@ dpaa2_sec_process_atomic_event(struct qbman_swp *swp
__rte_unused,
dqrr_index = qbman_get_dqrr_idx(dq);
*dpaa2_seqn(crypto_op->sym->m_src) = QBMAN_ENQUEUE_FLAG_DCA |
dqrr_index;
DPAA2_PER_LCORE_DQRR_SIZE++;
- DPAA2_PER_LCORE_DQRR_HELD |= 1 << dqrr_index;
+ DPAA2_PER_LCORE_DQRR_HELD |= UINT64_C(1) << dqrr_index;
DPAA2_PER_LCORE_DQRR_MBUF(dqrr_index) = crypto_op->sym->m_src;
ev->event_ptr = crypto_op;
}
--
2.45.2
[PATCH v2 02/19] net/cpfl: avoid calling log (printf) with null
The log message would always refer to str variable which
is NULL here. Looks like author intended to print original
parameter.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Signed-off-by: Stephen Hemminger
---
drivers/net/cpfl/cpfl_ethdev.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/cpfl/cpfl_ethdev.c b/drivers/net/cpfl/cpfl_ethdev.c
index 6f6707a0bd..1817221652 100644
--- a/drivers/net/cpfl/cpfl_ethdev.c
+++ b/drivers/net/cpfl/cpfl_ethdev.c
@@ -1580,7 +1580,7 @@ parse_repr(const char *key __rte_unused, const char
*value, void *args)
RTE_DIM(eth_da->representor_ports));
done:
if (str == NULL) {
- PMD_DRV_LOG(ERR, "wrong representor format: %s", str);
+ PMD_DRV_LOG(ERR, "wrong representor format: %s", value);
return -1;
}
--
2.45.2
[PATCH v2 00/19] minor fixes from PVS studio bug list
More bug fixes from PVS studio bug reports. And one other fix to ptpclient. Stephen Hemminger (19): common/cnxk: remove duplicate condition net/cpfl: avoid calling log (printf) with null raw/cnxk_gpio: fix file descriptor leak net/ntnic: remove dead code net/i40e: remove duplicate code eal: fix out of bounds access in devargs net/qede: fix missing debug string examples/ptpclient: replace rte_memcpy with assignment examples/ptpclient: fix self memcmp net/octeon_ep: remove duplicate code net/hinic: fix flow type bitmask overflow crypto/dpaa2_sec: fix bitmask truncation crypto/dpaa_sec: fix bitmask truncation event/dpaa: fix bitmask truncation net/dpaa: fix bitmask truncation net/dpaa2: fix bitmask truncation net/qede: don't use same loop variable twice examples/l3fwd: fix operator precedence bugs common/cnxk: fix null ptr check v2 - add a few more fixes, and rebase drivers/common/cnxk/cnxk_security.c | 16 -- drivers/common/cnxk/roc_bphy_cgx.c | 12 +-- drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c | 8 +++ drivers/crypto/dpaa_sec/dpaa_sec.c | 7 +++--- drivers/event/dpaa/dpaa_eventdev.c | 10 - drivers/net/cpfl/cpfl_ethdev.c | 2 +- drivers/net/dpaa/dpaa_rxtx.c| 7 +++--- drivers/net/dpaa2/dpaa2_rxtx.c | 6 +++--- drivers/net/hinic/hinic_pmd_flow.c | 14 ++-- drivers/net/i40e/i40e_fdir.c| 10 - drivers/net/ntnic/ntnic_ethdev.c| 8 --- drivers/net/octeon_ep/otx_ep_ethdev.c | 9 ++-- drivers/net/qede/base/ecore_dcbx.c | 8 +++ drivers/net/qede/qede_debug.c | 5 + drivers/raw/cnxk_gpio/cnxk_gpio_selftest.c | 24 + examples/l3fwd-power/main.c | 4 ++-- examples/l3fwd/main.c | 5 +++-- examples/ptpclient/ptpclient.c | 10 +++-- lib/eal/common/eal_common_devargs.c | 2 +- 19 files changed, 81 insertions(+), 86 deletions(-) -- 2.45.2
Re: [PATCH v2 1/6] dts: Remove build target config and list of devices
Hi Nick, I second the comments already made. A couple of nits on top of those: On 05/07/2024 18:13, Nicholas Pratte wrote: Remove the list of devices from the schema, as these are unuesed. Likewise, removed build-target information since these is not currently used, and it is unlikely to be used in the future. Adjustments to the dts.rst are made to reflect these changes. There are no adjustments made to the dts.rst here, so it shouldn't be mentioned. Bugzilla ID: 1360 There should be an empty line between Signed-off-by and the metatags. Signed-off-by: Nicholas Pratte --- @unique class Compiler(StrEnum): r"""The supported compilers of :class:`~framework.testbed_model.node.Node`\s.""" @@ -341,28 +325,20 @@ class BuildTargetConfiguration: The configuration used for building DPDK. Attributes: -arch: The target architecture to build for. -os: The target os to build for. -cpu: The target CPU to build for. compiler: The compiler executable to use. compiler_wrapper: This string will be put in front of the compiler when executing the build. Useful for adding wrapper commands, such as ``ccache``. name: The name of the compiler. I am a bit confused whether you are removing `name` or not, but if you are you forgot to remove this entry from the docstring. """ -arch: Architecture -os: OS -cpu: CPUType compiler: Compiler compiler_wrapper: str -name: str
RE: [PATCH 2/3] regex/cn9k: fix build warnings on ubuntu 24.04
> -Original Message-
> From: sk...@marvell.com
> Sent: Thursday, November 14, 2024 1:08 PM
> To: Liron Himi ; Jerin Jacob
> Cc: dev@dpdk.org; alia...@nvidia.com; Sunil Kumar Kori
> ; sta...@dpdk.org
> Subject: [PATCH 2/3] regex/cn9k: fix build warnings on ubuntu 24.04
>
> From: Sunil Kumar Kori
>
> Bugzilla ID: 1513
> Fixes: 72c00ae9dba7 ("regex/cn9k: use cnxk infrastructure")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Sunil Kumar Kori
> ---
> drivers/regex/cn9k/cn9k_regexdev.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/regex/cn9k/cn9k_regexdev.c
> b/drivers/regex/cn9k/cn9k_regexdev.c
> index aa809ab5bf..e635980183 100644
> --- a/drivers/regex/cn9k/cn9k_regexdev.c
> +++ b/drivers/regex/cn9k/cn9k_regexdev.c
> @@ -271,7 +271,7 @@ static inline int
> ree_enqueue(struct roc_ree_qp *qp, struct rte_regex_ops *op,
>struct roc_ree_pending_queue *pend_q) {
> - union roc_ree_inst inst;
> + union roc_ree_inst inst = {0};
This changes are in fastpath and it is costly. Please fix with other means or
suppress the warning if it not relevant.
I skipped this patch for the merged and 1/3, 3/3 patches are pplied to
dpdk-next-net-mrvl/for-main. Thanks
> union ree_res *res;
> uint32_t offset;
>
> --
> 2.43.0
Re: [PATCH v2 2/6] dts: Use First Core Logic Change
Same comments as Juraj again. Also the commit subject should be all lower case, and the bugzilla tag should be split from the signed-off-by here too.
[PATCH v2 1/2] app/testpmd: fix flow update
If actions provided to “flow update…“ command contained an age
action, then testpmd did not update the age action context
accordingly.
Thus "flow aged destroy" command can not
execute successfully.
Fix was done with next steps
1. Generate new port flow entry to add/replace action(s).
2. Set age context if age action is present.
3. Replace flow in the flow list.
Fixes: 2d9c7e56e52c ("app/testpmd: support updating flow rule actions")
Cc: sta...@dpdk.org
Signed-off-by: Danylo Vodopianov
---
app/test-pmd/config.c | 21 -
1 file changed, 20 insertions(+), 1 deletion(-)
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index 88770b4dfc..c831166431 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -3882,7 +3882,8 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
const struct rte_flow_action *actions, bool is_user_id)
{
struct rte_port *port;
- struct port_flow **flow_list;
+ struct port_flow **flow_list, *uf;
+ struct rte_flow_action_age *age = age_action_get(actions);
if (port_id_is_invalid(port_id, ENABLED_WARN) ||
port_id == (portid_t)RTE_PORT_ALL)
@@ -3897,6 +3898,16 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
flow_list = &flow->next;
continue;
}
+
+ /* Update flow action(s) with new action(s) */
+ uf = port_flow_new(flow->rule.attr_ro, flow->rule.pattern_ro,
actions, &error);
+ if (!uf)
+ return port_flow_complain(&error);
+ if (age) {
+ flow->age_type = ACTION_AGE_CONTEXT_TYPE_FLOW;
+ age->context = &flow->age_type;
+ }
+
/*
* Poisoning to make sure PMDs update it in case
* of error.
@@ -3913,6 +3924,14 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
printf("Flow rule #%"PRIu64
" updated with new actions\n",
flow->id);
+
+ uf->next = flow->next;
+ uf->id = flow->id;
+ uf->user_id = flow->user_id;
+ uf->flow = flow->flow;
+ *flow_list = uf;
+
+ free(flow);
return 0;
}
printf("Failed to find flow %"PRIu32"\n", rule_id);
--
2.43.5
[PATCH v2 2/2] app/testpmd: fix aged flow destroy
Avoid removal of additional flows after requested number of flows has
been already removed.
port_flow_destroy() function goes through all flows and compares
given flow ‘id’ with them. However in some cases it can advance pointer
with “given ID” and thus remove additional flow.
port_flow_destroy() function never assumed that rule array can be freed
when it's executing, and port_flow_aged() just violated that assumption.
Fixes: de956d5ecf08 ("app/testpmd: support age shared action context")
Cc: sta...@dpdk.org
Signed-off-by: Danylo Vodopianov
---
app/test-pmd/config.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index c831166431..04de2fe59d 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -4160,8 +4160,9 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
}
type = (enum age_action_context_type *)contexts[idx];
switch (*type) {
- case ACTION_AGE_CONTEXT_TYPE_FLOW:
+ case ACTION_AGE_CONTEXT_TYPE_FLOW: {
ctx.pf = container_of(type, struct port_flow, age_type);
+ uint64_t flow_id = ctx.pf->id;
printf("%-20s\t%" PRIu64 "\t%" PRIu32 "\t%" PRIu32
"\t%c%c%c\t\n",
"Flow",
@@ -4172,9 +4173,10 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
ctx.pf->rule.attr->egress ? 'e' : '-',
ctx.pf->rule.attr->transfer ? 't' : '-');
if (destroy && !port_flow_destroy(port_id, 1,
- &ctx.pf->id, false))
+ &flow_id, false))
total++;
break;
+ }
case ACTION_AGE_CONTEXT_TYPE_INDIRECT_ACTION:
ctx.pia = container_of(type,
struct port_indirect_action, age_type);
--
2.43.5
[PATCH v3 1/2] app/testpmd: fix flow update
If actions provided to “flow update…“ command contained an age
action, then testpmd did not update the age action context
accordingly.
Thus "flow aged destroy" command can not
execute successfully.
Fix was done with next steps
1. Generate new port flow entry to add/replace action(s).
2. Set age context if age action is present.
3. Replace flow in the flow list.
Fixes: 2d9c7e56e52c ("app/testpmd: support updating flow rule actions")
Cc: sta...@dpdk.org
Signed-off-by: Danylo Vodopianov
---
app/test-pmd/config.c | 21 -
1 file changed, 20 insertions(+), 1 deletion(-)
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index 88770b4dfc..c831166431 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -3882,7 +3882,8 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
const struct rte_flow_action *actions, bool is_user_id)
{
struct rte_port *port;
- struct port_flow **flow_list;
+ struct port_flow **flow_list, *uf;
+ struct rte_flow_action_age *age = age_action_get(actions);
if (port_id_is_invalid(port_id, ENABLED_WARN) ||
port_id == (portid_t)RTE_PORT_ALL)
@@ -3897,6 +3898,16 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
flow_list = &flow->next;
continue;
}
+
+ /* Update flow action(s) with new action(s) */
+ uf = port_flow_new(flow->rule.attr_ro, flow->rule.pattern_ro,
actions, &error);
+ if (!uf)
+ return port_flow_complain(&error);
+ if (age) {
+ flow->age_type = ACTION_AGE_CONTEXT_TYPE_FLOW;
+ age->context = &flow->age_type;
+ }
+
/*
* Poisoning to make sure PMDs update it in case
* of error.
@@ -3913,6 +3924,14 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
printf("Flow rule #%"PRIu64
" updated with new actions\n",
flow->id);
+
+ uf->next = flow->next;
+ uf->id = flow->id;
+ uf->user_id = flow->user_id;
+ uf->flow = flow->flow;
+ *flow_list = uf;
+
+ free(flow);
return 0;
}
printf("Failed to find flow %"PRIu32"\n", rule_id);
--
2.43.5
[PATCH v2 0/2] Testpmd flow update/destroy fixes
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. v2: 1. Rephase commit messages. 2. Copy user_id to the flow list for flow_update command. 3. Enclose the case's body for flow_destroy command in braces. Danylo Vodopianov (2): app/testpmd: fix flow update app/testpmd: fix aged flow destroy app/test-pmd/config.c | 27 --- 1 file changed, 24 insertions(+), 3 deletions(-) -- 2.43.5
Re: [PATCH v2 3/3] bus/dpaa: enabling clang thread safety for locks
14/11/2024 08:44, Hemant Agrawal: > To avoid regressions on this topic, could you look into enabling clang > thread safety check? This original question is replaced with a proper description of the change: bus/dpaa: enable clang thread safety check for FQ locks Enable "annotate_locks" for compile-time checks by clang. FQLOCK and FQUNLOCK need to be considered as lock functions that the clang analyzer can rely on. > > Signed-off-by: David Marchand > Signed-off-by: Hemant Agrawal Applied, thanks.
[PATCH v2 0/2] Testpmd flow update/destroy fixes
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. v2: 1. Rephase commit messages. 2. Copy user_id to the flow list for flow_update command. 3. Enclose the case's body for flow_destroy command in braces. Danylo Vodopianov (2): app/testpmd: fix flow update app/testpmd: fix aged flow destroy app/test-pmd/config.c | 27 --- 1 file changed, 24 insertions(+), 3 deletions(-) -- 2.43.5
[PATCH 0/3] bnxt patchset
bnxt patchset rebased against the latest dpdk repository. Ajit Khaparde (3): net/bnxt: update HWRM API net/bnxt: add support for Rx profile selection net/bnxt: set max VFs count for P7 drivers/net/bnxt/bnxt.h| 2 + drivers/net/bnxt/bnxt_hwrm.c | 15 +- drivers/net/bnxt/bnxt_vnic.c | 6 +++ drivers/net/bnxt/hsi_struct_def_dpdk.h | 71 +- 4 files changed, 91 insertions(+), 3 deletions(-) -- 2.39.5 (Apple Git-154)
[PATCH v3 2/2] app/testpmd: fix aged flow destroy
port_flow_destroy() function never assumed that rule array can be freed
when it's executing, and port_flow_aged() just violated that assumption.
In case of flow async create failure, it tries to do a cleanup, but it
wrongly removes a 1st flow (with id 0). pf->id is not set at this
moment and it always is 0, thus 1st flow is removed. A local copy of
flow->id must be used to call of port_flow_destroy() to avoid access
and processing of flow->id after the flow is removed.
Fixes: de956d5ecf08 ("app/testpmd: support age shared action context")
Cc: sta...@dpdk.org
Signed-off-by: Danylo Vodopianov
---
app/test-pmd/config.c | 7 +--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index c831166431..28d45568ac 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -4160,8 +4160,10 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
}
type = (enum age_action_context_type *)contexts[idx];
switch (*type) {
- case ACTION_AGE_CONTEXT_TYPE_FLOW:
+ case ACTION_AGE_CONTEXT_TYPE_FLOW: {
+ uint64_t flow_id;
ctx.pf = container_of(type, struct port_flow, age_type);
+ flow_id = ctx.pf->id;
printf("%-20s\t%" PRIu64 "\t%" PRIu32 "\t%" PRIu32
"\t%c%c%c\t\n",
"Flow",
@@ -4172,9 +4174,10 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
ctx.pf->rule.attr->egress ? 'e' : '-',
ctx.pf->rule.attr->transfer ? 't' : '-');
if (destroy && !port_flow_destroy(port_id, 1,
- &ctx.pf->id, false))
+ &flow_id, false))
total++;
break;
+ }
case ACTION_AGE_CONTEXT_TYPE_INDIRECT_ACTION:
ctx.pia = container_of(type,
struct port_indirect_action, age_type);
--
2.43.5
[PATCH v3 0/2] Testpmd flow update/destroy fixes
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. v2: 1. Rephase commit messages. 2. Copy user_id to the flow list for flow_update command. 3. Enclose the case's body for flow_destroy command in braces. v3: 1. Update commit message 2. Aligned variable declaration code style Danylo Vodopianov (2): app/testpmd: fix flow update app/testpmd: fix aged flow destroy app/test-pmd/config.c | 28 +--- 1 file changed, 25 insertions(+), 3 deletions(-) -- 2.43.5
[PATCH v2 06/19] eal: fix out of bounds access in devargs
The code for parsing layers in devargs could reference past
the end of layers[] array on stack.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: 9a1a9e4a2ddd ("devargs: support path value with global device syntax")
Cc: xuemi...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
lib/eal/common/eal_common_devargs.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/eal/common/eal_common_devargs.c
b/lib/eal/common/eal_common_devargs.c
index a64805b268..dd857fc839 100644
--- a/lib/eal/common/eal_common_devargs.c
+++ b/lib/eal/common/eal_common_devargs.c
@@ -88,7 +88,7 @@ rte_devargs_layers_parse(struct rte_devargs *devargs,
s = devargs->data;
while (s != NULL) {
- if (nblayer > RTE_DIM(layers)) {
+ if (nblayer >= RTE_DIM(layers)) {
ret = -E2BIG;
goto get_out;
}
--
2.45.2
[PATCH v2 07/19] net/qede: fix missing debug string
The array of debug status strings did not match possible enum
values. Add the missing element and a static assert to make sure
the table has all possible values.
For more complete description see.
Link: https://pvs-studio.com/en/blog/posts/cpp/1176/
Fixes: ec55c118792b ("net/qede: add infrastructure for debug data collection")
Cc: rm...@marvell.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
drivers/net/qede/qede_debug.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/drivers/net/qede/qede_debug.c b/drivers/net/qede/qede_debug.c
index 18f2d988fb..1d3147b724 100644
--- a/drivers/net/qede/qede_debug.c
+++ b/drivers/net/qede/qede_debug.c
@@ -4,6 +4,7 @@
* www.marvell.com
*/
+#include
#include
#include "base/bcm_osal.h"
#include "base/ecore.h"
@@ -82,6 +83,7 @@ static const char * const s_mem_group_names[] = {
"TM_MEM",
"TASK_CFC_MEM",
};
+static_assert(RTE_DIM(s_mem_group_names) == MEM_GROUPS_NUM, "memory group
string mismatch");
/* Idle check conditions */
@@ -5614,6 +5616,8 @@ static const char * const s_status_str[] = {
/* DBG_STATUS_INVALID_FILTER_TRIGGER_DWORDS */
"The filter/trigger constraint dword offsets are not enabled for
recording",
+ /* DBG_STATUS_NO_MATCHING_FRAMING_MODE */
+ "No matching frame mode",
/* DBG_STATUS_VFC_READ_ERROR */
"Error reading from VFC",
@@ -5759,6 +5763,7 @@ static const char * const s_status_str[] = {
/* DBG_STATUS_MISSING_TRIGGER_STATE_STORM */
"When triggering on Storm data, the Storm to trigger on must be
specified"
};
+static_assert(RTE_DIM(s_status_str) == MAX_DBG_STATUS, "status string table
mismatch");
/* Idle check severity names array */
static const char * const s_idle_chk_severity_str[] = {
--
2.45.2
[PATCH v2 05/19] net/i40e: remove duplicate code
There are two branches in the cascading if/else that have same
condition and code; remove one. Update the code to follow DPDK
style where all statements in if should have brackets if any
leg requires them.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: 2ab5c84605f0 ("net/i40e: fix ESP flow creation")
Signed-off-by: Stephen Hemminger
Acked-by: Bruce Richardson
---
drivers/net/i40e/i40e_fdir.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/net/i40e/i40e_fdir.c b/drivers/net/i40e/i40e_fdir.c
index 47f79ecf11..6861bea99a 100644
--- a/drivers/net/i40e/i40e_fdir.c
+++ b/drivers/net/i40e/i40e_fdir.c
@@ -599,18 +599,16 @@ i40e_flow_fdir_fill_eth_ip_head(struct i40e_pf *pf,
} else if (cus_pctype->index == I40E_CUSTOMIZED_ESP_IPV4_UDP) {
len = fill_ip4_head(fdir_input, raw_pkt, IPPROTO_UDP,
len, ether_type);
- } else if (cus_pctype->index == I40E_CUSTOMIZED_ESP_IPV4_UDP) {
- len = fill_ip4_head(fdir_input, raw_pkt, IPPROTO_UDP,
- len, ether_type);
- } else if (cus_pctype->index == I40E_CUSTOMIZED_ESP_IPV6)
+ } else if (cus_pctype->index == I40E_CUSTOMIZED_ESP_IPV6) {
len = fill_ip6_head(fdir_input, raw_pkt, IPPROTO_ESP,
len, ether_type);
- else if (cus_pctype->index == I40E_CUSTOMIZED_ESP_IPV6_UDP)
+ } else if (cus_pctype->index == I40E_CUSTOMIZED_ESP_IPV6_UDP) {
len = fill_ip6_head(fdir_input, raw_pkt, IPPROTO_UDP,
len, ether_type);
- else if (cus_pctype->index == I40E_CUSTOMIZED_IPV6_L2TPV3)
+ } else if (cus_pctype->index == I40E_CUSTOMIZED_IPV6_L2TPV3) {
len = fill_ip6_head(fdir_input, raw_pkt, IPPROTO_L2TP,
len, ether_type);
+ }
} else {
PMD_DRV_LOG(ERR, "unknown pctype %u.", fdir_input->pctype);
return -1;
--
2.45.2
[PATCH v2 08/19] examples/ptpclient: replace rte_memcpy with assignment
Don't use rte_memcpy() when not necessary. Structure assignment
is as fast and type safe.
Signed-off-by: Stephen Hemminger
---
examples/ptpclient/ptpclient.c | 8 ++--
1 file changed, 2 insertions(+), 6 deletions(-)
diff --git a/examples/ptpclient/ptpclient.c b/examples/ptpclient/ptpclient.c
index 23fa487081..2ec532d058 100644
--- a/examples/ptpclient/ptpclient.c
+++ b/examples/ptpclient/ptpclient.c
@@ -415,9 +415,7 @@ parse_sync(struct ptpv2_time_receiver_ordinary *ptp_data,
uint16_t rx_tstamp_idx
ptp_data->seqID_SYNC = rte_be_to_cpu_16(ptp_hdr->seq_id);
if (ptp_data->ptpset == 0) {
- rte_memcpy(&ptp_data->transmitter_clock_id,
- &ptp_hdr->source_port_id.clock_id,
- sizeof(struct clock_id));
+ ptp_data->transmitter_clock_id =
ptp_hdr->source_port_id.clock_id;
ptp_data->ptpset = 1;
}
@@ -522,9 +520,7 @@ parse_fup(struct ptpv2_time_receiver_ordinary *ptp_data)
client_clkid->id[6] = eth_hdr->src_addr.addr_bytes[4];
client_clkid->id[7] = eth_hdr->src_addr.addr_bytes[5];
- rte_memcpy(&ptp_data->client_clock_id,
- client_clkid,
- sizeof(struct clock_id));
+ ptp_data->client_clock_id = *client_clkid;
/* Enable flag for hardware timestamping. */
created_pkt->ol_flags |= RTE_MBUF_F_TX_IEEE1588_TMST;
--
2.45.2
[PATCH v2 03/19] raw/cnxk_gpio: fix file descriptor leak
The function would leak file if fscanf failed.
There is a working version in other file, clone that.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: 0e6557b448fa ("raw/cnxk_gpio: add self test")
Cc: tduszyn...@marvell.com
Signed-off-by: Stephen Hemminger
---
drivers/raw/cnxk_gpio/cnxk_gpio_selftest.c | 24 ++
1 file changed, 15 insertions(+), 9 deletions(-)
diff --git a/drivers/raw/cnxk_gpio/cnxk_gpio_selftest.c
b/drivers/raw/cnxk_gpio/cnxk_gpio_selftest.c
index 2f3973a7b5..a0d9942f20 100644
--- a/drivers/raw/cnxk_gpio/cnxk_gpio_selftest.c
+++ b/drivers/raw/cnxk_gpio/cnxk_gpio_selftest.c
@@ -34,24 +34,30 @@ cnxk_gpio_attr_exists(const char *attr)
static int
cnxk_gpio_read_attr(char *attr, char *val)
{
+ int ret, ret2;
FILE *fp;
- int ret;
fp = fopen(attr, "r");
if (!fp)
return -errno;
ret = fscanf(fp, "%s", val);
- if (ret < 0)
- return -errno;
- if (ret != 1)
- return -EIO;
+ if (ret < 0) {
+ ret = -errno;
+ goto out;
+ }
+ if (ret != 1) {
+ ret = -EIO;
+ goto out;
+ }
- ret = fclose(fp);
- if (ret)
- return -errno;
+ ret = 0;
+out:
+ ret2 = fclose(fp);
+ if (!ret)
+ ret = ret2;
- return 0;
+ return ret;
}
#define CNXK_GPIO_ERR_STR(err, str, ...) do {
\
--
2.45.2
[PATCH v2 10/19] net/octeon_ep: remove duplicate code
Both sides of the if in uninit are using same code.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Signed-off-by: Stephen Hemminger
---
drivers/net/octeon_ep/otx_ep_ethdev.c | 9 ++---
1 file changed, 2 insertions(+), 7 deletions(-)
diff --git a/drivers/net/octeon_ep/otx_ep_ethdev.c
b/drivers/net/octeon_ep/otx_ep_ethdev.c
index b4f8baf3b3..8b14734b0c 100644
--- a/drivers/net/octeon_ep/otx_ep_ethdev.c
+++ b/drivers/net/octeon_ep/otx_ep_ethdev.c
@@ -721,14 +721,9 @@ static const struct eth_dev_ops otx_ep_eth_dev_ops = {
static int
otx_ep_eth_dev_uninit(struct rte_eth_dev *eth_dev)
{
- if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
- eth_dev->dev_ops = NULL;
- eth_dev->rx_pkt_burst = NULL;
- eth_dev->tx_pkt_burst = NULL;
- return 0;
- }
+ if (rte_eal_process_type() == RTE_PROC_PRIMARY)
+ otx_ep_mbox_uninit(eth_dev);
- otx_ep_mbox_uninit(eth_dev);
eth_dev->dev_ops = NULL;
eth_dev->rx_pkt_burst = NULL;
eth_dev->tx_pkt_burst = NULL;
--
2.45.2
[PATCH v2 14/19] event/dpaa: fix bitmask truncation
More bitmask truncation from mask computation.
Fixes: 0ee17f79ebd0 ("event/dpaa: add enqueue/dequeue")
Cc: sunil.k...@nxp.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
Acked-by: Hemant Agrawal
---
drivers/event/dpaa/dpaa_eventdev.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/event/dpaa/dpaa_eventdev.c
b/drivers/event/dpaa/dpaa_eventdev.c
index 853cc1ecf9..400e0ecd1c 100644
--- a/drivers/event/dpaa/dpaa_eventdev.c
+++ b/drivers/event/dpaa/dpaa_eventdev.c
@@ -102,7 +102,7 @@ dpaa_event_enqueue_burst(void *port, const struct rte_event
ev[],
qman_dca_index(ev[i].impl_opaque, 0);
mbuf = DPAA_PER_LCORE_DQRR_MBUF(i);
*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
- DPAA_PER_LCORE_DQRR_HELD &= ~(1 << i);
+ DPAA_PER_LCORE_DQRR_HELD &= ~(UINT64_C(1) << i);
DPAA_PER_LCORE_DQRR_SIZE--;
break;
default:
@@ -199,11 +199,11 @@ dpaa_event_dequeue_burst(void *port, struct rte_event
ev[],
/* Check if there are atomic contexts to be released */
i = 0;
while (DPAA_PER_LCORE_DQRR_SIZE) {
- if (DPAA_PER_LCORE_DQRR_HELD & (1 << i)) {
+ if (DPAA_PER_LCORE_DQRR_HELD & (UINT64_C(1) << i)) {
qman_dca_index(i, 0);
mbuf = DPAA_PER_LCORE_DQRR_MBUF(i);
*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
- DPAA_PER_LCORE_DQRR_HELD &= ~(1 << i);
+ DPAA_PER_LCORE_DQRR_HELD &= ~(UINT64_C(1) << i);
DPAA_PER_LCORE_DQRR_SIZE--;
}
i++;
@@ -263,11 +263,11 @@ dpaa_event_dequeue_burst_intr(void *port, struct
rte_event ev[],
/* Check if there are atomic contexts to be released */
i = 0;
while (DPAA_PER_LCORE_DQRR_SIZE) {
- if (DPAA_PER_LCORE_DQRR_HELD & (1 << i)) {
+ if (DPAA_PER_LCORE_DQRR_HELD & (UINT64_C(1) << i)) {
qman_dca_index(i, 0);
mbuf = DPAA_PER_LCORE_DQRR_MBUF(i);
*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
- DPAA_PER_LCORE_DQRR_HELD &= ~(1 << i);
+ DPAA_PER_LCORE_DQRR_HELD &= ~(UINT64_C(1) << i);
DPAA_PER_LCORE_DQRR_SIZE--;
}
i++;
--
2.45.2
[PATCH v2 13/19] crypto/dpaa_sec: fix bitmask truncation
The dqrr_held mask is 64 bit but updates were getting truncated
because 1 is of type int (32 bit) and the result shift of int is of
type int (32 bit); therefore any value >= 32 would get truncated.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: fe3688ba7950 ("crypto/dpaa_sec: support event crypto adapter")
Cc: akhil.go...@nxp.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
Acked-by: Hemant Agrawal
---
drivers/crypto/dpaa_sec/dpaa_sec.c | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/crypto/dpaa_sec/dpaa_sec.c
b/drivers/crypto/dpaa_sec/dpaa_sec.c
index 3fa88ca968..e117cd77a6 100644
--- a/drivers/crypto/dpaa_sec/dpaa_sec.c
+++ b/drivers/crypto/dpaa_sec/dpaa_sec.c
@@ -1907,13 +1907,12 @@ dpaa_sec_enqueue_burst(void *qp, struct rte_crypto_op
**ops,
op = *(ops++);
if (*dpaa_seqn(op->sym->m_src) != 0) {
index = *dpaa_seqn(op->sym->m_src) - 1;
- if (DPAA_PER_LCORE_DQRR_HELD & (1 << index)) {
+ if (DPAA_PER_LCORE_DQRR_HELD & (UINT64_C(1) <<
index)) {
/* QM_EQCR_DCA_IDXMASK = 0x0f */
flags[loop] = ((index & 0x0f) << 8);
flags[loop] |= QMAN_ENQUEUE_FLAG_DCA;
DPAA_PER_LCORE_DQRR_SIZE--;
- DPAA_PER_LCORE_DQRR_HELD &=
- ~(1 << index);
+ DPAA_PER_LCORE_DQRR_HELD &=
~(UINT64_C(1) << index);
}
}
@@ -3500,7 +3499,7 @@ dpaa_sec_process_atomic_event(void *event,
/* Save active dqrr entries */
index = ((uintptr_t)dqrr >> 6) & (16/*QM_DQRR_SIZE*/ - 1);
DPAA_PER_LCORE_DQRR_SIZE++;
- DPAA_PER_LCORE_DQRR_HELD |= 1 << index;
+ DPAA_PER_LCORE_DQRR_HELD |= UINT64_C(1) << index;
DPAA_PER_LCORE_DQRR_MBUF(index) = ctx->op->sym->m_src;
ev->impl_opaque = index + 1;
*dpaa_seqn(ctx->op->sym->m_src) = (uint32_t)index + 1;
--
2.45.2
[PATCH v2 17/19] net/qede: don't use same loop variable twice
Using variable in outer loop, and inner loop is obvious bug.
This bug is in base code, so likely on other platforms as well.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: 81dba2b2ff61 ("net/qede/base: add LLDP support")
Cc: rasesh.m...@cavium.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
drivers/net/qede/base/ecore_dcbx.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/qede/base/ecore_dcbx.c
b/drivers/net/qede/base/ecore_dcbx.c
index 31234f18cf..72bbedd65a 100644
--- a/drivers/net/qede/base/ecore_dcbx.c
+++ b/drivers/net/qede/base/ecore_dcbx.c
@@ -1363,7 +1363,7 @@ ecore_lldp_mib_update_event(struct ecore_hwfn *p_hwfn,
struct ecore_ptt *p_ptt)
struct ecore_dcbx_mib_meta_data data;
enum _ecore_status_t rc = ECORE_SUCCESS;
struct lldp_received_tlvs_s tlvs;
- int i;
+ int i, j;
for (i = 0; i < LLDP_MAX_LLDP_AGENTS; i++) {
OSAL_MEM_ZERO(&data, sizeof(data));
@@ -1381,9 +1381,9 @@ ecore_lldp_mib_update_event(struct ecore_hwfn *p_hwfn,
struct ecore_ptt *p_ptt)
if (!tlvs.length)
continue;
- for (i = 0; i < MAX_TLV_BUFFER; i++)
- tlvs.tlvs_buffer[i] =
- OSAL_CPU_TO_BE32(tlvs.tlvs_buffer[i]);
+ for (j = 0; j < MAX_TLV_BUFFER; j++)
+ tlvs.tlvs_buffer[j] =
+ OSAL_CPU_TO_BE32(tlvs.tlvs_buffer[j]);
OSAL_LLDP_RX_TLVS(p_hwfn, tlvs.tlvs_buffer, tlvs.length);
}
--
2.45.2
[PATCH v2 11/19] net/hinic: fix flow type bitmask overflow
The type mask is 64 bit value, doing a shift of literal 1 (32 bit)
will result in int type (32 bit) and cause truncation.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: f4ca3fd54c4d ("net/hinic: create and destroy flow director filter")
Cc: cloud.wangxiao...@huawei.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
drivers/net/hinic/hinic_pmd_flow.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/net/hinic/hinic_pmd_flow.c
b/drivers/net/hinic/hinic_pmd_flow.c
index 8fdd5a35be..6b1ca6ff88 100644
--- a/drivers/net/hinic/hinic_pmd_flow.c
+++ b/drivers/net/hinic/hinic_pmd_flow.c
@@ -1979,8 +1979,8 @@ static int hinic_lookup_new_filter(struct
hinic_5tuple_filter *filter,
return -EINVAL;
}
- if (!(filter_info->type_mask & (1 << type_id))) {
- filter_info->type_mask |= 1 << type_id;
+ if (!(filter_info->type_mask & (UINT64_C(1) << type_id))) {
+ filter_info->type_mask |= UINT64_C(1) << type_id;
filter->index = type_id;
filter_info->pkt_filters[type_id].enable = true;
filter_info->pkt_filters[type_id].pkt_proto =
@@ -2138,7 +2138,7 @@ static void hinic_remove_5tuple_filter(struct rte_eth_dev
*dev,
filter_info->pkt_type = 0;
filter_info->qid = 0;
filter_info->pkt_filters[filter->index].qid = 0;
- filter_info->type_mask &= ~(1 << (filter->index));
+ filter_info->type_mask &= ~(UINT64_C(1) << filter->index);
TAILQ_REMOVE(&filter_info->fivetuple_list, filter, entries);
rte_free(filter);
@@ -2268,8 +2268,8 @@ hinic_ethertype_filter_insert(struct hinic_filter_info
*filter_info,
if (id < 0)
return -EINVAL;
- if (!(filter_info->type_mask & (1 << id))) {
- filter_info->type_mask |= 1 << id;
+ if (!(filter_info->type_mask & (UINT64_C(1) << id))) {
+ filter_info->type_mask |= UINT64_C(1) << id;
filter_info->pkt_filters[id].pkt_proto =
ethertype_filter->pkt_proto;
filter_info->pkt_filters[id].enable = ethertype_filter->enable;
@@ -2289,7 +2289,7 @@ hinic_ethertype_filter_remove(struct hinic_filter_info
*filter_info,
return;
filter_info->pkt_type = 0;
- filter_info->type_mask &= ~(1 << idx);
+ filter_info->type_mask &= ~(UINT64_C(1) << idx);
filter_info->pkt_filters[idx].pkt_proto = (uint16_t)0;
filter_info->pkt_filters[idx].enable = FALSE;
filter_info->pkt_filters[idx].qid = 0;
@@ -2355,7 +2355,7 @@ hinic_add_del_ethertype_filter(struct rte_eth_dev *dev,
if (i < 0)
return -EINVAL;
- if ((filter_info->type_mask & (1 << i))) {
+ if ((filter_info->type_mask & (UINT64_C(1) << i))) {
filter_info->pkt_filters[i].enable = FALSE;
(void)hinic_set_fdir_filter(nic_dev->hwdev,
filter_info->pkt_type,
--
2.45.2
[PATCH v2 18/19] examples/l3fwd: fix operator precedence bugs
The expression:
if ((socketid = rte_lcore_to_socket_id(lcore) != 0) &&
gets evaluated as sockeid = (rte_lcore_to_socket_id(lcore) != 0)
which is not what was intended. This is goes all the way back
to first release.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: af75078fece3 ("first public release")
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
examples/l3fwd-power/main.c | 4 ++--
examples/l3fwd/main.c | 5 +++--
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/examples/l3fwd-power/main.c b/examples/l3fwd-power/main.c
index ae8b55924e..7957ea6c95 100644
--- a/examples/l3fwd-power/main.c
+++ b/examples/l3fwd-power/main.c
@@ -1412,8 +1412,8 @@ check_lcore_params(void)
"mask\n", lcore);
return -1;
}
- if ((socketid = rte_lcore_to_socket_id(lcore) != 0) &&
- (numa_on == 0)) {
+ socketid = rte_lcore_to_socket_id(lcore);
+ if (socketid != 0 && numa_on == 0) {
printf("warning: lcore %u is on socket %d with numa "
"off\n", lcore, socketid);
}
diff --git a/examples/l3fwd/main.c b/examples/l3fwd/main.c
index 6e2155e005..14076e07cc 100644
--- a/examples/l3fwd/main.c
+++ b/examples/l3fwd/main.c
@@ -307,8 +307,9 @@ check_lcore_params(void)
printf("error: lcore %u is not enabled in lcore
mask\n", lcore);
return -1;
}
- if ((socketid = rte_lcore_to_socket_id(lcore) != 0) &&
- (numa_on == 0)) {
+
+ socketid = rte_lcore_to_socket_id(lcore);
+ if (socketid != 0 && numa_on == 0) {
printf("warning: lcore %u is on socket %d with numa
off\n",
lcore, socketid);
}
--
2.45.2
[PATCH v2 16/19] net/dpaa2: fix bitmask truncation
The dqrr_held mask is 64 bit but updates were getting truncated
because 1 is of type int (32 bit) and the result shift of int is of
type int (32 bit); therefore any value >= 32 would get truncated.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: 2d3788631862 ("net/dpaa2: support atomic queues")
Cc: nipun.gu...@nxp.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
Acked-by: Hemant Agrawal
---
drivers/net/dpaa2/dpaa2_rxtx.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/dpaa2/dpaa2_rxtx.c b/drivers/net/dpaa2/dpaa2_rxtx.c
index e3b6c7e460..e253bccecd 100644
--- a/drivers/net/dpaa2/dpaa2_rxtx.c
+++ b/drivers/net/dpaa2/dpaa2_rxtx.c
@@ -933,7 +933,7 @@ dpaa2_dev_process_atomic_event(struct qbman_swp *swp
__rte_unused,
dqrr_index = qbman_get_dqrr_idx(dq);
*dpaa2_seqn(ev->mbuf) = dqrr_index + 1;
DPAA2_PER_LCORE_DQRR_SIZE++;
- DPAA2_PER_LCORE_DQRR_HELD |= 1 << dqrr_index;
+ DPAA2_PER_LCORE_DQRR_HELD |= UINT64_C(1) << dqrr_index;
DPAA2_PER_LCORE_DQRR_MBUF(dqrr_index) = ev->mbuf;
}
@@ -1317,7 +1317,7 @@ dpaa2_dev_tx(void *queue, struct rte_mbuf **bufs,
uint16_t nb_pkts)
flags[loop] = QBMAN_ENQUEUE_FLAG_DCA |
dqrr_index;
DPAA2_PER_LCORE_DQRR_SIZE--;
- DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dqrr_index);
+ DPAA2_PER_LCORE_DQRR_HELD &= ~(UINT64_C(1) <<
dqrr_index);
*dpaa2_seqn(*bufs) = DPAA2_INVALID_MBUF_SEQN;
}
@@ -1575,7 +1575,7 @@ dpaa2_set_enqueue_descriptor(struct dpaa2_queue *dpaa2_q,
dq_idx = *dpaa2_seqn(m) - 1;
qbman_eq_desc_set_dca(eqdesc, 1, dq_idx, 0);
DPAA2_PER_LCORE_DQRR_SIZE--;
- DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dq_idx);
+ DPAA2_PER_LCORE_DQRR_HELD &= ~(UINT64_C(1) << dq_idx);
}
*dpaa2_seqn(m) = DPAA2_INVALID_MBUF_SEQN;
}
--
2.45.2
RE: [PATCH v3 1/2] app/testpmd: fix flow update
> -Original Message-
> From: Danylo Vodopianov
> Sent: Monday, November 18, 2024 19:03
> To: NBU-Contact-Thomas Monjalon (EXTERNAL) ;
> aman.deep.si...@intel.com; yuying.zh...@intel.com; Ori Kam
> ; mko-...@napatech.com; c...@napatech.com; Dariusz
> Sosnowski ; sil-...@napatech.com
> Cc: Gregory Etelson ; Alexander Kozyrev
> ; dev@dpdk.org; sta...@dpdk.org;
> ferruh.yi...@amd.com
> Subject: [PATCH v3 1/2] app/testpmd: fix flow update
>
> External email: Use caution opening links or attachments
>
>
> If actions provided to “flow update…“ command contained an age action, then
> testpmd did not update the age action context accordingly.
>
> Thus "flow aged destroy" command can not execute successfully.
>
> Fix was done with next steps
> 1. Generate new port flow entry to add/replace action(s).
> 2. Set age context if age action is present.
> 3. Replace flow in the flow list.
>
> Fixes: 2d9c7e56e52c ("app/testpmd: support updating flow rule actions")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Danylo Vodopianov
> ---
> app/test-pmd/config.c | 21 -
> 1 file changed, 20 insertions(+), 1 deletion(-)
>
> diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c index
> 88770b4dfc..c831166431 100644
> --- a/app/test-pmd/config.c
> +++ b/app/test-pmd/config.c
> @@ -3882,7 +3882,8 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
> const struct rte_flow_action *actions, bool is_user_id) {
> struct rte_port *port;
> - struct port_flow **flow_list;
> + struct port_flow **flow_list, *uf;
> + struct rte_flow_action_age *age = age_action_get(actions);
>
> if (port_id_is_invalid(port_id, ENABLED_WARN) ||
> port_id == (portid_t)RTE_PORT_ALL) @@ -3897,6 +3898,16 @@
> port_flow_update(portid_t port_id, uint32_t rule_id,
> flow_list = &flow->next;
> continue;
> }
> +
> + /* Update flow action(s) with new action(s) */
> + uf = port_flow_new(flow->rule.attr_ro, flow->rule.pattern_ro,
> actions,
> &error);
> + if (!uf)
> + return port_flow_complain(&error);
> + if (age) {
> + flow->age_type = ACTION_AGE_CONTEXT_TYPE_FLOW;
> + age->context = &flow->age_type;
> + }
> +
> /*
> * Poisoning to make sure PMDs update it in case
> * of error.
> @@ -3913,6 +3924,14 @@ port_flow_update(portid_t port_id, uint32_t
> rule_id,
> printf("Flow rule #%"PRIu64
>" updated with new actions\n",
>flow->id);
> +
> + uf->next = flow->next;
> + uf->id = flow->id;
> + uf->user_id = flow->user_id;
> + uf->flow = flow->flow;
> + *flow_list = uf;
> +
> + free(flow);
> return 0;
> }
> printf("Failed to find flow %"PRIu32"\n", rule_id);
> --
> 2.43.5
>
> Disclaimer: This email and any files transmitted with it may contain
> confidential
> information intended for the addressee(s) only. The information is not to be
> surrendered or copied to unauthorized persons. If you have received this
> communication in error, please notify the sender immediately and delete this
> e-
> mail from your system.
Acked-by: Dariusz Sosnowski
Best regards,
Dariusz Sosnowski
Re: [PATCH] eal: fix lcore variables documentation
17/11/2024 20:11, Mattias Rönnblom:
> On 2024-11-14 18:09, Thomas Monjalon wrote:
> > The lcore variables API is new in DPDK 24.11,
> > that's why the function rte_lcore_var_alloc() was marked experimental.
> > To be clearer, the whole header file (including all macros)
> > is marked experimental.
> >
> > A change decreased the default buffer size from 1 MB to 128 kB,
> > missing to update the documentation, which is fixed here.
> >
> > Fixes: 5bce9bed67ad ("eal: add static per-lcore memory allocation facility")
> > Fixes: f2fd6c2e080c ("config: limit lcore variable maximum size to 128k")
> >
> > Signed-off-by: Thomas Monjalon
>
> Acked-by: Mattias Rönnblom
Applied
RE: [PATCH v3 2/2] app/testpmd: fix aged flow destroy
> -Original Message-
> From: Danylo Vodopianov
> Sent: Monday, November 18, 2024 19:03
> To: NBU-Contact-Thomas Monjalon (EXTERNAL) ;
> aman.deep.si...@intel.com; yuying.zh...@intel.com; Ori Kam
> ; mko-...@napatech.com; c...@napatech.com; Dariusz
> Sosnowski ; sil-...@napatech.com
> Cc: Gregory Etelson ; Alexander Kozyrev
> ; dev@dpdk.org; sta...@dpdk.org;
> ferruh.yi...@amd.com
> Subject: [PATCH v3 2/2] app/testpmd: fix aged flow destroy
>
> External email: Use caution opening links or attachments
>
>
> port_flow_destroy() function never assumed that rule array can be freed when
> it's executing, and port_flow_aged() just violated that assumption.
>
> In case of flow async create failure, it tries to do a cleanup, but it
> wrongly removes
> a 1st flow (with id 0). pf->id is not set at this moment and it always is 0,
> thus 1st
> flow is removed. A local copy of
> flow->id must be used to call of port_flow_destroy() to avoid access
> and processing of flow->id after the flow is removed.
>
> Fixes: de956d5ecf08 ("app/testpmd: support age shared action context")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Danylo Vodopianov
> ---
> app/test-pmd/config.c | 7 +--
> 1 file changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c index
> c831166431..28d45568ac 100644
> --- a/app/test-pmd/config.c
> +++ b/app/test-pmd/config.c
> @@ -4160,8 +4160,10 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
> }
> type = (enum age_action_context_type *)contexts[idx];
> switch (*type) {
> - case ACTION_AGE_CONTEXT_TYPE_FLOW:
> + case ACTION_AGE_CONTEXT_TYPE_FLOW: {
> + uint64_t flow_id;
> ctx.pf = container_of(type, struct port_flow,
> age_type);
> + flow_id = ctx.pf->id;
> printf("%-20s\t%" PRIu64 "\t%" PRIu32 "\t%" PRIu32
>
> "\t%c%c%c\t\n",
>"Flow",
> @@ -4172,9 +4174,10 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
>ctx.pf->rule.attr->egress ? 'e' : '-',
>ctx.pf->rule.attr->transfer ? 't' : '-');
> if (destroy && !port_flow_destroy(port_id, 1,
> - &ctx.pf->id, false))
> + &flow_id,
> + false))
> total++;
> break;
> + }
> case ACTION_AGE_CONTEXT_TYPE_INDIRECT_ACTION:
> ctx.pia = container_of(type,
> struct port_indirect_action,
> age_type);
> --
> 2.43.5
>
> Disclaimer: This email and any files transmitted with it may contain
> confidential
> information intended for the addressee(s) only. The information is not to be
> surrendered or copied to unauthorized persons. If you have received this
> communication in error, please notify the sender immediately and delete this
> e-
> mail from your system.
Acked-by: Dariusz Sosnowski
Best regards,
Dariusz Sosnowski
[PATCH v2 01/19] common/cnxk: remove duplicate condition
The same condition is checked twice in an if statement.
Harmless, but redundant.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Signed-off-by: Stephen Hemminger
Acked-by: Anoob Joseph
---
drivers/common/cnxk/cnxk_security.c | 16 ++--
1 file changed, 10 insertions(+), 6 deletions(-)
diff --git a/drivers/common/cnxk/cnxk_security.c
b/drivers/common/cnxk/cnxk_security.c
index c2871ad2bd..9446c14ac8 100644
--- a/drivers/common/cnxk/cnxk_security.c
+++ b/drivers/common/cnxk/cnxk_security.c
@@ -174,9 +174,11 @@ ot_ipsec_sa_common_param_fill(union roc_ot_ipsec_sa_word2
*w2, uint8_t *cipher_k
}
/* Set AES key length */
- if (w2->s.enc_type == ROC_IE_SA_ENC_AES_CBC || w2->s.enc_type ==
ROC_IE_SA_ENC_AES_CCM ||
- w2->s.enc_type == ROC_IE_SA_ENC_AES_CTR || w2->s.enc_type ==
ROC_IE_SA_ENC_AES_GCM ||
- w2->s.enc_type == ROC_IE_SA_ENC_AES_CCM || w2->s.auth_type ==
ROC_IE_SA_AUTH_AES_GMAC) {
+ if (w2->s.enc_type == ROC_IE_SA_ENC_AES_CBC ||
+ w2->s.enc_type == ROC_IE_SA_ENC_AES_CTR ||
+ w2->s.enc_type == ROC_IE_SA_ENC_AES_GCM ||
+ w2->s.enc_type == ROC_IE_SA_ENC_AES_CCM ||
+ w2->s.auth_type == ROC_IE_SA_AUTH_AES_GMAC) {
switch (length) {
case ROC_CPT_AES128_KEY_LEN:
w2->s.aes_key_len = ROC_IE_SA_AES_KEY_LEN_128;
@@ -879,9 +881,11 @@ on_ipsec_sa_ctl_set(struct rte_security_ipsec_xform *ipsec,
}
/* Set AES key length */
- if (ctl->enc_type == ROC_IE_SA_ENC_AES_CBC || ctl->enc_type ==
ROC_IE_SA_ENC_AES_CCM ||
- ctl->enc_type == ROC_IE_SA_ENC_AES_CTR || ctl->enc_type ==
ROC_IE_SA_ENC_AES_GCM ||
- ctl->enc_type == ROC_IE_SA_ENC_AES_CCM || ctl->auth_type ==
ROC_IE_SA_AUTH_AES_GMAC) {
+ if (ctl->enc_type == ROC_IE_SA_ENC_AES_CBC ||
+ ctl->enc_type == ROC_IE_SA_ENC_AES_CTR ||
+ ctl->enc_type == ROC_IE_SA_ENC_AES_GCM ||
+ ctl->enc_type == ROC_IE_SA_ENC_AES_CCM ||
+ ctl->auth_type == ROC_IE_SA_AUTH_AES_GMAC) {
switch (aes_key_len) {
case 16:
ctl->aes_key_len = ROC_IE_SA_AES_KEY_LEN_128;
--
2.45.2
[PATCH v2 04/19] net/ntnic: remove dead code
The loop to update speed would not work because num_port_speeds
was always zero so it did nothing. And the array of pls_mbps
was only used inside the loop but never set.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Signed-off-by: Stephen Hemminger
---
drivers/net/ntnic/ntnic_ethdev.c | 8
1 file changed, 8 deletions(-)
diff --git a/drivers/net/ntnic/ntnic_ethdev.c b/drivers/net/ntnic/ntnic_ethdev.c
index 2a2643a106..467fea4bf2 100644
--- a/drivers/net/ntnic/ntnic_ethdev.c
+++ b/drivers/net/ntnic/ntnic_ethdev.c
@@ -2037,8 +2037,6 @@ nthw_pci_dev_init(struct rte_pci_device *pci_dev)
uint32_t exception_path = 0;
struct flow_queue_id_s queue_ids[MAX_QUEUES];
int n_phy_ports;
- struct port_link_speed pls_mbps[NUM_ADAPTER_PORTS_MAX] = { 0 };
- int num_port_speeds = 0;
enum flow_eth_dev_profile profile = FLOW_ETH_DEV_PROFILE_INLINE;
NT_LOG_DBGX(DBG, NTNIC, "Dev %s PF #%i Init : %02x:%02x:%i",
pci_dev->name,
@@ -2178,12 +2176,6 @@ nthw_pci_dev_init(struct rte_pci_device *pci_dev)
p_nt_drv->b_shutdown = false;
p_nt_drv->adapter_info.pb_shutdown = &p_nt_drv->b_shutdown;
- for (int i = 0; i < num_port_speeds; ++i) {
- struct adapter_info_s *p_adapter_info = &p_nt_drv->adapter_info;
- nt_link_speed_t link_speed =
convert_link_speed(pls_mbps[i].link_speed);
- port_ops->set_link_speed(p_adapter_info, i, link_speed);
- }
-
/* store context */
store_pdrv(p_drv);
--
2.45.2
[PATCH v2 19/19] common/cnxk: fix null ptr check
The pointer mode is used then checked which is a bug reported
by PVS studio and Coverity.
Fixes: bd2fd34ab86f ("common/cnxk: sync eth mode change command with firmware")
Cc: tduszyn...@marvell.com
Signed-off-by: Stephen Hemminger
---
drivers/common/cnxk/roc_bphy_cgx.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/common/cnxk/roc_bphy_cgx.c
b/drivers/common/cnxk/roc_bphy_cgx.c
index 4f43605e10..6e5cfc7b1d 100644
--- a/drivers/common/cnxk/roc_bphy_cgx.c
+++ b/drivers/common/cnxk/roc_bphy_cgx.c
@@ -369,20 +369,20 @@ roc_bphy_cgx_set_link_mode(struct roc_bphy_cgx *roc_cgx,
unsigned int lmac,
{
uint64_t scr1, scr0;
+ if (!mode)
+ return -EINVAL;
+
+ if (!roc_cgx)
+ return -EINVAL;
+
if (roc_model_is_cn9k() &&
(mode->use_portm_idx || mode->portm_idx || mode->mode_group_idx)) {
return -ENOTSUP;
}
- if (!roc_cgx)
- return -EINVAL;
-
if (!roc_bphy_cgx_lmac_exists(roc_cgx, lmac))
return -ENODEV;
- if (!mode)
- return -EINVAL;
-
scr1 = FIELD_PREP(SCR1_ETH_CMD_ID, ETH_CMD_MODE_CHANGE) |
FIELD_PREP(SCR1_ETH_MODE_CHANGE_ARGS_SPEED, mode->speed) |
FIELD_PREP(SCR1_ETH_MODE_CHANGE_ARGS_DUPLEX, mode->full_duplex) |
--
2.45.2
[PATCH v11] dts: port over queue start/stop suite
This suite tests the ability of the Poll Mode Driver to enable
and disable Rx/Tx queues on a port.
Depends-on: patch-12 ("dts: add port queue modification
and forwarding stats to testpmd")
Signed-off-by: Dean Marx
Reviewed-by: Jeremy Spewock
---
dts/framework/config/conf_yaml_schema.json | 3 +-
dts/tests/TestSuite_queue_start_stop.py| 89 ++
2 files changed, 91 insertions(+), 1 deletion(-)
create mode 100644 dts/tests/TestSuite_queue_start_stop.py
diff --git a/dts/framework/config/conf_yaml_schema.json
b/dts/framework/config/conf_yaml_schema.json
index df390e8ae2..12a4a26dc8 100644
--- a/dts/framework/config/conf_yaml_schema.json
+++ b/dts/framework/config/conf_yaml_schema.json
@@ -187,7 +187,8 @@
"enum": [
"hello_world",
"os_udp",
-"pmd_buffer_scatter"
+"pmd_buffer_scatter",
+"queue_start_stop"
]
},
"test_target": {
diff --git a/dts/tests/TestSuite_queue_start_stop.py
b/dts/tests/TestSuite_queue_start_stop.py
new file mode 100644
index 00..389030ae8c
--- /dev/null
+++ b/dts/tests/TestSuite_queue_start_stop.py
@@ -0,0 +1,89 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2024 University of New Hampshire
+
+"""Rx/Tx queue start and stop functionality suite.
+
+This suite tests the ability of the poll mode driver to start and
+stop either the Rx or Tx queue (depending on the port) during runtime,
+and verify that packets are not received when one is disabled.
+
+Given a paired port topology, the Rx queue will be disabled on port 0,
+and the Tx queue will be disabled on port 1.
+
+"""
+
+from scapy.layers.inet import IP # type: ignore[import-untyped]
+from scapy.layers.l2 import Ether # type: ignore[import-untyped]
+from scapy.packet import Raw # type: ignore[import-untyped]
+
+from framework.remote_session.testpmd_shell import SimpleForwardingModes,
TestPmdShell
+from framework.test_suite import TestSuite, func_test
+from framework.testbed_model.capability import NicCapability, TopologyType,
requires
+
+
+@requires(topology_type=TopologyType.two_links)
+@requires(NicCapability.RUNTIME_RX_QUEUE_SETUP)
+@requires(NicCapability.RUNTIME_TX_QUEUE_SETUP)
+class TestQueueStartStop(TestSuite):
+"""DPDK Queue start/stop test suite.
+
+Ensures Rx/Tx queue on a port can be disabled and enabled.
+Verifies packets are not received when either queue is disabled.
+The suite contains two test cases, Rx queue start/stop and
+Tx queue start/stop, which each disable the corresponding
+queue and verify that packets are not received/forwarded.
+"""
+
+def send_packet_and_verify(self, should_receive: bool = True) -> None:
+"""Generate a packet, send to the DUT, and verify it is forwarded back.
+
+Args:
+should_receive: Indicate whether the packet should be received.
+"""
+packet = Ether() / IP() / Raw(load="x")
+received = self.send_packet_and_capture(packet)
+contains_packet = any(
+packet.haslayer(Raw) and b"x" in packet.load for packet in
received
+)
+self.verify(
+should_receive == contains_packet,
+f"Packet was {'dropped' if should_receive else 'received'}",
+)
+
+@func_test
+def test_rx_queue_start_stop(self) -> None:
+"""Verify packets are not received by port 0 when Rx queue is disabled.
+
+Test:
+Create an interactive testpmd session, stop Rx queue on port 0,
verify
+packets are not received.
+"""
+with TestPmdShell(node=self.sut_node) as testpmd:
+testpmd.set_forward_mode(SimpleForwardingModes.mac)
+testpmd.stop_port_queue(0, 0, True)
+testpmd.start()
+self.send_packet_and_verify(should_receive=False)
+stats = testpmd.show_port_stats(port_id=0)
+self.verify(
+stats.rx_packets == 0,
+"Packets were received on Rx queue when it should've been
disabled",
+)
+
+@func_test
+def test_tx_queue_start_stop(self) -> None:
+"""Verify packets are not forwarded by port 1 when Tx queue is
disabled.
+
+Test:
+Create an interactive testpmd session, stop Tx queue on port 1,
verify
+packets are not forwarded.
+"""
+with TestPmdShell(node=self.sut_node) as testpmd:
+testpmd.set_forward_mode(SimpleForwardingModes.mac)
+testpmd.stop_port_queue(1, 0, False)
+testpmd.start()
+self.send_packet_and_verify(should_receive=False)
+stats = testpmd.show_port_stats(port_id=1)
+self.verify(
+stats.tx_packets == 0,
+"Packets were forwarded on Tx queue when it should've been
disabled",
+)
--
2.44.0
Re: [PATCH] net/txgbe: fix a mass of interrupts
On 11/15/2024 8:33 AM, Jiawen Wu wrote: > Since firmware version 0x20010, GPIO interrupt enable is set to 0xd by > default, which means enable bit 0 'tx_fault'. And GPIO interrupt polarity > is set to 0xd by default too, which means these interrupts are rising-edge > sensitive. > > So when unplug the SFP module, GPIO line 0 'tx_fault' is 0 -> 1 triggers > the interrupt. However, the interrupt is not cleared. And GPIO interrupt > mask is enabled and disabled to trigger the MISC interrupt repeatedly. > > Since this 'tx_fault' interrupt does not make much sense, simply clear it > to fix the issue. > > Signed-off-by: Jiawen Wu > Hi Jiawen, Can you please provide a fixes tag? Also I believe you would like to backport this fix, if so please add the stable tag as well.
RE: [PATCH v2 1/2] app/testpmd: fix flow update
> -Original Message-
> From: Danylo Vodopianov
> Sent: Monday, November 18, 2024 11:25
> To: NBU-Contact-Thomas Monjalon (EXTERNAL) ;
> aman.deep.si...@intel.com; yuying.zh...@intel.com; Ori Kam
> ; mko-...@napatech.com; c...@napatech.com; Dariusz
> Sosnowski ; sil-...@napatech.com
> Cc: Gregory Etelson ; Alexander Kozyrev
> ; dev@dpdk.org; sta...@dpdk.org;
> ferruh.yi...@amd.com
> Subject: [PATCH v2 1/2] app/testpmd: fix flow update
>
> External email: Use caution opening links or attachments
>
>
> If actions provided to “flow update…“ command contained an age action, then
> testpmd did not update the age action context accordingly.
>
> Thus "flow aged destroy" command can not execute successfully.
>
> Fix was done with next steps
> 1. Generate new port flow entry to add/replace action(s).
> 2. Set age context if age action is present.
> 3. Replace flow in the flow list.
>
> Fixes: 2d9c7e56e52c ("app/testpmd: support updating flow rule actions")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Danylo Vodopianov
> ---
> app/test-pmd/config.c | 21 -
> 1 file changed, 20 insertions(+), 1 deletion(-)
>
> diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c index
> 88770b4dfc..c831166431 100644
> --- a/app/test-pmd/config.c
> +++ b/app/test-pmd/config.c
> @@ -3882,7 +3882,8 @@ port_flow_update(portid_t port_id, uint32_t
> rule_id,
> const struct rte_flow_action *actions, bool is_user_id) {
> struct rte_port *port;
> - struct port_flow **flow_list;
> + struct port_flow **flow_list, *uf;
> + struct rte_flow_action_age *age = age_action_get(actions);
>
> if (port_id_is_invalid(port_id, ENABLED_WARN) ||
> port_id == (portid_t)RTE_PORT_ALL) @@ -3897,6 +3898,16 @@
> port_flow_update(portid_t port_id, uint32_t rule_id,
> flow_list = &flow->next;
> continue;
> }
> +
> + /* Update flow action(s) with new action(s) */
> + uf = port_flow_new(flow->rule.attr_ro, flow->rule.pattern_ro,
> actions, &error);
> + if (!uf)
> + return port_flow_complain(&error);
> + if (age) {
> + flow->age_type = ACTION_AGE_CONTEXT_TYPE_FLOW;
> + age->context = &flow->age_type;
> + }
> +
> /*
> * Poisoning to make sure PMDs update it in case
> * of error.
> @@ -3913,6 +3924,14 @@ port_flow_update(portid_t port_id, uint32_t
> rule_id,
> printf("Flow rule #%"PRIu64
>" updated with new actions\n",
>flow->id);
> +
> + uf->next = flow->next;
> + uf->id = flow->id;
> + uf->user_id = flow->user_id;
> + uf->flow = flow->flow;
> + *flow_list = uf;
> +
> + free(flow);
> return 0;
> }
> printf("Failed to find flow %"PRIu32"\n", rule_id);
> --
> 2.43.5
Acked-by: Dariusz Sosnowski
Best regards,
Dariusz Sosnowski
Re: [PATCH v6 1/2] dts: add port queue modification and forwarding stats to testpmd
Reviewed-by: Patrick Robb
RE: [PATCH] net/txgbe: fix a mass of interrupts
On Tue, Nov 19, 2024 3:59 AM, Ferruh Yigit wrote:
> On 11/15/2024 8:33 AM, Jiawen Wu wrote:
> > Since firmware version 0x20010, GPIO interrupt enable is set to 0xd by
> > default, which means enable bit 0 'tx_fault'. And GPIO interrupt polarity
> > is set to 0xd by default too, which means these interrupts are rising-edge
> > sensitive.
> >
> > So when unplug the SFP module, GPIO line 0 'tx_fault' is 0 -> 1 triggers
> > the interrupt. However, the interrupt is not cleared. And GPIO interrupt
> > mask is enabled and disabled to trigger the MISC interrupt repeatedly.
> >
> > Since this 'tx_fault' interrupt does not make much sense, simply clear it
> > to fix the issue.
> >
> > Signed-off-by: Jiawen Wu
> >
>
> Hi Jiawen,
>
> Can you please provide a fixes tag?
> Also I believe you would like to backport this fix, if so please add the
> stable tag as well.
I have thought about it. But there are two changes based on two commits:
1. d3bb4a04eac1 ("net/txgbe: add SFP hotplug identification")
2. 12011b11a3d6 ("net/txgbe: adapt to MNG veto bit setting")
Should I split the patch and add fixes tags separately?
Re: [PATCH v1 0/4] TruFlow fixes for Thor2
On Mon, Nov 18, 2024 at 8:07 AM Sriharsha Basavapatna wrote: > > This patch series fixes a few issues in TruFlow for Broadcom Thor2 NIC. Patchset applied to dpdk-next-net-brcm. Fixed up a couple of typos during merge. > > Kishore Padmanabha (2): > net/bnxt/tf_ulp: fix vfr clean up and stats lockup > net/bnxt/tf_ulp: update template files > > Peter Spreadborough (1): > net/bnxt/tf_ulp: performance and tuning changes for thor2 stats cache > > Shuanglin Wang (1): > net/bnxt/tf_ulp: fix F1F2 vxlan counter acccumulation for Thor2 > > drivers/net/bnxt/tf_core/v3/tfc.h |12 +- > drivers/net/bnxt/tf_core/v3/tfc_act.c |26 +- > drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c|12 +- > drivers/net/bnxt/tf_ulp/bnxt_ulp.c|17 + > drivers/net/bnxt/tf_ulp/bnxt_ulp.h| 3 + > .../generic_templates/ulp_template_db_enum.h | 8 +- > .../ulp_template_db_thor2_class.c | 10126 > .../ulp_template_db_thor_class.c | 8736 +++-- > drivers/net/bnxt/tf_ulp/ulp_fc_mgr_tfc.c | 5 +- > drivers/net/bnxt/tf_ulp/ulp_flow_db.c | 4 +- > drivers/net/bnxt/tf_ulp/ulp_flow_db.h |13 + > drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c | 309 +- > drivers/net/bnxt/tf_ulp/ulp_sc_mgr.h |33 +- > drivers/net/bnxt/tf_ulp/ulp_sc_mgr_tfc.c | 4 +- > 14 files changed, 9627 insertions(+), 9681 deletions(-) > > -- > 2.39.3 > smime.p7s Description: S/MIME Cryptographic Signature
[DPDK/DTS Bug 1584] dts: rxq info dataclass fields were dropped
https://bugs.dpdk.org/show_bug.cgi?id=1584 Bug ID: 1584 Summary: dts: rxq info dataclass fields were dropped Product: DPDK Version: unspecified Hardware: All OS: All Status: UNCONFIRMED Severity: minor Priority: Normal Component: DTS Assignee: dev@dpdk.org Reporter: pr...@iol.unh.edu CC: juraj.lin...@pantheon.tech, pr...@iol.unh.edu Target Milestone: --- https://patchwork.dpdk.org/project/dpdk/patch/20241105165830.15881-2-dm...@iol.unh.edu/ In the series above, any fields which were previously a part of TestPmdRxqInfo dataclass, and were not needed for our current testsuites, were dropped when we shifted to a rxqinfo base class and 2 child classes for txq and rxq respectively. I can see why this was done, but it is worth discussing at an upcoming meeting whether we should continue with this practice in other cases, or attempt to provide complete structures for testpmd output, even before they're needed for a testsuite. In my view it is simpler to just support all fields possible in the given dataclass instead of add them in piecemeal as testsuites are added. For context, the original dataclass, which represents all the values returned from "show rxq info ": --- @dataclass class TestPmdRxqInfo(TextParser): """Representation of testpmd's ``show rxq info `` command. References: testpmd command function: ``app/test-pmd/cmdline.c:cmd_showqueue()`` testpmd display function: ``app/test-pmd/config.c:rx_queue_infos_display()`` """ #: port_id: int = field(metadata=TextParser.find_int(r"Infos for port (\d+)\b ?, RX queue \d+\b")) #: queue_id: int = field(metadata=TextParser.find_int(r"Infos for port \d+\b ?, RX queue (\d+)\b")) #: Mempool used by that queue mempool: str = field(metadata=TextParser.find(r"Mempool: ([^\r\n]+)")) #: Ring prefetch threshold rx_prefetch_threshold: int = field( metadata=TextParser.find_int(r"RX prefetch threshold: (\d+)\b") ) #: Ring host threshold rx_host_threshold: int = field(metadata=TextParser.find_int(r"RX host threshold: (\d+)\b")) #: Ring writeback threshold rx_writeback_threshold: int = field( metadata=TextParser.find_int(r"RX writeback threshold: (\d+)\b") ) #: Drives the freeing of Rx descriptors rx_free_threshold: int = field(metadata=TextParser.find_int(r"RX free threshold: (\d+)\b")) #: Drop packets if no descriptors are available rx_drop_packets: bool = field(metadata=TextParser.find(r"RX drop packets: on")) #: Do not start queue with rte_eth_dev_start() rx_deferred_start: bool = field(metadata=TextParser.find(r"RX deferred start: on")) #: Scattered packets Rx enabled rx_scattered_packets: bool = field(metadata=TextParser.find(r"RX scattered packets: on")) #: The state of the queue rx_queue_state: str = field(metadata=RxQueueState.make_parser()) #: Configured number of RXDs number_of_rxds: int = field(metadata=TextParser.find_int(r"Number of RXDs: (\d+)\b")) #: Hardware receive buffer size rx_buffer_size: int | None = field( default=None, metadata=TextParser.find_int(r"RX buffer size: (\d+)\b") ) #: Burst mode information burst_mode: str | None = field( default=None, metadata=TextParser.find(r"Burst mode: ([^\r\n]+)") ) -- You are receiving this mail because: You are the assignee for the bug.
Re: [PATCH v6 2/2] dts: add dynamic queue test suite
Applied to next-dts, thanks.
Re: [PATCH v6 2/2] dts: add dynamic queue test suite
Reviewed-by: Patrick Robb
Re: [PATCH v3 0/2] bnxt patchset
On Mon, Nov 18, 2024 at 1:13 PM Ajit Khaparde wrote: > > v1->v2: Rebasing patchset based on latest dpdk tree. > v2->v3: Dropped an incomplete patch which ended up on mailing list. Patches rebased and applied to dpdk-next-net-brcm. > > Please apply. > > Ajit Khaparde (2): > net/bnxt: update HWRM API > net/bnxt: add support for Rx profile selection > > drivers/net/bnxt/bnxt.h| 1 + > drivers/net/bnxt/bnxt_hwrm.c | 15 +- > drivers/net/bnxt/hsi_struct_def_dpdk.h | 71 +- > 3 files changed, 84 insertions(+), 3 deletions(-) > > -- > 2.39.5 (Apple Git-154) > smime.p7s Description: S/MIME Cryptographic Signature
Re: [PATCH v6 1/2] dts: add port queue modification and forwarding stats to testpmd
Although I don't think it's necessarily wrong, I wanted to note for the broader group that this series is indeed dropping rxq info dataclass fields which were previously supported. It is worth discussing at the next DTS meeting so we have better clarity on how to handle these situations going forward. I've created a ticket: https://bugs.dpdk.org/show_bug.cgi?id=1584 Applied to next-dts, thanks.
答复: [PATCH v8 12/17] net/r8169: implement Tx path
Dear Thomas,
I will consider that. By the way, should I make the changes and then upload
patch v9?
Thanks!
-邮件原件-
发件人: Thomas Monjalon
发送时间: 2024年11月18日 22:59
收件人: pro_nic_d...@realtek.com; 王颢
抄送: dev@dpdk.org; Ferruh Yigit
主题: Re: [PATCH v8 12/17] net/r8169: implement Tx path
External mail.
Hello,
13/11/2024 10:28, Howard Wang:
> +static u32
> +rtl_get_opts1(struct rtl_tx_desc *txd) {
> + rte_smp_rmb();
> +
> + return rte_le_to_cpu_32(txd->opts1); }
We should avoid using such heavy memory barrier.
Is it possible to use a lighter fence or atomic API?
[PATCH v2 09/19] examples/ptpclient: fix self memcmp
Calling memcmp on same structure will always be true.
Replace with same conditional used elsewhere.
Link: https://pvs-studio.com/en/blog/posts/cpp/1183/
Fixes: ab129e9065a5 ("examples/ptpclient: add minimal PTP client")
Cc: danielx.t.mrzyg...@intel.com
Cc: sta...@dpdk.org
Signed-off-by: Stephen Hemminger
---
examples/ptpclient/ptpclient.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/examples/ptpclient/ptpclient.c b/examples/ptpclient/ptpclient.c
index 2ec532d058..d6dff2eb7e 100644
--- a/examples/ptpclient/ptpclient.c
+++ b/examples/ptpclient/ptpclient.c
@@ -419,7 +419,7 @@ parse_sync(struct ptpv2_time_receiver_ordinary *ptp_data,
uint16_t rx_tstamp_idx
ptp_data->ptpset = 1;
}
- if (memcmp(&ptp_hdr->source_port_id.clock_id,
+ if (memcmp(&ptp_data->transmitter_clock_id,
&ptp_hdr->source_port_id.clock_id,
sizeof(struct clock_id)) == 0) {
--
2.45.2
[PATCH v5 03/16] net: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
lib/net/rte_arp.h | 2 ++
lib/net/rte_dtls.h | 1 +
lib/net/rte_esp.h | 2 ++
lib/net/rte_geneve.h | 1 +
lib/net/rte_gre.h | 4
lib/net/rte_gtp.h | 5 +
lib/net/rte_ib.h | 1 +
lib/net/rte_icmp.h | 3 +++
lib/net/rte_ip4.h | 1 +
lib/net/rte_ip6.h | 4
lib/net/rte_l2tpv2.h | 4
lib/net/rte_macsec.h | 2 ++
lib/net/rte_mpls.h | 1 +
lib/net/rte_pdcp_hdr.h | 4
lib/net/rte_ppp.h | 1 +
lib/net/rte_sctp.h | 1 +
lib/net/rte_tcp.h | 1 +
lib/net/rte_tls.h | 1 +
lib/net/rte_udp.h | 1 +
lib/net/rte_vxlan.h| 7 +++
20 files changed, 47 insertions(+)
diff --git a/lib/net/rte_arp.h b/lib/net/rte_arp.h
index 668cea1704..35d68c108d 100644
--- a/lib/net/rte_arp.h
+++ b/lib/net/rte_arp.h
@@ -21,6 +21,7 @@ extern "C" {
/**
* ARP header IPv4 payload.
*/
+__rte_msvc_pack
struct __rte_aligned(2) rte_arp_ipv4 {
struct rte_ether_addr arp_sha; /**< sender hardware address */
rte_be32_tarp_sip; /**< sender IP address */
@@ -31,6 +32,7 @@ struct __rte_aligned(2) rte_arp_ipv4 {
/**
* ARP header.
*/
+__rte_msvc_pack
struct __rte_aligned(2) rte_arp_hdr {
rte_be16_t arp_hardware; /**< format of hardware address */
#define RTE_ARP_HRD_ETHER 1 /**< ARP Ethernet address format */
diff --git a/lib/net/rte_dtls.h b/lib/net/rte_dtls.h
index 246cd8a72d..a6243a2791 100644
--- a/lib/net/rte_dtls.h
+++ b/lib/net/rte_dtls.h
@@ -30,6 +30,7 @@
* DTLS Header
*/
__extension__
+__rte_msvc_pack
struct rte_dtls_hdr {
/** Content type of DTLS packet. Defined as RTE_DTLS_TYPE_*. */
uint8_t type;
diff --git a/lib/net/rte_esp.h b/lib/net/rte_esp.h
index 745a9847fe..ff7b5fe585 100644
--- a/lib/net/rte_esp.h
+++ b/lib/net/rte_esp.h
@@ -16,6 +16,7 @@
/**
* ESP Header
*/
+__rte_msvc_pack
struct rte_esp_hdr {
rte_be32_t spi; /**< Security Parameters Index */
rte_be32_t seq; /**< packet sequence number */
@@ -24,6 +25,7 @@ struct rte_esp_hdr {
/**
* ESP Trailer
*/
+__rte_msvc_pack
struct rte_esp_tail {
uint8_t pad_len; /**< number of pad bytes (0-255) */
uint8_t next_proto; /**< IPv4 or IPv6 or next layer header */
diff --git a/lib/net/rte_geneve.h b/lib/net/rte_geneve.h
index eb2c85f1e9..8bea7d288b 100644
--- a/lib/net/rte_geneve.h
+++ b/lib/net/rte_geneve.h
@@ -34,6 +34,7 @@
* More-bits (optional) variable length options.
*/
__extension__
+__rte_msvc_pack
struct rte_geneve_hdr {
#if RTE_BYTE_ORDER == RTE_BIG_ENDIAN
uint8_t ver:2; /**< Version. */
diff --git a/lib/net/rte_gre.h b/lib/net/rte_gre.h
index 1483e1b42d..48bd5b1621 100644
--- a/lib/net/rte_gre.h
+++ b/lib/net/rte_gre.h
@@ -23,6 +23,7 @@
* GRE Header
*/
__extension__
+__rte_msvc_pack
struct rte_gre_hdr {
#if RTE_BYTE_ORDER == RTE_LITTLE_ENDIAN
uint16_t res2:4; /**< Reserved */
@@ -47,6 +48,7 @@ struct rte_gre_hdr {
/**
* Optional field checksum in GRE header
*/
+__rte_msvc_pack
struct rte_gre_hdr_opt_checksum_rsvd {
rte_be16_t checksum;
rte_be16_t reserved1;
@@ -55,6 +57,7 @@ struct rte_gre_hdr_opt_checksum_rsvd {
/**
* Optional field key in GRE header
*/
+__rte_msvc_pack
struct rte_gre_hdr_opt_key {
rte_be32_t key;
} __rte_packed;
@@ -62,6 +65,7 @@ struct rte_gre_hdr_opt_key {
/**
* Optional field sequence in GRE header
*/
+__rte_msvc_pack
struct rte_gre_hdr_opt_sequence {
rte_be32_t sequence;
} __rte_packed;
diff --git a/lib/net/rte_gtp.h b/lib/net/rte_gtp.h
index ab06e23a6e..09d7d8e3a9 100644
--- a/lib/net/rte_gtp.h
+++ b/lib/net/rte_gtp.h
@@ -24,6 +24,7 @@
* No optional fields and next extension header.
*/
__extension__
+__rte_msvc_pack
struct rte_gtp_hdr {
union {
uint8_t gtp_hdr_info; /**< GTP header info */
@@ -51,6 +52,7 @@ struct rte_gtp_hdr {
} __rte_packed;
/* Optional word of GTP header, present if any of E, S, PN is set. */
+__rte_msvc_pack
struct rte_gtp_hdr_ext_word {
rte_be16_t sqn; /**< Sequence Number. */
uint8_t npdu; /**< N-PDU number. */
@@ -62,6 +64,7 @@ struct rte_gtp_hdr_ext_word {
* defined based on RFC 38415-g30.
*/
__extension__
+__rte_msvc_pack
struct rte_gtp_psc_generic_hdr {
uint8_t ext_hdr_len;/**< PDU ext hdr len in multiples of 4 bytes */
#if RTE_BYTE_ORDER == RTE_BIG_ENDIAN
@@ -85,6 +88,7 @@ struct rte_gtp_psc_generic_hdr {
* type0 defined based on RFC 38415-g30
*/
__extension__
+__rte_msvc_pack
struct rte_gtp_psc_type0_hdr {
uint8_t ext_hdr_len;/**< PDU ext hdr len in multiples of 4 bytes */
#if RTE_BYTE_ORDER == RTE_BIG_ENDIAN
@@ -112,6 +116,7 @@ struct rte_gtp_psc_type0_hdr {
* type1 defined based on RFC 38415-g30
*/
__extension__
+__rte_msvc_pack
struct rte_gtp_psc_type1_hdr {
[PATCH v5 06/16] common/mlx5: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
drivers/common/mlx5/mlx5_common_mr.h| 4
drivers/common/mlx5/mlx5_common_utils.h | 1 +
drivers/common/mlx5/mlx5_prm.h | 30 +
3 files changed, 35 insertions(+)
diff --git a/drivers/common/mlx5/mlx5_common_mr.h
b/drivers/common/mlx5/mlx5_common_mr.h
index a7f1042037..a1ea68504e 100644
--- a/drivers/common/mlx5/mlx5_common_mr.h
+++ b/drivers/common/mlx5/mlx5_common_mr.h
@@ -49,6 +49,7 @@ struct mlx5_mr {
};
/* Cache entry for Memory Region. */
+__rte_msvc_pack
struct mr_cache_entry {
uintptr_t start; /* Start address of MR. */
uintptr_t end; /* End address of MR. */
@@ -56,6 +57,7 @@ struct mr_cache_entry {
} __rte_packed;
/* MR Cache table for Binary search. */
+__rte_msvc_pack
struct mlx5_mr_btree {
uint32_t len; /* Number of entries. */
uint32_t size; /* Total number of entries. */
@@ -65,6 +67,7 @@ struct mlx5_mr_btree {
struct mlx5_common_device;
/* Per-queue MR control descriptor. */
+__rte_msvc_pack
struct mlx5_mr_ctrl {
uint32_t *dev_gen_ptr; /* Generation number of device to poll. */
uint32_t cur_gen; /* Generation number saved to flush caches. */
@@ -78,6 +81,7 @@ LIST_HEAD(mlx5_mr_list, mlx5_mr);
LIST_HEAD(mlx5_mempool_reg_list, mlx5_mempool_reg);
/* Global per-device MR cache. */
+__rte_msvc_pack
struct mlx5_mr_share_cache {
uint32_t dev_gen; /* Generation number to flush local caches. */
rte_rwlock_t rwlock; /* MR cache Lock. */
diff --git a/drivers/common/mlx5/mlx5_common_utils.h
b/drivers/common/mlx5/mlx5_common_utils.h
index 9139bc6829..b93df0b27d 100644
--- a/drivers/common/mlx5/mlx5_common_utils.h
+++ b/drivers/common/mlx5/mlx5_common_utils.h
@@ -27,6 +27,7 @@ struct mlx5_list;
* Structure of the entry in the mlx5 list, user should define its own struct
* that contains this in order to store the data.
*/
+__rte_msvc_pack
struct mlx5_list_entry {
LIST_ENTRY(mlx5_list_entry) next; /* Entry pointers in the list. */
alignas(8) RTE_ATOMIC(uint32_t) ref_cnt; /* 0 means, entry is invalid.
*/
diff --git a/drivers/common/mlx5/mlx5_prm.h b/drivers/common/mlx5/mlx5_prm.h
index 2d82807bc2..72f87ad30b 100644
--- a/drivers/common/mlx5/mlx5_prm.h
+++ b/drivers/common/mlx5/mlx5_prm.h
@@ -323,6 +323,7 @@ enum mlx5_mpw_mode {
};
/* WQE Control segment. */
+__rte_msvc_pack
struct __rte_aligned(MLX5_WSEG_SIZE) mlx5_wqe_cseg {
uint32_t opcode;
uint32_t sq_ds;
@@ -340,10 +341,12 @@ struct __rte_aligned(MLX5_WSEG_SIZE) mlx5_wqe_cseg {
#define WQE_CSEG_WQE_INDEX_OFFSET 8
/* Header of data segment. Minimal size Data Segment */
+__rte_msvc_pack
struct mlx5_wqe_dseg {
uint32_t bcount;
union {
uint8_t inline_data[MLX5_DSEG_MIN_INLINE_SIZE];
+ __rte_msvc_pack
struct {
uint32_t lkey;
uint64_t pbuf;
@@ -352,8 +355,10 @@ struct mlx5_wqe_dseg {
} __rte_packed;
/* Subset of struct WQE Ethernet Segment. */
+__rte_msvc_pack
struct mlx5_wqe_eseg {
union {
+ __rte_msvc_pack
struct {
uint32_t swp_offs;
uint8_t cs_flags;
@@ -366,6 +371,7 @@ struct mlx5_wqe_eseg {
uint16_t vlan_tag;
};
} __rte_packed;
+ __rte_msvc_pack
struct {
uint32_t offsets;
uint32_t flags;
@@ -375,6 +381,7 @@ struct mlx5_wqe_eseg {
};
} __rte_packed;
+__rte_msvc_pack
struct mlx5_wqe_qseg {
uint32_t reserved0;
uint32_t reserved1;
@@ -382,6 +389,7 @@ struct mlx5_wqe_qseg {
uint32_t qpn_cqn;
} __rte_packed;
+__rte_msvc_pack
struct mlx5_wqe_wseg {
uint32_t operation;
uint32_t lkey;
@@ -392,6 +400,7 @@ struct mlx5_wqe_wseg {
} __rte_packed;
/* The title WQEBB, header of WQE. */
+__rte_msvc_pack
struct mlx5_wqe {
union {
struct mlx5_wqe_cseg cseg;
@@ -464,6 +473,7 @@ struct mlx5_cqe {
uint8_t lro_num_seg;
union {
uint8_t user_index_bytes[3];
+ __rte_msvc_pack
struct {
uint8_t user_index_hi;
uint16_t user_index_low;
@@ -487,6 +497,7 @@ struct mlx5_cqe_ts {
uint8_t op_own;
};
+__rte_msvc_pack
struct mlx5_wqe_rseg {
uint64_t raddr;
uint32_t rkey;
@@ -506,6 +517,7 @@ struct mlx5_wqe_rseg {
#define MLX5_UMR_KLM_NUM_ALIGN \
(MLX5_UMR_KLM_PTR_ALIGN / sizeof(struct mlx5_klm))
+__rte_msvc_pack
struct mlx5_wqe_umr_cseg {
uint32_t if_cf_toe_cq_res;
uint32_t ko_to_bs;
@@ -513,6 +525,7 @@ struct mlx5_wqe_umr_cseg {
uint32_t rsvd1[8];
} __rte_packe
[PATCH v5 05/16] common/idpf: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
drivers/common/idpf/base/idpf_osdep.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/common/idpf/base/idpf_osdep.h
b/drivers/common/idpf/base/idpf_osdep.h
index cf9e553906..95449fbf19 100644
--- a/drivers/common/idpf/base/idpf_osdep.h
+++ b/drivers/common/idpf/base/idpf_osdep.h
@@ -180,6 +180,7 @@ static inline uint64_t idpf_read_addr64(volatile void *addr)
#define BITS_PER_BYTE 8
/* memory allocation tracking */
+__rte_msvc_pack
struct idpf_dma_mem {
void *va;
u64 pa;
@@ -187,6 +188,7 @@ struct idpf_dma_mem {
const void *zone;
} __rte_packed;
+__rte_msvc_pack
struct idpf_virt_mem {
void *va;
u32 size;
--
2.47.0.vfs.0.3
[PATCH v5 10/16] net/ice: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
drivers/net/ice/base/ice_osdep.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ice/base/ice_osdep.h b/drivers/net/ice/base/ice_osdep.h
index 5f7347023f..d0280dd2c2 100644
--- a/drivers/net/ice/base/ice_osdep.h
+++ b/drivers/net/ice/base/ice_osdep.h
@@ -187,6 +187,7 @@ do {
\
#define BITS_PER_BYTE 8
/* memory allocation tracking */
+__rte_msvc_pack
struct ice_dma_mem {
void *va;
u64 pa;
@@ -194,6 +195,7 @@ struct ice_dma_mem {
const void *zone;
} __rte_packed;
+__rte_msvc_pack
struct ice_virt_mem {
void *va;
u32 size;
--
2.47.0.vfs.0.3
[PATCH v5 11/16] net/mlx5: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
drivers/net/mlx5/hws/mlx5dr.h | 1 +
drivers/net/mlx5/mlx5.h | 1 +
drivers/net/mlx5/mlx5_flow.h| 4
drivers/net/mlx5/mlx5_hws_cnt.h | 1 +
drivers/net/mlx5/mlx5_utils.h | 4
5 files changed, 11 insertions(+)
diff --git a/drivers/net/mlx5/hws/mlx5dr.h b/drivers/net/mlx5/hws/mlx5dr.h
index 3668ab9fcf..6539af27bb 100644
--- a/drivers/net/mlx5/hws/mlx5dr.h
+++ b/drivers/net/mlx5/hws/mlx5dr.h
@@ -328,6 +328,7 @@ union mlx5dr_crc_encap_entropy_hash_ip_field {
};
};
+__rte_msvc_pack
struct mlx5dr_crc_encap_entropy_hash_fields {
union mlx5dr_crc_encap_entropy_hash_ip_field dst;
union mlx5dr_crc_encap_entropy_hash_ip_field src;
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 89d277b523..1ca52bfe69 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -235,6 +235,7 @@ mlx5_is_probed_port_on_mpesw_device(struct
mlx5_dev_spawn_data *spawn)
}
/** Data associated with socket messages. */
+__rte_msvc_pack
struct mlx5_flow_dump_req {
uint32_t port_id; /**< There are plans in DPDK to extend port_id. */
uint64_t flow_id;
diff --git a/drivers/net/mlx5/mlx5_flow.h b/drivers/net/mlx5/mlx5_flow.h
index bcc2782460..4d40866539 100644
--- a/drivers/net/mlx5/mlx5_flow.h
+++ b/drivers/net/mlx5/mlx5_flow.h
@@ -673,6 +673,7 @@ struct mlx5_flow_dv_tag_resource {
};
/* Modify resource structure */
+__rte_msvc_pack
struct mlx5_flow_dv_modify_hdr_resource {
struct mlx5_list_entry entry;
void *action; /**< Modify header action object. */
@@ -831,6 +832,7 @@ struct mlx5_flow_dv_dest_array_resource {
/** Device flow handle structure for DV mode only. */
+__rte_msvc_pack
struct mlx5_flow_handle_dv {
/* Flow DV api: */
struct mlx5_flow_dv_matcher *matcher; /**< Cache to matcher. */
@@ -849,6 +851,7 @@ struct mlx5_flow_handle_dv {
} __rte_packed;
/** Device flow handle structure: used both for creating & destroying. */
+__rte_msvc_pack
struct mlx5_flow_handle {
SILIST_ENTRY(uint32_t)next;
struct mlx5_vf_vlan vf_vlan; /**< Structure for VF VLAN workaround. */
@@ -1250,6 +1253,7 @@ struct mlx5_flow_attr {
};
/* Flow structure. */
+__rte_msvc_pack
struct rte_flow {
uint32_t dev_handles;
/**< Device flow handles that are part of the flow. */
diff --git a/drivers/net/mlx5/mlx5_hws_cnt.h b/drivers/net/mlx5/mlx5_hws_cnt.h
index d8da9dfcdd..7479da7c5e 100644
--- a/drivers/net/mlx5/mlx5_hws_cnt.h
+++ b/drivers/net/mlx5/mlx5_hws_cnt.h
@@ -133,6 +133,7 @@ enum {
};
/* HWS counter age parameter. */
+__rte_msvc_pack
struct __rte_cache_aligned mlx5_hws_age_param {
RTE_ATOMIC(uint32_t) timeout; /* Aging timeout in seconds (atomically
accessed). */
RTE_ATOMIC(uint32_t) sec_since_last_hit;
diff --git a/drivers/net/mlx5/mlx5_utils.h b/drivers/net/mlx5/mlx5_utils.h
index f933daf03c..081678fa43 100644
--- a/drivers/net/mlx5/mlx5_utils.h
+++ b/drivers/net/mlx5/mlx5_utils.h
@@ -115,6 +115,7 @@ struct mlx5_l3t_level_tbl {
};
/* L3 word entry table data structure. */
+__rte_msvc_pack
struct mlx5_l3t_entry_word {
uint32_t idx; /* Table index. */
uint64_t ref_cnt; /* Table ref_cnt. */
@@ -125,6 +126,7 @@ struct mlx5_l3t_entry_word {
} __rte_packed;
/* L3 double word entry table data structure. */
+__rte_msvc_pack
struct mlx5_l3t_entry_dword {
uint32_t idx; /* Table index. */
uint64_t ref_cnt; /* Table ref_cnt. */
@@ -135,6 +137,7 @@ struct mlx5_l3t_entry_dword {
} __rte_packed;
/* L3 quad word entry table data structure. */
+__rte_msvc_pack
struct mlx5_l3t_entry_qword {
uint32_t idx; /* Table index. */
uint64_t ref_cnt; /* Table ref_cnt. */
@@ -145,6 +148,7 @@ struct mlx5_l3t_entry_qword {
} __rte_packed;
/* L3 pointer entry table data structure. */
+__rte_msvc_pack
struct mlx5_l3t_entry_ptr {
uint32_t idx; /* Table index. */
uint64_t ref_cnt; /* Table ref_cnt. */
--
2.47.0.vfs.0.3
[PATCH v5 04/16] common/iavf: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
drivers/common/iavf/iavf_osdep.h| 2 ++
drivers/common/iavf/virtchnl_inline_ipsec.h | 11 +++
2 files changed, 13 insertions(+)
diff --git a/drivers/common/iavf/iavf_osdep.h b/drivers/common/iavf/iavf_osdep.h
index 1f2b7889cb..8e9d93ccaa 100644
--- a/drivers/common/iavf/iavf_osdep.h
+++ b/drivers/common/iavf/iavf_osdep.h
@@ -158,6 +158,7 @@ do {
\
} while (0)
/* memory allocation tracking */
+__rte_msvc_pack
struct iavf_dma_mem {
void *va;
u64 pa;
@@ -165,6 +166,7 @@ struct iavf_dma_mem {
const void *zone;
} __rte_packed;
+__rte_msvc_pack
struct iavf_virt_mem {
void *va;
u32 size;
diff --git a/drivers/common/iavf/virtchnl_inline_ipsec.h
b/drivers/common/iavf/virtchnl_inline_ipsec.h
index 2f4bf15725..feb0f44381 100644
--- a/drivers/common/iavf/virtchnl_inline_ipsec.h
+++ b/drivers/common/iavf/virtchnl_inline_ipsec.h
@@ -109,6 +109,7 @@ enum inline_ipsec_ops {
};
/* Not all valid, if certain field is invalid, set 1 for all bits */
+__rte_msvc_pack
struct virtchnl_algo_cap {
u32 algo_type;
@@ -132,6 +133,7 @@ struct virtchnl_algo_cap {
} __rte_packed;
/* vf record the capability of crypto from the virtchnl */
+__rte_msvc_pack
struct virtchnl_sym_crypto_cap {
u8 crypto_type;
u8 algo_cap_num;
@@ -142,6 +144,7 @@ struct virtchnl_sym_crypto_cap {
* VF pass virtchnl_ipsec_cap to PF
* and PF return capability of ipsec from virtchnl.
*/
+__rte_msvc_pack
struct virtchnl_ipsec_cap {
/* max number of SA per VF */
u16 max_sa_num;
@@ -172,6 +175,7 @@ struct virtchnl_ipsec_cap {
} __rte_packed;
/* configuration of crypto function */
+__rte_msvc_pack
struct virtchnl_ipsec_crypto_cfg_item {
u8 crypto_type;
@@ -203,6 +207,7 @@ struct virtchnl_ipsec_sym_crypto_cfg {
* PF create SA as configuration and PF driver will return
* an unique index (sa_idx) for the created SA.
*/
+__rte_msvc_pack
struct virtchnl_ipsec_sa_cfg {
/* IPsec SA Protocol - AH/ESP */
u8 virtchnl_protocol_type;
@@ -298,6 +303,7 @@ struct virtchnl_ipsec_sa_cfg {
* VF send configuration of index of SA to PF
* PF will update SA according to configuration
*/
+__rte_msvc_pack
struct virtchnl_ipsec_sa_update {
u32 sa_index; /* SA to update */
u32 esn_hi; /* high 32 bits of esn */
@@ -310,6 +316,7 @@ struct virtchnl_ipsec_sa_update {
* flag bitmap indicate all SA or just selected SA will
* be destroyed
*/
+__rte_msvc_pack
struct virtchnl_ipsec_sa_destroy {
/* All zero bitmap indicates all SA will be destroyed.
* Non-zero bitmap indicates the selected SA in
@@ -325,6 +332,7 @@ struct virtchnl_ipsec_sa_destroy {
* VF send this SA configuration to PF using virtchnl;
* PF read SA and will return configuration for the created SA.
*/
+__rte_msvc_pack
struct virtchnl_ipsec_sa_read {
/* SA valid - invalid/valid */
u8 valid;
@@ -431,6 +439,7 @@ struct virtchnl_ipsec_sa_read {
#define VIRTCHNL_IPSEC_INBOUND_SPD_TBL_IPV6(1)
/* Add allowlist entry in IES */
+__rte_msvc_pack
struct virtchnl_ipsec_sp_cfg {
u32 spi;
u32 dip[4];
@@ -459,6 +468,7 @@ struct virtchnl_ipsec_sp_cfg {
/* Delete allowlist entry in IES */
+__rte_msvc_pack
struct virtchnl_ipsec_sp_destroy {
/* 0 for IPv4 table, 1 for IPv6 table. */
u8 table_id;
@@ -494,6 +504,7 @@ struct virtchnl_ipsec_resp {
};
/* Internal message descriptor for VF <-> IPsec communication */
+__rte_msvc_pack
struct inline_ipsec_msg {
u16 ipsec_opcode;
u16 req_id;
--
2.47.0.vfs.0.3
[PATCH v5 01/16] eal: provide pack start macro for MSVC
From: Tyler Retzlaff MSVC struct packing is not compatible with GCC. Provide a macro that can be used to push existing pack value and sets packing to 1-byte. The existing __rte_packed macro is then used to restore the pack value prior to the push. Instead of providing macros exclusively for MSVC and for GCC the existing macro is deliberately utilized to trigger a warning if no existing packing has been pushed allowing easy identification of locations where the __rte_msvc_pack is missing. Signed-off-by: Tyler Retzlaff --- lib/eal/include/rte_common.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/lib/eal/include/rte_common.h b/lib/eal/include/rte_common.h index 4d299f2b36..409890863e 100644 --- a/lib/eal/include/rte_common.h +++ b/lib/eal/include/rte_common.h @@ -103,8 +103,10 @@ typedef uint16_t unaligned_uint16_t; * Force a structure to be packed */ #ifdef RTE_TOOLCHAIN_MSVC -#define __rte_packed +#define __rte_msvc_pack __pragma(pack(push, 1)) +#define __rte_packed __pragma(pack(pop)) #else +#define __rte_msvc_pack #define __rte_packed __attribute__((__packed__)) #endif -- 2.47.0.vfs.0.3
[PATCH v5 09/16] net/iavf: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Remove __rte_packed where it appears natural alignment without
packing would produce the same layout or the struct is internal.
Signed-off-by: Tyler Retzlaff
---
drivers/net/iavf/iavf_ipsec_crypto.h | 4 +++-
drivers/net/iavf/iavf_rxtx.c | 2 +-
2 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/iavf/iavf_ipsec_crypto.h
b/drivers/net/iavf/iavf_ipsec_crypto.h
index 49f9202aca..0637ce21d6 100644
--- a/drivers/net/iavf/iavf_ipsec_crypto.h
+++ b/drivers/net/iavf/iavf_ipsec_crypto.h
@@ -11,12 +11,14 @@
+__rte_msvc_pack
struct iavf_tx_ipsec_desc {
union {
struct {
__le64 qw0;
__le64 qw1;
};
+ __rte_msvc_pack
struct {
__le16 l4payload_length;
__le32 esn;
@@ -97,7 +99,7 @@ struct iavf_ipsec_crypto_pkt_metadata {
uint8_t next_proto; /* Next Protocol (8b/1B) */
uint32_t esn; /* Extended Sequence Number (32b/4B) */
-} __rte_packed;
+};
/**
* Inline IPsec Crypto offload is supported
diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c
index 4850b9e381..7efbee7c3e 100644
--- a/drivers/net/iavf/iavf_rxtx.c
+++ b/drivers/net/iavf/iavf_rxtx.c
@@ -53,7 +53,7 @@ static uint16_t geneve_udp_port = RTE_GENEVE_DEFAULT_PORT;
struct simple_gre_hdr {
uint16_t flags;
uint16_t proto;
-} __rte_packed;
+};
/* structure that caches offload info for the current packet */
struct offload_info {
--
2.47.0.vfs.0.3
[PATCH v5 08/16] net/i40e: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
drivers/net/i40e/base/i40e_osdep.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/i40e/base/i40e_osdep.h
b/drivers/net/i40e/base/i40e_osdep.h
index 8960206d8b..af8756b7d8 100644
--- a/drivers/net/i40e/base/i40e_osdep.h
+++ b/drivers/net/i40e/base/i40e_osdep.h
@@ -177,6 +177,7 @@ static inline uint64_t i40e_read64_addr(volatile void *addr)
#define ARRAY_SIZE(arr) RTE_DIM(arr)
/* memory allocation tracking */
+__rte_msvc_pack
struct i40e_dma_mem {
void *va;
u64 pa;
@@ -188,6 +189,7 @@ struct i40e_dma_mem {
i40e_allocate_dma_mem_d(h, m, s, a)
#define i40e_free_dma_mem(h, m) i40e_free_dma_mem_d(h, m)
+__rte_msvc_pack
struct i40e_virt_mem {
void *va;
u32 size;
--
2.47.0.vfs.0.3
[PATCH v5 16/16] crypto/mlx5: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
drivers/crypto/mlx5/mlx5_crypto.h | 2 ++
drivers/crypto/mlx5/mlx5_crypto_gcm.c | 1 +
2 files changed, 3 insertions(+)
diff --git a/drivers/crypto/mlx5/mlx5_crypto.h
b/drivers/crypto/mlx5/mlx5_crypto.h
index 547bb490e2..80a7a38f17 100644
--- a/drivers/crypto/mlx5/mlx5_crypto.h
+++ b/drivers/crypto/mlx5/mlx5_crypto.h
@@ -32,6 +32,7 @@ enum mlx5_crypto_mode {
MLX5_CRYPTO_IPSEC_OPT,
};
+__rte_msvc_pack
struct mlx5_crypto_ipsec_mem {
uint8_t mem[MLX5_CRYPTO_GCM_IPSEC_IV_SIZE];
} __rte_packed;
@@ -108,6 +109,7 @@ struct mlx5_crypto_devarg_params {
enum mlx5_crypto_mode crypto_mode;
};
+__rte_msvc_pack
struct mlx5_crypto_session {
union {
/**< AES-XTS configuration. */
diff --git a/drivers/crypto/mlx5/mlx5_crypto_gcm.c
b/drivers/crypto/mlx5/mlx5_crypto_gcm.c
index cd21605bd2..005be4d4bf 100644
--- a/drivers/crypto/mlx5/mlx5_crypto_gcm.c
+++ b/drivers/crypto/mlx5/mlx5_crypto_gcm.c
@@ -62,6 +62,7 @@ struct mlx5_crypto_gcm_data {
uint32_t dst_mkey;
};
+__rte_msvc_pack
struct mlx5_crypto_gcm_tag_cpy_info {
void *digest;
uint8_t tag_len;
--
2.47.0.vfs.0.3
[PATCH v5 02/16] eal: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Remove __rte_packed where it appears natural alignment without
packing would produce the same layout or the struct is internal.
Signed-off-by: Tyler Retzlaff
---
lib/eal/common/eal_private.h | 2 +-
lib/eal/include/rte_memory.h | 1 +
lib/eal/include/rte_memzone.h | 1 +
lib/eal/include/rte_trace_point.h | 2 +-
lib/eal/x86/include/rte_memcpy.h | 3 +++
5 files changed, 7 insertions(+), 2 deletions(-)
diff --git a/lib/eal/common/eal_private.h b/lib/eal/common/eal_private.h
index bb315dab04..3b94e01b5b 100644
--- a/lib/eal/common/eal_private.h
+++ b/lib/eal/common/eal_private.h
@@ -62,7 +62,7 @@ struct rte_config {
* DPDK instances
*/
struct rte_mem_config *mem_config;
-} __rte_packed;
+};
/**
* Get the global configuration structure.
diff --git a/lib/eal/include/rte_memory.h b/lib/eal/include/rte_memory.h
index dbd0a6bedc..fffbd2ed65 100644
--- a/lib/eal/include/rte_memory.h
+++ b/lib/eal/include/rte_memory.h
@@ -46,6 +46,7 @@ extern "C" {
/**
* Physical memory segment descriptor.
*/
+__rte_msvc_pack
struct rte_memseg {
rte_iova_t iova;/**< Start IO address. */
union {
diff --git a/lib/eal/include/rte_memzone.h b/lib/eal/include/rte_memzone.h
index e1563994d5..e52bdc91b8 100644
--- a/lib/eal/include/rte_memzone.h
+++ b/lib/eal/include/rte_memzone.h
@@ -45,6 +45,7 @@ extern "C" {
* A structure describing a memzone, which is a contiguous portion of
* physical memory identified by a name.
*/
+__rte_msvc_pack
struct rte_memzone {
#define RTE_MEMZONE_NAMESIZE 32 /**< Maximum length of memory zone
name.*/
diff --git a/lib/eal/include/rte_trace_point.h
b/lib/eal/include/rte_trace_point.h
index 87b5f43c3c..b24db3b6da 100644
--- a/lib/eal/include/rte_trace_point.h
+++ b/lib/eal/include/rte_trace_point.h
@@ -298,7 +298,7 @@ struct __rte_trace_stream_header {
rte_uuid_t uuid;
uint32_t lcore_id;
char thread_name[__RTE_TRACE_EMIT_STRING_LEN_MAX];
-} __rte_packed;
+};
struct __rte_trace_header {
uint32_t offset;
diff --git a/lib/eal/x86/include/rte_memcpy.h b/lib/eal/x86/include/rte_memcpy.h
index 42058e4a3f..fed715ca33 100644
--- a/lib/eal/x86/include/rte_memcpy.h
+++ b/lib/eal/x86/include/rte_memcpy.h
@@ -67,12 +67,15 @@ rte_mov15_or_less(void *dst, const void *src, size_t n)
* Use the following structs to avoid violating C standard
* alignment requirements and to avoid strict aliasing bugs
*/
+ __rte_msvc_pack
struct rte_uint64_alias {
uint64_t val;
} __rte_packed __rte_may_alias;
+ __rte_msvc_pack
struct rte_uint32_alias {
uint32_t val;
} __rte_packed __rte_may_alias;
+ __rte_msvc_pack
struct rte_uint16_alias {
uint16_t val;
} __rte_packed __rte_may_alias;
--
2.47.0.vfs.0.3
[PATCH v5 00/16] fix packing of structs when building with MSVC
From: Andre Muezerie MSVC struct packing is not compatible with GCC provide a macro that can be used to push existing pack value and sets packing to 1-byte. The existing __rte_packed macro is then used to restore the pack value prior to the push. Instead of providing macros exclusively for MSVC and for GCC the existing macro is deliberately utilized to trigger a warning if no existing packing has been pushed allowing easy identification of locations where the __rte_msvc_pack is missing. I've decided to only add the macro to packed structs that are built for Windows. It seems there is little value in adding the macro tree wide and if new code arrives that is built on Windows the __rte_packed will flag where the preamble macro is required. If as a part of review maintainers identify structs they believe don't require packing so long as they are explicitly identified i'll remove the __rte_packed as a part of this series. v5: * rebase on top of latest main v4: * add another missing __rte_msvc_pack to crypto/mlx5 patch * correct commit message for duplicated packing packing in crypto/mlx5 patch v3: * add missing __rte_msvc_pack to crypto/mlx5 * fix commit messages to reference __rte_msvc_pack macro instead of __rte_msvc_pushpack(1) v2: * app/testpmd, remove packing from simple_gre_hdr * net/iavf, remove packing from iavf_ipsec_crypto_pkt_metadata, simple_gre_hdr * examples, remove packing from pkt_key_qinq, pkt_key_ipv4_5tuple, pkt_key_ipv6_5tuple, pkt_key_ipv4_addr, pkt_key_ipv6_addr * eal, remove packing from rte_config, __rte_trace_stream_header Tyler Retzlaff (16): eal: provide pack start macro for MSVC eal: pack structures when building with MSVC net: pack structures when building with MSVC common/iavf: pack structures when building with MSVC common/idpf: pack structures when building with MSVC common/mlx5: pack structures when building with MSVC dma/ioat: pack structures when building with MSVC net/i40e: pack structures when building with MSVC net/iavf: pack structures when building with MSVC net/ice: pack structures when building with MSVC net/mlx5: pack structures when building with MSVC net/octeon_ep: pack structures when building with MSVC app/testpmd: pack structures when building with MSVC app/test: pack structures when building with MSVC examples: pack structures when building with MSVC crypto/mlx5: pack structures when building with MSVC app/test-pmd/csumonly.c | 2 +- app/test/test_efd.c | 1 + app/test/test_hash.c| 1 + app/test/test_member.c | 1 + drivers/common/iavf/iavf_osdep.h| 2 ++ drivers/common/iavf/virtchnl_inline_ipsec.h | 11 drivers/common/idpf/base/idpf_osdep.h | 2 ++ drivers/common/mlx5/mlx5_common_mr.h| 4 +++ drivers/common/mlx5/mlx5_common_utils.h | 1 + drivers/common/mlx5/mlx5_prm.h | 30 + drivers/crypto/mlx5/mlx5_crypto.h | 2 ++ drivers/crypto/mlx5/mlx5_crypto_gcm.c | 1 + drivers/dma/ioat/ioat_hw_defs.h | 1 + drivers/net/i40e/base/i40e_osdep.h | 2 ++ drivers/net/iavf/iavf_ipsec_crypto.h| 4 ++- drivers/net/iavf/iavf_rxtx.c| 2 +- drivers/net/ice/base/ice_osdep.h| 2 ++ drivers/net/mlx5/hws/mlx5dr.h | 1 + drivers/net/mlx5/mlx5.h | 1 + drivers/net/mlx5/mlx5_flow.h| 4 +++ drivers/net/mlx5/mlx5_hws_cnt.h | 1 + drivers/net/mlx5/mlx5_utils.h | 4 +++ drivers/net/octeon_ep/otx_ep_mbox.h | 1 + examples/common/neon/port_group.h | 1 + examples/ip_pipeline/cli.c | 10 +++ examples/ipsec-secgw/ipsec.h| 1 + examples/l3fwd-power/main.c | 4 ++- examples/l3fwd/l3fwd_route.h| 2 ++ examples/ptpclient/ptpclient.c | 8 ++ examples/vhost_blk/blk_spec.h | 1 + lib/eal/common/eal_private.h| 2 +- lib/eal/include/rte_common.h| 4 ++- lib/eal/include/rte_memory.h| 1 + lib/eal/include/rte_memzone.h | 1 + lib/eal/include/rte_trace_point.h | 2 +- lib/eal/x86/include/rte_memcpy.h| 3 +++ lib/net/rte_arp.h | 2 ++ lib/net/rte_dtls.h | 1 + lib/net/rte_esp.h | 2 ++ lib/net/rte_geneve.h| 1 + lib/net/rte_gre.h | 4 +++ lib/net/rte_gtp.h | 5 lib/net/rte_ib.h| 1 + lib/net/rte_icmp.h | 3 +++ lib/net/rte_ip4.h | 1 + lib/net/rte_ip6.h | 4 +++ lib/net/rte_l2tpv2.h| 4 +++ lib/net/rte_m
[PATCH v5 07/16] dma/ioat: pack structures when building with MSVC
From: Tyler Retzlaff
Add __rte_msvc_pack to all __rte_packed structs to cause packing
when building with MSVC.
Signed-off-by: Tyler Retzlaff
---
drivers/dma/ioat/ioat_hw_defs.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/dma/ioat/ioat_hw_defs.h b/drivers/dma/ioat/ioat_hw_defs.h
index 11893951f2..b67e252d6c 100644
--- a/drivers/dma/ioat/ioat_hw_defs.h
+++ b/drivers/dma/ioat/ioat_hw_defs.h
@@ -52,6 +52,7 @@ extern "C" {
#defineIOAT_DMACAP_PQ (1 << 9)
#defineIOAT_DMACAP_DMA_DIF (1 << 10)
+__rte_msvc_pack
struct ioat_registers {
uint8_t chancnt;
uint8_t xfercap;
--
2.47.0.vfs.0.3
[PATCH v16 3/4] pmu: support reading Intel x86_64 PMU events in runtime
Add support for reading Intel x86_64 PMU events in runtime.
Signed-off-by: Tomasz Duszynski
---
app/test/test_pmu.c | 2 ++
lib/pmu/meson.build | 1 +
lib/pmu/rte_pmu.h| 2 ++
lib/pmu/rte_pmu_pmc_x86_64.h | 24
4 files changed, 29 insertions(+)
create mode 100644 lib/pmu/rte_pmu_pmc_x86_64.h
diff --git a/app/test/test_pmu.c b/app/test/test_pmu.c
index 0dc33eeccc..c6e798f52d 100644
--- a/app/test/test_pmu.c
+++ b/app/test/test_pmu.c
@@ -15,6 +15,8 @@ test_pmu_read(void)
#if defined(RTE_ARCH_ARM64)
name = "cpu_cycles";
+#elif defined(RTE_ARCH_X86_64)
+ name = "cpu-cycles";
#endif
if (name == NULL) {
diff --git a/lib/pmu/meson.build b/lib/pmu/meson.build
index 94619c63c2..5230ca3e69 100644
--- a/lib/pmu/meson.build
+++ b/lib/pmu/meson.build
@@ -12,6 +12,7 @@ sources = files('rte_pmu.c')
indirect_headers += files(
'rte_pmu_pmc_arm64.h',
+'rte_pmu_pmc_x86_64.h',
)
if dpdk_conf.has('RTE_ARCH_ARM64')
diff --git a/lib/pmu/rte_pmu.h b/lib/pmu/rte_pmu.h
index 6e073f11ea..e96655e4d2 100644
--- a/lib/pmu/rte_pmu.h
+++ b/lib/pmu/rte_pmu.h
@@ -39,6 +39,8 @@ extern "C" {
#if defined(RTE_ARCH_ARM64)
#include "rte_pmu_pmc_arm64.h"
+#elif defined(RTE_ARCH_X86_64)
+#include "rte_pmu_pmc_x86_64.h"
#endif
/** Maximum number of events in a group */
diff --git a/lib/pmu/rte_pmu_pmc_x86_64.h b/lib/pmu/rte_pmu_pmc_x86_64.h
new file mode 100644
index 00..5eb8f73c11
--- /dev/null
+++ b/lib/pmu/rte_pmu_pmc_x86_64.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2024 Marvell.
+ */
+#ifndef _RTE_PMU_PMC_X86_64_H_
+#define _RTE_PMU_PMC_X86_64_H_
+
+#include
+
+static __rte_always_inline uint64_t
+rte_pmu_pmc_read(int index)
+{
+ uint64_t low, high;
+
+ asm volatile(
+ "rdpmc\n"
+ : "=a" (low), "=d" (high)
+ : "c" (index)
+ );
+
+ return low | (high << 32);
+}
+#define rte_pmu_pmc_read rte_pmu_pmc_read
+
+#endif /* _RTE_PMU_PMC_X86_64_H_ */
--
2.34.1
RE: [PATCH v2 1/2] app/testpmd: fix flow update
> -Original Message-
> From: Danylo Vodopianov
> Sent: Monday, November 18, 2024 12:26
> To: NBU-Contact-Thomas Monjalon (EXTERNAL) ;
> aman.deep.si...@intel.com; yuying.zh...@intel.com; Ori Kam
> ; mko-...@napatech.com; c...@napatech.com; Dariusz
> Sosnowski ; sil-...@napatech.com
> Cc: Gregory Etelson ; Alexander Kozyrev
> ; dev@dpdk.org; sta...@dpdk.org;
> ferruh.yi...@amd.com
> Subject: [PATCH v2 1/2] app/testpmd: fix flow update
>
> If actions provided to “flow update…“ command contained an age action, then
> testpmd did not update the age action context accordingly.
>
> Thus "flow aged destroy" command can not execute successfully.
>
> Fix was done with next steps
> 1. Generate new port flow entry to add/replace action(s).
> 2. Set age context if age action is present.
> 3. Replace flow in the flow list.
>
> Fixes: 2d9c7e56e52c ("app/testpmd: support updating flow rule actions")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Danylo Vodopianov
> ---
> app/test-pmd/config.c | 21 -
> 1 file changed, 20 insertions(+), 1 deletion(-)
>
> diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c index
> 88770b4dfc..c831166431 100644
> --- a/app/test-pmd/config.c
> +++ b/app/test-pmd/config.c
> @@ -3882,7 +3882,8 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
> const struct rte_flow_action *actions, bool is_user_id) {
> struct rte_port *port;
> - struct port_flow **flow_list;
> + struct port_flow **flow_list, *uf;
> + struct rte_flow_action_age *age = age_action_get(actions);
>
> if (port_id_is_invalid(port_id, ENABLED_WARN) ||
> port_id == (portid_t)RTE_PORT_ALL) @@ -3897,6 +3898,16 @@
> port_flow_update(portid_t port_id, uint32_t rule_id,
> flow_list = &flow->next;
> continue;
> }
> +
> + /* Update flow action(s) with new action(s) */
> + uf = port_flow_new(flow->rule.attr_ro, flow->rule.pattern_ro,
> actions,
> &error);
> + if (!uf)
> + return port_flow_complain(&error);
> + if (age) {
> + flow->age_type = ACTION_AGE_CONTEXT_TYPE_FLOW;
> + age->context = &flow->age_type;
> + }
> +
> /*
> * Poisoning to make sure PMDs update it in case
> * of error.
> @@ -3913,6 +3924,14 @@ port_flow_update(portid_t port_id, uint32_t
> rule_id,
> printf("Flow rule #%"PRIu64
>" updated with new actions\n",
>flow->id);
> +
> + uf->next = flow->next;
> + uf->id = flow->id;
> + uf->user_id = flow->user_id;
> + uf->flow = flow->flow;
> + *flow_list = uf;
> +
> + free(flow);
> return 0;
> }
> printf("Failed to find flow %"PRIu32"\n", rule_id);
> --
> 2.43.5
Acked-by: Dariusz Sosnowski
Best regards,
Dariusz Sosnowski
RE: [PATCH 2/3] regex/cn9k: fix build warnings on ubuntu 24.04
> -Original Message-
> From: Jerin Jacob
> Sent: Monday, November 18, 2024 2:11 PM
> To: Sunil Kumar Kori ; Liron Himi
> Cc: dev@dpdk.org; alia...@nvidia.com; Sunil Kumar Kori
> ; sta...@dpdk.org
> Subject: RE: [PATCH 2/3] regex/cn9k: fix build warnings on ubuntu 24.04
>
>
>
> > -Original Message-
> > From: sk...@marvell.com
> > Sent: Thursday, November 14, 2024 1:08 PM
> > To: Liron Himi ; Jerin Jacob
> > Cc: dev@dpdk.org; alia...@nvidia.com; Sunil Kumar Kori
> > ; sta...@dpdk.org
> > Subject: [PATCH 2/3] regex/cn9k: fix build warnings on ubuntu 24.04
> >
> > From: Sunil Kumar Kori
> >
> > Bugzilla ID: 1513
> > Fixes: 72c00ae9dba7 ("regex/cn9k: use cnxk infrastructure")
> > Cc: sta...@dpdk.org
> >
> > Signed-off-by: Sunil Kumar Kori
> > ---
> > drivers/regex/cn9k/cn9k_regexdev.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/regex/cn9k/cn9k_regexdev.c
> > b/drivers/regex/cn9k/cn9k_regexdev.c
> > index aa809ab5bf..e635980183 100644
> > --- a/drivers/regex/cn9k/cn9k_regexdev.c
> > +++ b/drivers/regex/cn9k/cn9k_regexdev.c
> > @@ -271,7 +271,7 @@ static inline int
> > ree_enqueue(struct roc_ree_qp *qp, struct rte_regex_ops *op,
> > struct roc_ree_pending_queue *pend_q) {
> > - union roc_ree_inst inst;
> > + union roc_ree_inst inst = {0};
>
> This changes are in fastpath and it is costly. Please fix with other means or
> suppress the warning if it not relevant.
> I skipped this patch for the merged and 1/3, 3/3 patches are pplied to dpdk-
> next-net-mrvl/for-main. Thanks
Ack.
>
> > union ree_res *res;
> > uint32_t offset;
> >
> > --
> > 2.43.0
[PATCH v2] doc: add security document
This is a new document covering security protocols implemented in DPDK. Signed-off-by: Nandini Persad Signed-off-by: Thomas Monjalon Reviewed-by: Stephen Hemminger --- V2 - incorporate review feedback doc/guides/index.rst | 1 + doc/guides/security/security.rst | 337 +++ 2 files changed, 338 insertions(+) create mode 100644 doc/guides/security/security.rst diff --git a/doc/guides/index.rst b/doc/guides/index.rst index 244b99624c..b8fddc56ae 100644 --- a/doc/guides/index.rst +++ b/doc/guides/index.rst @@ -13,6 +13,7 @@ DPDK documentation sample_app_ug/index prog_guide/index howto/index + security/index tools/index testpmd_app_ug/index nics/index diff --git a/doc/guides/security/security.rst b/doc/guides/security/security.rst new file mode 100644 index 00..ab2dfa4a4a --- /dev/null +++ b/doc/guides/security/security.rst @@ -0,0 +1,337 @@ +.. SPDX-License-Identifier: BSD-3-Clause + +Security Support Guide +== + +This document describes the security features available in the DPDK. +This guide will provides information on each protocol, +including supported algorithms, practical implementation details, and references. + +By detailing the supported algorithms and providing insights into each +security protocol, this document serves as a resource for anyone looking +to implement or enhance security measures within their DPDK-based environments. + + + +Related Documentation +- + +Here is a list of related documents that provide detail of each library, +its capabilities and what level of support it currently has within DPDK. + +* :doc:`Crypto Device Drivers <../cryptodevs/index>` + This section contains information about all the crypto drivers in DPDK, + such as feature support availability, cipher algorithms and authentication + algorithms. + +* :doc:`Security Library <../prog_guide/rte_security>` + This library is the glue between ethdev and crypto dev. It includes low-level supported protocols such as MACsec, TLS, IPSec, and PDCP. + +* Protocols: These include two supported protocols in DPDK. + * :doc:`IPSec Library <../prog_guide/ipsec_lib>` + * :doc:`PDCP Library <../prog_guide/pdcp_lib>` + + +Protocols +- + + +MACSec +~~ + +MACsec (accelerated by Marvell) is a network security standard that operates +at the medium access control layer and defines connectionless data confidentiality +and integrity for media access independent protocols. It is standardized by the +IEEE 802.1 working group. + + +**Wikipedia Link** +* https://en.wikipedia.org/wiki/IEEE_802.1AE + +**Standard Link** +* https://1.ieee802.org/security/802-1ae/ + +**Level of Support in DPDK** +* Supported in DPDK + Sample Application :doc:`MACSec Sample Application <../sample_app_ug/l2_forward_macsec.html>>` + +**Supported Algorithms** +* As specified by MACsec specification: AES-128-GCM, AES-256-GCM + +**Drivers** +* Marvell cnxk Ethernet PMD which supports inline MACsec + +**Facts** +* Uses the AES-GCM cryptography algorithm +* Works on layer 2 and protects all DHCP and ARP traffic +* Each MAC frame has a separate integrity verification code +* Prevents attackers from resending copied MAC frames into the network without being detected +* Commonly used in environments where securing Ethernet traffic between devices is critical, such as in enterprise networks, data centers and service provider networks +* Applications do not need modification to work with IPsec + +**Cons** +* Only operates at Layer 2, so it doesn't protect traffic beyond the local Ethernet segment or over Layer 3 networks or the internet +* Data is decrypted and re-encrypted at each network device, +which could expose data at each point +* Can't detect rogue devices that operate on Layer 1 +* Relies on hardware for encryption and decryption, so not all network devices can use it + + +IPSec +~ + +IPsec (accelerated by Intel, Marvell, Netronome, NXP) allows secure communication +over the internet by encrypting data traffic between two or more devices or networks. +IPsec works on a different layer than MACsec, at layer 3. + +**Wikipedia Link** +* https://en.wikipedia.org/wiki/IPsec + +**Standard Link** +* https://datatracker.ietf.org/wg/ipsec/about/ + +**Level of Support in DPDK** +* Supported +* High-level library and sample application +* :doc:`IPSec Library <../prog_guide/ipsec_lib>` +* :doc:`IPSec Sample Application <../sample_app_ug/ipsec_secgw>` + +**Supported Algorithms** +* AES-GCM and ChaCha20-Poly1305 +* AES CBC and AES-CTR +* HMAC-SHA1/SHA2 for integrity protection and authenticity + +**Pros** +* Uses public keys to create an encrypted, authenticated tunnel to resources +* Offers strong security, scalability, and inter
Re: [PATCH v2 3/6] dts: Self-Discovering Architecture Change
Here again I am on the same level as Juraj. Repeating my previous
comments on commit subject and Bugzilla ID.
Moreover, the subject should be of imperative form according to the
contributing guidelines. In other words, the first word is always an
imperative verb. Something like this could work:
dts: enable self-discovering architecture
About the commit body indicating uncertainty ("I believe..."), there
shouldn't be space for uncertainty on the tree, but just backed-up
facts. If you are uncertain about something, make sure to find out if
the statement stands true and provide the reasoning around it.
You can find out if the call is a standard by looking up the manual[1].
In this case it doesn't mention anything except of being related to the
syscall[2], which is indeed backed up by the POSIX.1 standard. But you
can attempt to compare it to any other POSIX OS and draw some
conclusions. For example FreeBSD's manual[3], explicitly states that
their command is conform to the POSIX.2 standard. And you'll notice that
FreeBSD's has more options than Linux's. The conclusion I can gather
here is that Linux's version is not entirely conform and implements only
a subset. For our usage, this is good enough as it still falls under the
POSIX umbrella. Therefore, you can change your paragraph into something
like:
The POSIX-compliant options of the command `uname` are used
to determine the system architecture.
Hope this helps!
Best,
Luca
[1] https://man7.org/linux/man-pages/man1/uname.1.html
[2] https://man7.org/linux/man-pages/man2/uname.2.html
[3] https://man.freebsd.org/cgi/man.cgi?uname
Re: [PATCH v2 4/6] dts: Rework DPDK Attributes In SUT Node Config
Good one. My only comment – aside the usual ones – on top of Juraj's is could try to make the commit subject more concise. For example: dts: isolate config of dpdk arguments
Re: [PATCH v2 5/6] dts: add conditional behavior for test suite
The issue you mention should be resolved now with the Pydantic changes. Do we still have reason to make this change? I guess we could make the test suites in the config file optional, in which case you can just add a default to the field and remove the constraint.
Re: [PATCH v2 6/6] doc: dpdk documentation changes for new dts config
As Juraj already said this is no longer needed.
RE: bug in cryptodev enqueue/dequeue callbacks?
> > On Nov 14, 2024, at 8:41 AM, Konstantin Ananyev
> > wrote:
> >
> > Hi everyone,
> >
> > Looking at implementation of cryptodev callbacks
> > (it uses DPDK RCU), it seems like there is a bug here:
> >
> > at init time we don't call rte_rcu_qsbr_thread_register().
> > As I understand without it rte_rcu_qsbr_check() wouldn't
> > work properly for that thread.
> Yes, this understanding is correct. However, the responsibility of calling
> the rte_rcu_qsbr_thread_register lies with the application.
> The roles and responsibilities are documented at [1]
>
> [1]
> https://doc.dpdk.org/guides/prog_guide/rcu_lib.html#resource-reclamation-framework-for-dpdk
I don't think it is an option here.
cryptodev callbacks adding/removing/invoking mechanism it totally opaque to the
user.
So these RCU instances are internal for cryptodev devices and shouldn't be
updated for the user.
In fact, it is sort of 'speacial-case' where we have rcu with just one
thread_id(==0),
and doing online/offline only.
Here is a link to original discussion:
https://mails.dpdk.org/archives/dev/2020-September/183401.html
So I still think we just forgot to call
rte_rcu_qsbr_thread_register(qsbr, 0);
here.
> >
> > Probably need to add:
> > static int
> > cryptodev_cb_init(struct rte_cryptodev *dev)
> > {
> >
> > if (rte_rcu_qsbr_init(qsbr, max_threads)) {...}
> > + rte_rcu_qsbr_thread_register(qsbr, 0);
> >
> > Unless I am missing something obvious here?
> > Konstantin
> >
Re: [PATCH v2] dts: Testbed And Node Configuration Split
This is a good change, but I'd like to review it once you adapt it to the Pydantic-based config, as it may look quite different. While we are splitting things I suggest to create two new source files under `config` for each sections: conf/ __init__.py # where Configurations stays and brings everything together test_run.py # to hold the `test_run` and all the related models node.py # as above but for `node` This will hopefully make it easier to read. Best, Luca
[PATCH v2 2/2] app/testpmd: fix aged flow destroy
Avoid removal of additional flows after requested number of flows has
been already removed.
port_flow_destroy() function goes through all flows and compares
given flow ‘id’ with them. However in some cases it can advance pointer
with “given ID” and thus remove additional flow.
port_flow_destroy() function never assumed that rule array can be freed
when it's executing, and port_flow_aged() just violated that assumption.
Fixes: de956d5ecf08 ("app/testpmd: support age shared action context")
Cc: sta...@dpdk.org
Signed-off-by: Danylo Vodopianov
---
app/test-pmd/config.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index c831166431..04de2fe59d 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -4160,8 +4160,9 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
}
type = (enum age_action_context_type *)contexts[idx];
switch (*type) {
- case ACTION_AGE_CONTEXT_TYPE_FLOW:
+ case ACTION_AGE_CONTEXT_TYPE_FLOW: {
ctx.pf = container_of(type, struct port_flow, age_type);
+ uint64_t flow_id = ctx.pf->id;
printf("%-20s\t%" PRIu64 "\t%" PRIu32 "\t%" PRIu32
"\t%c%c%c\t\n",
"Flow",
@@ -4172,9 +4173,10 @@ port_flow_aged(portid_t port_id, uint8_t destroy)
ctx.pf->rule.attr->egress ? 'e' : '-',
ctx.pf->rule.attr->transfer ? 't' : '-');
if (destroy && !port_flow_destroy(port_id, 1,
- &ctx.pf->id, false))
+ &flow_id, false))
total++;
break;
+ }
case ACTION_AGE_CONTEXT_TYPE_INDIRECT_ACTION:
ctx.pia = container_of(type,
struct port_indirect_action, age_type);
--
2.43.5
[PATCH 1/3] net/bnxt: update HWRM API
Update HWRM API to select ring profile.
Signed-off-by: Ajit Khaparde
---
drivers/net/bnxt/hsi_struct_def_dpdk.h | 71 +-
1 file changed, 69 insertions(+), 2 deletions(-)
diff --git a/drivers/net/bnxt/hsi_struct_def_dpdk.h
b/drivers/net/bnxt/hsi_struct_def_dpdk.h
index 8f348c20fb..737bf2693b 100644
--- a/drivers/net/bnxt/hsi_struct_def_dpdk.h
+++ b/drivers/net/bnxt/hsi_struct_def_dpdk.h
@@ -15617,7 +15617,46 @@ struct hwrm_func_qcaps_output {
* (SR-IOV) disabled or on a VF.
*/
uint32_troce_vf_max_gid;
- uint8_t unused_3[3];
+ uint32_tflags_ext3;
+ /*
+* When this bit is '1', firmware supports the driver using
+* FUNC_CFG (or FUNC_VF_CFG) to decrease resource reservations
+* while some resources are still allocated. An error is returned
+* if the driver tries to set the reservation to be less than the
+* number of allocated resources.
+*/
+ #define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_RM_RSV_WHILE_ALLOC_CAP \
+ UINT32_C(0x1)
+ /*
+* When this bit is '1', the PF requires an L2 filter to be
+* allocated by the driver using HWRM_CFA_L2_FILTER_ALLOC after
+* bringing the interface up, before traffic is sent.
+*/
+ #define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_REQUIRE_L2_FILTER \
+ UINT32_C(0x2)
+ /*
+* When set to 1, indicates the field max_roce_vfs in the structure
+* is valid. If this bit is 0, the driver should not use the
+* 'max_roce_vfs' field.
+*/
+ #define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_MAX_ROCE_VFS_SUPPORTED \
+ UINT32_C(0x4)
+ /*
+* When set to 1, indicates the field 'rx_rate_profile_sel' in
+* RING_ALLOC can specify a valid RX rate profile when allocating
+* RX or RX aggregation rings. If this bit is 0, the driver
+* should not use the 'rx_rate_profile_sel' field.
+*/
+ #define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT3_RX_RATE_PROFILE_SEL_SUPPORTED
\
+ UINT32_C(0x8)
+ /*
+* The number of VFs that can be used for RoCE on the function. If less
+* than max_vfs, roce vfs will be assigned to the first VF of the
+* function and be contiguous.
+* This is valid only on the PF with SR-IOV and RDMA enabled.
+*/
+ uint16_tmax_roce_vfs;
+ uint8_t unused_3[5];
/*
* This field is used in Output records to indicate that the output
* is completely written to RAM. This field should be read as '1'
@@ -45026,6 +45065,14 @@ struct hwrm_ring_alloc_input {
*/
#define HWRM_RING_ALLOC_INPUT_ENABLES_STEERING_TAG_VALID \
UINT32_C(0x800)
+ /*
+* This bit must be '1' for the rx_rate_profile_sel field to
+* be configured. This should only be used when
+* 'rx_rate_profile_sel_supported' bit is set in flags_ext3
+* field of FUNC_QCAPS response.
+*/
+ #define HWRM_RING_ALLOC_INPUT_ENABLES_RX_RATE_PROFILE_VALID \
+ UINT32_C(0x1000)
/* Ring Type. */
uint8_t ring_type;
/* L2 Completion Ring (CR) */
@@ -45362,7 +45409,27 @@ struct hwrm_ring_alloc_input {
#define HWRM_RING_ALLOC_INPUT_MPC_CHNLS_TYPE_PRIMATE UINT32_C(0x4)
#define HWRM_RING_ALLOC_INPUT_MPC_CHNLS_TYPE_LAST \
HWRM_RING_ALLOC_INPUT_MPC_CHNLS_TYPE_PRIMATE
- uint8_t unused_4[2];
+ /* RX rate profile select */
+ uint8_t rx_rate_profile_sel;
+ /*
+* Indicate default RX rate profile when allocating
+* RX or RX aggregation rings. This should only be
+* used when 'rx_rate_profile_sel_supported' bit is
+* set in flags_ext3 field of FUNC_QCAPS response.
+*/
+ #define HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_DEFAULT \
+ UINT32_C(0x0)
+ /*
+* Indicate poll_mode RX rate profile when allocating
+* RX or RX aggregation rings. This should only be
+* used when 'rx_rate_profile_sel_supported' bit is
+* set in flags_ext3 field of FUNC_QCAPS response.
+*/
+ #define HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_POLL_MODE \
+ UINT32_C(0x1)
+ #define HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_LAST \
+ HWRM_RING_ALLOC_INPUT_RX_RATE_PROFILE_SEL_POLL_MODE
+ uint8_t unused_4;
/*
* The cq_handle is specified when allocating a completion ring. For
* devices that support NQs, this cq_handle will be included in the
--
2.39.5 (Apple Git-154)
[PATCH 3/3] net/bnxt: set max VFs count for P7
The number of max VFs per PF is 128 for P7 devices.
Cap the vnic hash table creation if the number of max VNICs
is less than 8.
Signed-off-by: Kishore Padmanabha
Reviewed-by: Ajit Khaparde
---
drivers/net/bnxt/bnxt.h | 1 +
drivers/net/bnxt/bnxt_vnic.c | 6 ++
2 files changed, 7 insertions(+)
diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h
index 0402de3eb9..b8b0ecc7e7 100644
--- a/drivers/net/bnxt/bnxt.h
+++ b/drivers/net/bnxt/bnxt.h
@@ -260,6 +260,7 @@ struct bnxt_pf_info {
#define BNXT_MAX_VFS(bp) ((bp)->pf->max_vfs)
#define BNXT_MAX_VF_REPS_P4 64
#define BNXT_MAX_VF_REPS_P5 256
+#define BNXT_MAX_VF_REPS_P7 128
#define BNXT_MAX_VF_REPS(bp) \
(BNXT_CHIP_P5(bp) ? BNXT_MAX_VF_REPS_P5 : \
BNXT_MAX_VF_REPS_P4)
diff --git a/drivers/net/bnxt/bnxt_vnic.c b/drivers/net/bnxt/bnxt_vnic.c
index 7b028f2ee5..4b5ac84d70 100644
--- a/drivers/net/bnxt/bnxt_vnic.c
+++ b/drivers/net/bnxt/bnxt_vnic.c
@@ -978,6 +978,12 @@ int32_t bnxt_vnic_queue_db_init(struct bnxt *bp)
hash_tbl_params.name = hash_tbl_name;
hash_tbl_params.entries = (bp->max_vnics > BNXT_VNIC_MAX_SUPPORTED_ID) ?
BNXT_VNIC_MAX_SUPPORTED_ID : bp->max_vnics;
+
+ /* if the number of max vnics is less than bucket size */
+ /* then let the max entries size be the least value */
+ if (hash_tbl_params.entries <= RTE_HASH_BUCKET_ENTRIES)
+ hash_tbl_params.entries = RTE_HASH_BUCKET_ENTRIES;
+
hash_tbl_params.key_len = BNXT_VNIC_MAX_QUEUE_SZ_IN_8BITS;
hash_tbl_params.socket_id = rte_socket_id();
bp->vnic_queue_db.rss_q_db = rte_hash_create(&hash_tbl_params);
--
2.39.5 (Apple Git-154)
Re: [PATCH v8 12/17] net/r8169: implement Tx path
Hello,
13/11/2024 10:28, Howard Wang:
> +static u32
> +rtl_get_opts1(struct rtl_tx_desc *txd)
> +{
> + rte_smp_rmb();
> +
> + return rte_le_to_cpu_32(txd->opts1);
> +}
We should avoid using such heavy memory barrier.
Is it possible to use a lighter fence or atomic API?
[PATCH] net/mlx5: fix the leak of action data list
In the actions construction of NT2HWS, the `masks` parameter is
always set to NULL and all the actions will be translated in the
"construct" stage as non-fixed ones.
In the stage of translating actions template, the actions data would
be allocated from the pool and managed in a list. The list would be
released when destroying the template with the actions. In the NT2HWS
implementation, the temporary template was freed directly and the
actions will be destroyed with the flow rule deletion. No other rule
would use this list anymore.
The actions data in the list should be freed when the actions
construction is done.
Fixes: ff4064d5b1fe ("net/mlx5: support bulk actions in non-template flow")
Cc: mkash...@nvidia.com
Signed-off-by: Bing Zhao
---
drivers/net/mlx5/mlx5_flow_hw.c | 63 -
1 file changed, 39 insertions(+), 24 deletions(-)
diff --git a/drivers/net/mlx5/mlx5_flow_hw.c b/drivers/net/mlx5/mlx5_flow_hw.c
index 50dbaa27ab..7233ac46c4 100644
--- a/drivers/net/mlx5/mlx5_flow_hw.c
+++ b/drivers/net/mlx5/mlx5_flow_hw.c
@@ -966,18 +966,15 @@ __flow_hw_actions_release(struct rte_eth_dev *dev, struct
mlx5_hw_actions *acts)
}
/**
- * Destroy DR actions created by action template.
- *
- * For DR actions created during table creation's action translate.
- * Need to destroy the DR action when destroying the table.
+ * Release the action data back into the pool without destory any action.
*
* @param[in] dev
* Pointer to the rte_eth_dev structure.
* @param[in] acts
* Pointer to the template HW steering DR actions.
*/
-static void
-__flow_hw_action_template_destroy(struct rte_eth_dev *dev, struct
mlx5_hw_actions *acts)
+static inline void
+__flow_hw_act_data_flush(struct rte_eth_dev *dev, struct mlx5_hw_actions *acts)
{
struct mlx5_priv *priv = dev->data->dev_private;
struct mlx5_action_construct_data *data;
@@ -987,7 +984,23 @@ __flow_hw_action_template_destroy(struct rte_eth_dev *dev,
struct mlx5_hw_action
LIST_REMOVE(data, next);
mlx5_ipool_free(priv->acts_ipool, data->idx);
}
+}
+/*
+ * Destroy DR actions created by action template.
+ *
+ * For DR actions created during table creation's action translate.
+ * Need to destroy the DR action when destroying the table.
+ *
+ * @param[in] dev
+ * Pointer to the rte_eth_dev structure.
+ * @param[in] acts
+ * Pointer to the template HW steering DR actions.
+ */
+static void
+__flow_hw_action_template_destroy(struct rte_eth_dev *dev, struct
mlx5_hw_actions *acts)
+{
+ __flow_hw_act_data_flush(dev, acts);
__flow_hw_actions_release(dev, acts);
}
@@ -13492,14 +13505,14 @@ flow_nta_build_template_mask(const struct
rte_flow_action actions[],
static int
flow_hw_translate_flow_actions(struct rte_eth_dev *dev,
- const struct rte_flow_attr *attr,
- const struct rte_flow_action actions[],
- struct rte_flow_hw *flow,
- struct mlx5_flow_hw_action_params *ap,
- struct mlx5_hw_actions *hw_acts,
- uint64_t item_flags, uint64_t action_flags,
- bool external,
- struct rte_flow_error *error)
+ const struct rte_flow_attr *attr,
+ const struct rte_flow_action actions[],
+ struct rte_flow_hw *flow,
+ struct mlx5_flow_hw_action_params *ap,
+ struct mlx5_hw_actions *hw_acts,
+ uint64_t item_flags, uint64_t action_flags,
+ bool external,
+ struct rte_flow_error *error)
{
int ret = 0;
uint32_t src_group = 0;
@@ -13542,7 +13555,7 @@ flow_hw_translate_flow_actions(struct rte_eth_dev *dev,
table = mlx5_malloc(MLX5_MEM_ZERO, sizeof(*table), 0, SOCKET_ID_ANY);
if (!table)
return rte_flow_error_set(error, ENOMEM,
RTE_FLOW_ERROR_TYPE_ACTION,
- actions, "Failed to allocate dummy table");
+ actions, "Failed to allocate dummy
table");
at = __flow_hw_actions_template_create(dev, &template_attr, actions,
masks, true, error);
if (!at) {
ret = -rte_errno;
@@ -13556,27 +13569,29 @@ flow_hw_translate_flow_actions(struct rte_eth_dev
*dev,
memcpy(&table->cfg.attr.flow_attr, attr, sizeof(*attr));
table->ats[0].action_template = at;
ret = __flow_hw_translate_actions_template(dev, &table->cfg, hw_acts,
at,
- &table->mpctx, true, error);
+ &table->mpctx, true, error);
if (ret)
goto end;
/* handle bulk actions register. */
ret = flow_hw_encap_decap_resource_register(dev, table, hw
[PATCH v1 0/4] TruFlow fixes for Thor2
This patch series fixes a few issues in TruFlow for Broadcom Thor2 NIC. Kishore Padmanabha (2): net/bnxt/tf_ulp: fix vfr clean up and stats lockup net/bnxt/tf_ulp: update template files Peter Spreadborough (1): net/bnxt/tf_ulp: performance and tuning changes for thor2 stats cache Shuanglin Wang (1): net/bnxt/tf_ulp: fix F1F2 vxlan counter acccumulation for Thor2 drivers/net/bnxt/tf_core/v3/tfc.h |12 +- drivers/net/bnxt/tf_core/v3/tfc_act.c |26 +- drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c|12 +- drivers/net/bnxt/tf_ulp/bnxt_ulp.c|17 + drivers/net/bnxt/tf_ulp/bnxt_ulp.h| 3 + .../generic_templates/ulp_template_db_enum.h | 8 +- .../ulp_template_db_thor2_class.c | 10126 .../ulp_template_db_thor_class.c | 8736 +++-- drivers/net/bnxt/tf_ulp/ulp_fc_mgr_tfc.c | 5 +- drivers/net/bnxt/tf_ulp/ulp_flow_db.c | 4 +- drivers/net/bnxt/tf_ulp/ulp_flow_db.h |13 + drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c | 309 +- drivers/net/bnxt/tf_ulp/ulp_sc_mgr.h |33 +- drivers/net/bnxt/tf_ulp/ulp_sc_mgr_tfc.c | 4 +- 14 files changed, 9627 insertions(+), 9681 deletions(-) -- 2.39.3
[PATCH v1 1/4] net/bnxt/tf_ulp: fix F1F2 vxlan counter acccumulation for Thor2
From: Shuanglin Wang
Add code for Thor2 to support the counter accumulation for
F1F2 vxlan parent-child flows.
Also, add a check for device state in the TF tunnel free API,
if it is zero then TF will skip the resource free as it
was already done by bnxt_free_all_hwrm_resources.
Fixes: 0513f0af034d ("net/bnxt/tf_ulp: add stats cache for Thor2")
Reviewed-by: Kishore Padmanabha
Signed-off-by: Shuanglin Wang
Signed-off-by: Sangtani Parag Satishbhai
Signed-off-by: Sriharsha Basavapatna
---
drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c | 12 ++-
drivers/net/bnxt/tf_ulp/ulp_flow_db.c | 4 +-
drivers/net/bnxt/tf_ulp/ulp_flow_db.h | 13 +++
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c | 110 +++--
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.h | 18 +++-
5 files changed, 142 insertions(+), 15 deletions(-)
diff --git a/drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c
b/drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c
index c8f3a5abc2..5854f13447 100644
--- a/drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c
+++ b/drivers/net/bnxt/tf_ulp/bnxt_tf_pmd_shim.c
@@ -587,10 +587,16 @@ bnxt_pmd_global_tunnel_set(struct bnxt_ulp_context
*ulp_ctx,
return -EINVAL;
}
- if (udp_port)
+ if (udp_port) {
rc = bnxt_udp_tunnel_port_add_op(eth_dev, &udp_tunnel);
- else
- rc = bnxt_udp_tunnel_port_del_op(eth_dev, &udp_tunnel);
+ } else {
+ /* TODO: Make the counters shareable so the resource
+* free can be synced up between core dpdk path and
+* the tf path.
+*/
+ if (eth_dev->data->dev_started != 0)
+ rc = bnxt_udp_tunnel_port_del_op(eth_dev,
&udp_tunnel);
+ }
} else {
bp = bnxt_pmd_get_bp(port_id);
if (!bp) {
diff --git a/drivers/net/bnxt/tf_ulp/ulp_flow_db.c
b/drivers/net/bnxt/tf_ulp/ulp_flow_db.c
index d6fd653b4a..943c9b799c 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_flow_db.c
+++ b/drivers/net/bnxt/tf_ulp/ulp_flow_db.c
@@ -1176,7 +1176,7 @@ ulp_default_flow_db_cfa_action_get(struct
bnxt_ulp_context *ulp_ctx,
}
/* internal validation function for parent flow tbl */
-static struct ulp_fdb_parent_info *
+struct ulp_fdb_parent_info *
ulp_flow_db_pc_db_entry_get(struct bnxt_ulp_context *ulp_ctxt,
uint32_t pc_idx)
{
@@ -1634,6 +1634,8 @@ ulp_flow_db_parent_flow_create(struct
bnxt_ulp_mapper_parms *parms)
}
}
+ /* Set parent flow entry idx in stats cache entry */
+ ulp_sc_mgr_set_pc_idx(parms->ulp_ctx, parms->flow_id, pc_idx);
return 0;
}
diff --git a/drivers/net/bnxt/tf_ulp/ulp_flow_db.h
b/drivers/net/bnxt/tf_ulp/ulp_flow_db.h
index 762ebb039d..4301094cfe 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_flow_db.h
+++ b/drivers/net/bnxt/tf_ulp/ulp_flow_db.h
@@ -417,4 +417,17 @@ void ulp_flow_db_shared_session_set(struct
ulp_flow_db_res_params *res,
*/
enum bnxt_ulp_session_type
ulp_flow_db_shared_session_get(struct ulp_flow_db_res_params *res);
+
+/*
+ * Get the parent flow table info
+ *
+ * ulp_ctxt [in] Ptr to ulp_context
+ * pc_idx [in] The index to parent child db
+ *
+ * returns Pointer of parent flow tbl
+ */
+struct ulp_fdb_parent_info *
+ulp_flow_db_pc_db_entry_get(struct bnxt_ulp_context *ulp_ctxt,
+ uint32_t pc_idx);
+
#endif /* _ULP_FLOW_DB_H_ */
diff --git a/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
b/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
index 13069126f0..1317668555 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
+++ b/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
@@ -441,8 +441,17 @@ int ulp_sc_mgr_query_count_get(struct bnxt_ulp_context
*ctxt,
{
struct ulp_sc_tfc_stats_cache_entry *sce;
struct bnxt_ulp_sc_info *ulp_sc_info;
+ struct ulp_fdb_parent_info *pc_entry;
+ struct bnxt_ulp_flow_db *flow_db;
+ uint32_t max_array;
+ uint32_t child_fid;
+ uint32_t a_idx;
+ uint32_t f2_cnt;
+ uint64_t *t;
+ uint64_t bs;
int rc = 0;
+ /* Get stats cache info */
ulp_sc_info = bnxt_ulp_cntxt_ptr2_sc_info_get(ctxt);
if (!ulp_sc_info)
return -ENODEV;
@@ -450,18 +459,66 @@ int ulp_sc_mgr_query_count_get(struct bnxt_ulp_context
*ctxt,
sce = ulp_sc_info->stats_cache_tbl;
sce += flow_id;
- /* If entry is not valid return an error */
- if (!(sce->flags & ULP_SC_ENTRY_FLAG_VALID))
- return -EBUSY;
-
- count->hits = sce->packet_count;
- count->hits_set = 1;
- count->bytes = sce->byte_count;
- count->bytes_set = 1;
+ /* To handle the parent flow */
+ if (sce->flags & ULP_SC_ENTRY_FLAG_PARENT) {
+ flow_db = bnxt_ulp_cntxt_ptr2_flow_db_get(ctxt);
+ if (!flow_db) {
+
[PATCH v1 2/4] net/bnxt/tf_ulp: fix vfr clean up and stats lockup
From: Kishore Padmanabha
The representor flows were not being deleted as part of the
vfr clean up. Added code to delete flows related to vfr interface.
Also fixed the stats counter thread lockup.
Fixes: 0513f0af034d ("net/bnxt/tf_ulp: add stats cache for Thor2")
Reviewed-by: Peter Spreadborough
Reviewed-by: Shuanglin Wang
Signed-off-by: Kishore Padmanabha
Signed-off-by: Sriharsha Basavapatna
---
drivers/net/bnxt/tf_ulp/bnxt_ulp.c | 17 +++
drivers/net/bnxt/tf_ulp/bnxt_ulp.h | 3 +
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c | 164 ---
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.h | 12 +-
4 files changed, 74 insertions(+), 122 deletions(-)
diff --git a/drivers/net/bnxt/tf_ulp/bnxt_ulp.c
b/drivers/net/bnxt/tf_ulp/bnxt_ulp.c
index e28a481f5e..1bfc88cf79 100644
--- a/drivers/net/bnxt/tf_ulp/bnxt_ulp.c
+++ b/drivers/net/bnxt/tf_ulp/bnxt_ulp.c
@@ -596,6 +596,9 @@ bnxt_ulp_port_deinit(struct bnxt *bp)
/* close the session associated with this port */
bp->ulp_ctx->ops->ulp_ctx_detach(bp, session);
} else {
+ /* Free the ulp context in the context entry list */
+ bnxt_ulp_cntxt_list_del(bp->ulp_ctx);
+
/* clean up default flows */
bnxt_ulp_destroy_df_rules(bp, true);
@@ -662,6 +665,20 @@ bnxt_ulp_cntxt_list_del(struct bnxt_ulp_context *ulp_ctx)
rte_spinlock_unlock(&bnxt_ulp_ctxt_lock);
}
+int
+bnxt_ulp_cntxt_list_count(void)
+{
+ struct ulp_context_list_entry *entry, *temp;
+ int count_1 = 0;
+
+ rte_spinlock_lock(&bnxt_ulp_ctxt_lock);
+ RTE_TAILQ_FOREACH_SAFE(entry, &ulp_cntx_list, next, temp) {
+ count_1++;
+ }
+ rte_spinlock_unlock(&bnxt_ulp_ctxt_lock);
+ return count_1;
+}
+
struct bnxt_ulp_context *
bnxt_ulp_cntxt_entry_acquire(void *arg)
{
diff --git a/drivers/net/bnxt/tf_ulp/bnxt_ulp.h
b/drivers/net/bnxt/tf_ulp/bnxt_ulp.h
index 83fb205f68..e0e31532fd 100644
--- a/drivers/net/bnxt/tf_ulp/bnxt_ulp.h
+++ b/drivers/net/bnxt/tf_ulp/bnxt_ulp.h
@@ -297,6 +297,9 @@ bnxt_ulp_cntxt_list_add(struct bnxt_ulp_context *ulp_ctx);
void
bnxt_ulp_cntxt_list_del(struct bnxt_ulp_context *ulp_ctx);
+int
+bnxt_ulp_cntxt_list_count(void);
+
struct bnxt_ulp_context *
bnxt_ulp_cntxt_entry_acquire(void *arg);
diff --git a/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
b/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
index 1317668555..c82fdaf6dd 100644
--- a/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
+++ b/drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c
@@ -89,12 +89,6 @@ int32_t ulp_sc_mgr_init(struct bnxt_ulp_context *ctxt)
ulp_sc_info->sc_ops = sc_ops;
ulp_sc_info->flags = 0;
- rc = pthread_mutex_init(&ulp_sc_info->sc_lock, NULL);
- if (rc) {
- BNXT_DRV_DBG(ERR, "Failed to initialize sc mutex\n");
- goto error;
- }
-
/* Add the SC info tbl to the ulp context. */
bnxt_ulp_cntxt_ptr2_sc_info_set(ctxt, ulp_sc_info);
@@ -109,9 +103,10 @@ int32_t ulp_sc_mgr_init(struct bnxt_ulp_context *ctxt)
* Size is determined by the number of flows + 10% to cover IDs
* used for resources.
*/
+ ulp_sc_info->cache_tbl_size = ulp_sc_info->num_counters +
+ (ulp_sc_info->num_counters / 10);
stats_cache_tbl_sz = sizeof(struct ulp_sc_tfc_stats_cache_entry) *
- (ulp_sc_info->num_counters +
-(ulp_sc_info->num_counters / 10));
+ ulp_sc_info->cache_tbl_size;
ulp_sc_info->stats_cache_tbl = rte_zmalloc("ulp_stats_cache_tbl",
stats_cache_tbl_sz, 0);
@@ -153,12 +148,6 @@ ulp_sc_mgr_deinit(struct bnxt_ulp_context *ctxt)
if (!ulp_sc_info)
return -EINVAL;
- pthread_mutex_lock(&ulp_sc_info->sc_lock);
-
- ulp_sc_mgr_thread_cancel(ctxt);
-
- pthread_mutex_destroy(&ulp_sc_info->sc_lock);
-
if (ulp_sc_info->stats_cache_tbl)
rte_free(ulp_sc_info->stats_cache_tbl);
@@ -173,13 +162,13 @@ ulp_sc_mgr_deinit(struct bnxt_ulp_context *ctxt)
return 0;
}
-#define ULP_SC_PERIOD_S 1
-#define ULP_SC_PERIOD_MS (ULP_SC_PERIOD_S * 1000)
+#define ULP_SC_PERIOD_US 256
+#define ULP_SC_CTX_DELAY 1
static uint32_t ulp_stats_cache_main_loop(void *arg)
{
struct ulp_sc_tfc_stats_cache_entry *count;
- const struct bnxt_ulp_sc_core_ops *sc_ops;
+ const struct bnxt_ulp_sc_core_ops *sc_ops = NULL;
struct ulp_sc_tfc_stats_cache_entry *sce;
struct ulp_sc_tfc_stats_cache_entry *sce_end;
struct tfc_mpc_batch_info_t batch_info;
@@ -188,95 +177,68 @@ static uint32_t ulp_stats_cache_main_loop(void *arg)
uint16_t words = (ULP_TFC_CNTR_READ_BYTES + ULP_TFC_ACT_WORD_SZ - 1) /
ULP_TFC_ACT_WORD_SZ;
uint32_t batch_size;
struct tfc *tfcp = NULL;
- uint32_t batch;
- uint32_t delay = ULP
[PATCH v1 3/4] net/bnxt/tf_ulp: performance and tuning changes for thor2 stats cache
From: Peter Spreadborough
This change adds changes the act get API to expect a physical
address for the host memory rather than a pointer to a local
virtual address. The change was made because the virt->phys
API call has a very high overhead. Also included in this change
to the mutex locks and delays within the main loop.
Fixes: 0513f0af034d ("net/bnxt/tf_ulp: add stats cache for Thor2")
Reviewed-by: Farah Smith
Reviewed-by: Jay Ding
Signed-off-by: Peter Spreadborough
Signed-off-by: Sriharsha Basavapatna
---
drivers/net/bnxt/tf_core/v3/tfc.h| 12 ---
drivers/net/bnxt/tf_core/v3/tfc_act.c| 26 ++
drivers/net/bnxt/tf_ulp/ulp_fc_mgr_tfc.c | 5 ++-
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.c | 45
drivers/net/bnxt/tf_ulp/ulp_sc_mgr.h | 3 +-
drivers/net/bnxt/tf_ulp/ulp_sc_mgr_tfc.c | 4 +--
6 files changed, 65 insertions(+), 30 deletions(-)
diff --git a/drivers/net/bnxt/tf_core/v3/tfc.h
b/drivers/net/bnxt/tf_core/v3/tfc.h
index 98725b32c2..4176aa01bd 100644
--- a/drivers/net/bnxt/tf_core/v3/tfc.h
+++ b/drivers/net/bnxt/tf_core/v3/tfc.h
@@ -705,7 +705,9 @@ enum tfc_tbl_scope_bucket_factor {
TFC_TBL_SCOPE_BUCKET_FACTOR_4 = 4,
TFC_TBL_SCOPE_BUCKET_FACTOR_8 = 8,
TFC_TBL_SCOPE_BUCKET_FACTOR_16 = 16,
- TFC_TBL_SCOPE_BUCKET_FACTOR_MAX = TFC_TBL_SCOPE_BUCKET_FACTOR_16
+ TFC_TBL_SCOPE_BUCKET_FACTOR_32 = 32,
+ TFC_TBL_SCOPE_BUCKET_FACTOR_64 = 64,
+ TFC_TBL_SCOPE_BUCKET_FACTOR_MAX = TFC_TBL_SCOPE_BUCKET_FACTOR_64
};
/**
@@ -1420,8 +1422,9 @@ int tfc_act_set(struct tfc *tfcp,
* @param[in] cmm_clr
* Pointer to cmm clr
*
- * @param[in,out] data
- * Data read. Must be word aligned, i.e. [1:0] must be 0.
+ * @param[in,out] host_address
+ * Data read. Must be word aligned, i.e. [1:0] must be 0. The address
+ * must be the ret_mem_virt2iova() version of the virt address.
*
* @param[in,out] data_sz_words
* Data buffer size in words.Size could be 8/16/24/32/64B
@@ -1437,7 +1440,8 @@ int tfc_act_get(struct tfc *tfcp,
struct tfc_mpc_batch_info_t *batch_info,
const struct tfc_cmm_info *cmm_info,
struct tfc_cmm_clr *clr,
- uint8_t *data, uint16_t *data_sz_words);
+ uint64_t *host_address,
+ uint16_t *data_sz_words);
/**
* Free a CMM Resource
diff --git a/drivers/net/bnxt/tf_core/v3/tfc_act.c
b/drivers/net/bnxt/tf_core/v3/tfc_act.c
index 3e215f4881..0e98bd30d7 100644
--- a/drivers/net/bnxt/tf_core/v3/tfc_act.c
+++ b/drivers/net/bnxt/tf_core/v3/tfc_act.c
@@ -367,7 +367,7 @@ int tfc_act_get_only_response(struct cfa_bld_mpcinfo
*mpc_info,
static int tfc_act_get_only(struct tfc *tfcp,
struct tfc_mpc_batch_info_t *batch_info,
const struct tfc_cmm_info *cmm_info,
- uint8_t *data,
+ uint64_t *host_address,
uint16_t *data_sz_words)
{
int rc = 0;
@@ -378,7 +378,6 @@ static int tfc_act_get_only(struct tfc *tfcp,
uint32_t buff_len;
struct cfa_mpc_data_obj fields_cmd[CFA_BLD_MPC_READ_CMD_MAX_FLD] = {
{0} };
uint32_t entry_offset;
- uint64_t host_address;
struct bnxt_mpc_mbuf mpc_msg_in;
struct bnxt_mpc_mbuf mpc_msg_out;
uint32_t record_size;
@@ -405,13 +404,11 @@ static int tfc_act_get_only(struct tfc *tfcp,
}
/* Check that data pointer is word aligned */
- if (unlikely(((uint64_t)data) & 0x3ULL)) {
+ if (unlikely(*host_address & 0x3ULL)) {
PMD_DRV_LOG_LINE(ERR, "data pointer not word aligned");
return -EINVAL;
}
- host_address = (uint64_t)rte_mem_virt2iova(data);
-
/* Check that MPC APIs are bound */
if (unlikely(mpc_info->mpcops == NULL)) {
PMD_DRV_LOG_LINE(ERR, "MPC not initialized");
@@ -450,7 +447,7 @@ static int tfc_act_get_only(struct tfc *tfcp,
fields_cmd[CFA_BLD_MPC_READ_CMD_HOST_ADDRESS_FLD].field_id =
CFA_BLD_MPC_READ_CMD_HOST_ADDRESS_FLD;
- fields_cmd[CFA_BLD_MPC_READ_CMD_HOST_ADDRESS_FLD].val = host_address;
+ fields_cmd[CFA_BLD_MPC_READ_CMD_HOST_ADDRESS_FLD].val = *host_address;
buff_len = TFC_MPC_MAX_TX_BYTES;
@@ -547,7 +544,7 @@ int tfc_act_get_clear_response(struct cfa_bld_mpcinfo
*mpc_info,
static int tfc_act_get_clear(struct tfc *tfcp,
struct tfc_mpc_batch_info_t *batch_info,
const struct tfc_cmm_info *cmm_info,
-uint8_t *data,
+uint64_t *host_address,
uint16_t *data_sz_words,
uint8_t clr_offset,
uint8_t clr_size)
@@ -560,7 +557,6 @@ static int tfc_act_get_clear(struct tfc *tfcp,
uint32_t buff_len;
struct cfa_mpc_d
[PATCH v2 1/2] app/testpmd: fix flow update
If actions provided to “flow update…“ command contained an age
action, then testpmd did not update the age action context
accordingly.
Thus "flow aged destroy" command can not
execute successfully.
Fix was done with next steps
1. Generate new port flow entry to add/replace action(s).
2. Set age context if age action is present.
3. Replace flow in the flow list.
Fixes: 2d9c7e56e52c ("app/testpmd: support updating flow rule actions")
Cc: sta...@dpdk.org
Signed-off-by: Danylo Vodopianov
---
app/test-pmd/config.c | 21 -
1 file changed, 20 insertions(+), 1 deletion(-)
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index 88770b4dfc..c831166431 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -3882,7 +3882,8 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
const struct rte_flow_action *actions, bool is_user_id)
{
struct rte_port *port;
- struct port_flow **flow_list;
+ struct port_flow **flow_list, *uf;
+ struct rte_flow_action_age *age = age_action_get(actions);
if (port_id_is_invalid(port_id, ENABLED_WARN) ||
port_id == (portid_t)RTE_PORT_ALL)
@@ -3897,6 +3898,16 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
flow_list = &flow->next;
continue;
}
+
+ /* Update flow action(s) with new action(s) */
+ uf = port_flow_new(flow->rule.attr_ro, flow->rule.pattern_ro,
actions, &error);
+ if (!uf)
+ return port_flow_complain(&error);
+ if (age) {
+ flow->age_type = ACTION_AGE_CONTEXT_TYPE_FLOW;
+ age->context = &flow->age_type;
+ }
+
/*
* Poisoning to make sure PMDs update it in case
* of error.
@@ -3913,6 +3924,14 @@ port_flow_update(portid_t port_id, uint32_t rule_id,
printf("Flow rule #%"PRIu64
" updated with new actions\n",
flow->id);
+
+ uf->next = flow->next;
+ uf->id = flow->id;
+ uf->user_id = flow->user_id;
+ uf->flow = flow->flow;
+ *flow_list = uf;
+
+ free(flow);
return 0;
}
printf("Failed to find flow %"PRIu32"\n", rule_id);
--
2.43.5
