* [dpdk-dev] [PATCH] net/ice: cleanup RSS/FDIR profile when device init @ 2019-09-27 18:32 Ying Wang 2019-09-30 9:25 ` Ye Xiaolong 2019-10-17 18:31 ` [dpdk-dev] [PATCH v2] " Ying Wang 0 siblings, 2 replies; 5+ messages in thread From: Ying Wang @ 2019-09-27 18:32 UTC (permalink / raw) To: xiaolong.ye, qi.z.zhang; +Cc: dev, qiming.yang, ying.a.wang, stable The patch cleanup RSS/FDIR profile resources when device init to fix profile ID increase issue. Fixes: d7d150b93070 ("net/ice: enable RSS when device init") Cc: stable@dpdk.org Signed-off-by: Ying Wang <ying.a.wang@intel.com> --- --- This patch depends on the following patches in patchwork: (1) http://patchwork.dpdk.org/project/dpdk/list/?series=6557 (2) http://patches.dpdk.org/project/dpdk/list/?series=6579 (3) http://patches.dpdk.org/project/dpdk/list/?series=6577 --- drivers/net/ice/ice_ethdev.c | 115 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 115 insertions(+) diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c index 3abdaffbc..04102728f 100644 --- a/drivers/net/ice/ice_ethdev.c +++ b/drivers/net/ice/ice_ethdev.c @@ -13,6 +13,7 @@ #include "base/ice_sched.h" #include "base/ice_flow.h" #include "base/ice_dcb.h" +#include "base/ice_common.h" #include "ice_ethdev.h" #include "ice_rxtx.h" #include "ice_generic_flow.h" @@ -40,6 +41,7 @@ static const char * const ice_valid_args[] = { #define ICE_OS_DEFAULT_PKG_NAME "ICE OS Default Package" #define ICE_COMMS_PKG_NAME "ICE COMMS Package" #define ICE_MAX_PKG_FILENAME_SIZE 256 +#define ICE_MAX_RES_DESC_NUM 1024 int ice_logtype_init; int ice_logtype_driver; @@ -1851,6 +1853,113 @@ ice_vsi_config_sw_lldp(struct ice_vsi *vsi, bool on) return ret; } +static enum ice_status +ice_aq_get_res_profs(struct ice_hw *hw, uint16_t num_entries, + struct ice_aqc_get_allocd_res_desc_resp *resp_buf, + uint16_t buf_size, uint16_t res_type, + bool res_shared, uint16_t *desc_id, + struct ice_sq_cd *cd, uint16_t *num_prof) +{ + struct ice_aq_desc aq_desc; + int ret; + + struct ice_aqc_get_allocd_res_desc *cmd = + &aq_desc.params.get_res_desc; + + if (!resp_buf) + return ICE_ERR_PARAM; + + if (buf_size != (num_entries * sizeof(*resp_buf))) + return ICE_ERR_PARAM; + + ice_fill_dflt_direct_cmd_desc(&aq_desc, + ice_aqc_opc_get_allocd_res_desc); + + cmd->ops.cmd.res = CPU_TO_LE16(((res_type << ICE_AQC_RES_TYPE_S) & + ICE_AQC_RES_TYPE_M) | (res_shared ? + ICE_AQC_RES_TYPE_FLAG_SHARED : 0)); + cmd->ops.cmd.first_desc = CPU_TO_LE16(*desc_id); + ret = ice_aq_send_cmd(hw, &aq_desc, resp_buf, buf_size, cd); + if (!ret) { + *desc_id = LE16_TO_CPU(cmd->ops.resp.next_desc); + *num_prof = LE16_TO_CPU(cmd->ops.resp.num_desc); + } + return ret; +} + +static enum ice_status +ice_get_hw_res(struct ice_hw *hw, uint16_t type, uint16_t num, uint16_t *res, + uint16_t *prof_buf, uint16_t *num_prof) +{ + struct ice_aqc_get_allocd_res_desc_resp *resp_buf; + int ret; + uint16_t buf_len; + + buf_len = sizeof(*resp_buf) + sizeof(resp_buf->elem) * (num - 1); + resp_buf = (struct ice_aqc_get_allocd_res_desc_resp *) + ice_malloc(hw, buf_len); + if (!resp_buf) + return -ENOMEM; + + ret = ice_aq_get_res_profs(hw, num, resp_buf, buf_len, type, 1, + res, NULL, num_prof); + if (ret) + goto exit; + + ice_memcpy(prof_buf, resp_buf->elem, sizeof(resp_buf->elem) * + (*num_prof), ICE_NONDMA_TO_NONDMA); +exit: + rte_free(resp_buf); + return ret; +} + +static int +ice_clearup_resource(struct ice_hw *hw, uint16_t res_type) +{ + int ret; + uint16_t prof_id; + uint16_t prof_buf[ICE_MAX_RES_DESC_NUM]; + uint16_t first_desc = 1; + uint16_t num_prof = 0; + + ret = ice_get_hw_res(hw, res_type, ICE_MAX_RES_DESC_NUM, + &first_desc, prof_buf, &num_prof); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to get fxp resource"); + return ret; + } + + for (prof_id = 0; prof_id < num_prof; prof_id++) { + ret = ice_free_hw_res(hw, res_type, 1, &prof_buf[prof_id]); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to free fxp resource"); + return ret; + } + } + + return 0; +} + +static int +ice_reset_fxp_resource(struct ice_hw *hw) +{ + int ret; + + ret = ice_clearup_resource(hw, ICE_AQC_RES_TYPE_FD_PROF_BLDR_PROFID); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to clearup fdir resource"); + return ret; + } + + ret = ice_clearup_resource(hw, ICE_AQC_RES_TYPE_HASH_PROF_BLDR_PROFID); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to clearup rss resource"); + return ret; + } + + return 0; +} + static int ice_dev_init(struct rte_eth_dev *dev) { @@ -1983,6 +2092,12 @@ ice_dev_init(struct rte_eth_dev *dev) return ret; } + ret = ice_reset_fxp_resource(hw); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to reset fxp resource"); + return ret; + } + return 0; err_pf_setup: -- 2.15.1 ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [dpdk-dev] [PATCH] net/ice: cleanup RSS/FDIR profile when device init 2019-09-27 18:32 [dpdk-dev] [PATCH] net/ice: cleanup RSS/FDIR profile when device init Ying Wang @ 2019-09-30 9:25 ` Ye Xiaolong 2019-10-18 0:42 ` Wang, Ying A 2019-10-17 18:31 ` [dpdk-dev] [PATCH v2] " Ying Wang 1 sibling, 1 reply; 5+ messages in thread From: Ye Xiaolong @ 2019-09-30 9:25 UTC (permalink / raw) To: Ying Wang; +Cc: qi.z.zhang, dev, qiming.yang, stable On 09/28, Ying Wang wrote: >The patch cleanup RSS/FDIR profile resources when device init >to fix profile ID increase issue. Could you elaborate the profile ID increase issue in the commit log? > >Fixes: d7d150b93070 ("net/ice: enable RSS when device init") >Cc: stable@dpdk.org > >Signed-off-by: Ying Wang <ying.a.wang@intel.com> >--- >--- >This patch depends on the following patches in patchwork: >(1) http://patchwork.dpdk.org/project/dpdk/list/?series=6557 >(2) http://patches.dpdk.org/project/dpdk/list/?series=6579 >(3) http://patches.dpdk.org/project/dpdk/list/?series=6577 >--- > > drivers/net/ice/ice_ethdev.c | 115 +++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 115 insertions(+) > >diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c >index 3abdaffbc..04102728f 100644 >--- a/drivers/net/ice/ice_ethdev.c >+++ b/drivers/net/ice/ice_ethdev.c >@@ -13,6 +13,7 @@ > #include "base/ice_sched.h" > #include "base/ice_flow.h" > #include "base/ice_dcb.h" >+#include "base/ice_common.h" > #include "ice_ethdev.h" > #include "ice_rxtx.h" > #include "ice_generic_flow.h" >@@ -40,6 +41,7 @@ static const char * const ice_valid_args[] = { > #define ICE_OS_DEFAULT_PKG_NAME "ICE OS Default Package" > #define ICE_COMMS_PKG_NAME "ICE COMMS Package" > #define ICE_MAX_PKG_FILENAME_SIZE 256 >+#define ICE_MAX_RES_DESC_NUM 1024 > > int ice_logtype_init; > int ice_logtype_driver; >@@ -1851,6 +1853,113 @@ ice_vsi_config_sw_lldp(struct ice_vsi *vsi, bool on) > return ret; > } > >+static enum ice_status >+ice_aq_get_res_profs(struct ice_hw *hw, uint16_t num_entries, >+ struct ice_aqc_get_allocd_res_desc_resp *resp_buf, >+ uint16_t buf_size, uint16_t res_type, >+ bool res_shared, uint16_t *desc_id, >+ struct ice_sq_cd *cd, uint16_t *num_prof) May consider use struct to reduce the number of parameters. >+{ >+ struct ice_aq_desc aq_desc; >+ int ret; >+ >+ struct ice_aqc_get_allocd_res_desc *cmd = >+ &aq_desc.params.get_res_desc; >+ >+ if (!resp_buf) >+ return ICE_ERR_PARAM; >+ >+ if (buf_size != (num_entries * sizeof(*resp_buf))) >+ return ICE_ERR_PARAM; >+ >+ ice_fill_dflt_direct_cmd_desc(&aq_desc, >+ ice_aqc_opc_get_allocd_res_desc); >+ >+ cmd->ops.cmd.res = CPU_TO_LE16(((res_type << ICE_AQC_RES_TYPE_S) & >+ ICE_AQC_RES_TYPE_M) | (res_shared ? >+ ICE_AQC_RES_TYPE_FLAG_SHARED : 0)); >+ cmd->ops.cmd.first_desc = CPU_TO_LE16(*desc_id); >+ ret = ice_aq_send_cmd(hw, &aq_desc, resp_buf, buf_size, cd); >+ if (!ret) { >+ *desc_id = LE16_TO_CPU(cmd->ops.resp.next_desc); >+ *num_prof = LE16_TO_CPU(cmd->ops.resp.num_desc); >+ } >+ return ret; >+} >+ >+static enum ice_status >+ice_get_hw_res(struct ice_hw *hw, uint16_t type, uint16_t num, uint16_t *res, >+ uint16_t *prof_buf, uint16_t *num_prof) >+{ >+ struct ice_aqc_get_allocd_res_desc_resp *resp_buf; >+ int ret; >+ uint16_t buf_len; >+ >+ buf_len = sizeof(*resp_buf) + sizeof(resp_buf->elem) * (num - 1); why (num - 1) here? >+ resp_buf = (struct ice_aqc_get_allocd_res_desc_resp *) >+ ice_malloc(hw, buf_len); Cast is unnecessary for void * >+ if (!resp_buf) >+ return -ENOMEM; >+ >+ ret = ice_aq_get_res_profs(hw, num, resp_buf, buf_len, type, 1, >+ res, NULL, num_prof); >+ if (ret) >+ goto exit; >+ >+ ice_memcpy(prof_buf, resp_buf->elem, sizeof(resp_buf->elem) * >+ (*num_prof), ICE_NONDMA_TO_NONDMA); >+exit: >+ rte_free(resp_buf); >+ return ret; >+} >+ >+static int >+ice_clearup_resource(struct ice_hw *hw, uint16_t res_type) s/clearup/cleanup >+{ >+ int ret; >+ uint16_t prof_id; >+ uint16_t prof_buf[ICE_MAX_RES_DESC_NUM]; >+ uint16_t first_desc = 1; >+ uint16_t num_prof = 0; >+ >+ ret = ice_get_hw_res(hw, res_type, ICE_MAX_RES_DESC_NUM, >+ &first_desc, prof_buf, &num_prof); What is purpose of first_desc? Seems there is no reference of it afterwards. >+ if (ret) { >+ PMD_INIT_LOG(ERR, "Failed to get fxp resource"); >+ return ret; >+ } >+ >+ for (prof_id = 0; prof_id < num_prof; prof_id++) { >+ ret = ice_free_hw_res(hw, res_type, 1, &prof_buf[prof_id]); >+ if (ret) { >+ PMD_INIT_LOG(ERR, "Failed to free fxp resource"); >+ return ret; >+ } >+ } >+ >+ return 0; >+} >+ >+static int >+ice_reset_fxp_resource(struct ice_hw *hw) >+{ >+ int ret; >+ >+ ret = ice_clearup_resource(hw, ICE_AQC_RES_TYPE_FD_PROF_BLDR_PROFID); >+ if (ret) { >+ PMD_INIT_LOG(ERR, "Failed to clearup fdir resource"); >+ return ret; >+ } >+ >+ ret = ice_clearup_resource(hw, ICE_AQC_RES_TYPE_HASH_PROF_BLDR_PROFID); >+ if (ret) { >+ PMD_INIT_LOG(ERR, "Failed to clearup rss resource"); >+ return ret; >+ } >+ >+ return 0; >+} s/clearup/cleanup Thanks, Xiaolong >+ > static int > ice_dev_init(struct rte_eth_dev *dev) > { >@@ -1983,6 +2092,12 @@ ice_dev_init(struct rte_eth_dev *dev) > return ret; > } > >+ ret = ice_reset_fxp_resource(hw); >+ if (ret) { >+ PMD_INIT_LOG(ERR, "Failed to reset fxp resource"); >+ return ret; >+ } >+ > return 0; > > err_pf_setup: >-- >2.15.1 > ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [dpdk-dev] [PATCH] net/ice: cleanup RSS/FDIR profile when device init 2019-09-30 9:25 ` Ye Xiaolong @ 2019-10-18 0:42 ` Wang, Ying A 0 siblings, 0 replies; 5+ messages in thread From: Wang, Ying A @ 2019-10-18 0:42 UTC (permalink / raw) To: Ye, Xiaolong; +Cc: Zhang, Qi Z, dev, Yang, Qiming, stable Hi, Xiaolong Thanks for your comments, will fix them in v2. > -----Original Message----- > From: Ye, Xiaolong > Sent: Monday, September 30, 2019 5:26 PM > To: Wang, Ying A <ying.a.wang@intel.com> > Cc: Zhang, Qi Z <qi.z.zhang@intel.com>; dev@dpdk.org; Yang, Qiming > <qiming.yang@intel.com>; stable@dpdk.org > Subject: Re: [PATCH] net/ice: cleanup RSS/FDIR profile when device init > > On 09/28, Ying Wang wrote: > >The patch cleanup RSS/FDIR profile resources when device init to fix > >profile ID increase issue. > > Could you elaborate the profile ID increase issue in the commit log? > > > > >Fixes: d7d150b93070 ("net/ice: enable RSS when device init") > >Cc: stable@dpdk.org > > > >Signed-off-by: Ying Wang <ying.a.wang@intel.com> > >--- > >--- > >This patch depends on the following patches in patchwork: > >(1) http://patchwork.dpdk.org/project/dpdk/list/?series=6557 > >(2) http://patches.dpdk.org/project/dpdk/list/?series=6579 > >(3) http://patches.dpdk.org/project/dpdk/list/?series=6577 > >--- > > > > drivers/net/ice/ice_ethdev.c | 115 > > +++++++++++++++++++++++++++++++++++++++++++ > > 1 file changed, 115 insertions(+) > > > >diff --git a/drivers/net/ice/ice_ethdev.c > >b/drivers/net/ice/ice_ethdev.c index 3abdaffbc..04102728f 100644 > >--- a/drivers/net/ice/ice_ethdev.c > >+++ b/drivers/net/ice/ice_ethdev.c > >@@ -13,6 +13,7 @@ > > #include "base/ice_sched.h" > > #include "base/ice_flow.h" > > #include "base/ice_dcb.h" > >+#include "base/ice_common.h" > > #include "ice_ethdev.h" > > #include "ice_rxtx.h" > > #include "ice_generic_flow.h" > >@@ -40,6 +41,7 @@ static const char * const ice_valid_args[] = { > > #define ICE_OS_DEFAULT_PKG_NAME "ICE OS Default Package" > > #define ICE_COMMS_PKG_NAME "ICE COMMS Package" > > #define ICE_MAX_PKG_FILENAME_SIZE 256 > >+#define ICE_MAX_RES_DESC_NUM 1024 > > > > int ice_logtype_init; > > int ice_logtype_driver; > >@@ -1851,6 +1853,113 @@ ice_vsi_config_sw_lldp(struct ice_vsi *vsi, bool > on) > > return ret; > > } > > > >+static enum ice_status > >+ice_aq_get_res_profs(struct ice_hw *hw, uint16_t num_entries, > >+ struct ice_aqc_get_allocd_res_desc_resp *resp_buf, > >+ uint16_t buf_size, uint16_t res_type, > >+ bool res_shared, uint16_t *desc_id, > >+ struct ice_sq_cd *cd, uint16_t *num_prof) > > May consider use struct to reduce the number of parameters. > > >+{ > >+ struct ice_aq_desc aq_desc; > >+ int ret; > >+ > >+ struct ice_aqc_get_allocd_res_desc *cmd = > >+ &aq_desc.params.get_res_desc; > >+ > >+ if (!resp_buf) > >+ return ICE_ERR_PARAM; > >+ > >+ if (buf_size != (num_entries * sizeof(*resp_buf))) > >+ return ICE_ERR_PARAM; > >+ > >+ ice_fill_dflt_direct_cmd_desc(&aq_desc, > >+ ice_aqc_opc_get_allocd_res_desc); > >+ > >+ cmd->ops.cmd.res = CPU_TO_LE16(((res_type << ICE_AQC_RES_TYPE_S) > & > >+ ICE_AQC_RES_TYPE_M) | (res_shared ? > >+ ICE_AQC_RES_TYPE_FLAG_SHARED : 0)); > >+ cmd->ops.cmd.first_desc = CPU_TO_LE16(*desc_id); > >+ ret = ice_aq_send_cmd(hw, &aq_desc, resp_buf, buf_size, cd); > >+ if (!ret) { > >+ *desc_id = LE16_TO_CPU(cmd->ops.resp.next_desc); > >+ *num_prof = LE16_TO_CPU(cmd->ops.resp.num_desc); > >+ } > >+ return ret; > >+} > >+ > >+static enum ice_status > >+ice_get_hw_res(struct ice_hw *hw, uint16_t type, uint16_t num, uint16_t > *res, > >+ uint16_t *prof_buf, uint16_t *num_prof) { > >+ struct ice_aqc_get_allocd_res_desc_resp *resp_buf; > >+ int ret; > >+ uint16_t buf_len; > >+ > >+ buf_len = sizeof(*resp_buf) + sizeof(resp_buf->elem) * (num - 1); > > why (num - 1) here? > > >+ resp_buf = (struct ice_aqc_get_allocd_res_desc_resp *) > >+ ice_malloc(hw, buf_len); > > Cast is unnecessary for void * > > >+ if (!resp_buf) > >+ return -ENOMEM; > >+ > >+ ret = ice_aq_get_res_profs(hw, num, resp_buf, buf_len, type, 1, > >+ res, NULL, num_prof); > >+ if (ret) > >+ goto exit; > >+ > >+ ice_memcpy(prof_buf, resp_buf->elem, sizeof(resp_buf->elem) * > >+ (*num_prof), ICE_NONDMA_TO_NONDMA); > >+exit: > >+ rte_free(resp_buf); > >+ return ret; > >+} > >+ > >+static int > >+ice_clearup_resource(struct ice_hw *hw, uint16_t res_type) > > s/clearup/cleanup > > >+{ > >+ int ret; > >+ uint16_t prof_id; > >+ uint16_t prof_buf[ICE_MAX_RES_DESC_NUM]; > >+ uint16_t first_desc = 1; > >+ uint16_t num_prof = 0; > >+ > >+ ret = ice_get_hw_res(hw, res_type, ICE_MAX_RES_DESC_NUM, > >+ &first_desc, prof_buf, &num_prof); > > What is purpose of first_desc? Seems there is no reference of it afterwards. > > >+ if (ret) { > >+ PMD_INIT_LOG(ERR, "Failed to get fxp resource"); > >+ return ret; > >+ } > >+ > >+ for (prof_id = 0; prof_id < num_prof; prof_id++) { > >+ ret = ice_free_hw_res(hw, res_type, 1, &prof_buf[prof_id]); > >+ if (ret) { > >+ PMD_INIT_LOG(ERR, "Failed to free fxp resource"); > >+ return ret; > >+ } > >+ } > >+ > >+ return 0; > >+} > >+ > >+static int > >+ice_reset_fxp_resource(struct ice_hw *hw) { > >+ int ret; > >+ > >+ ret = ice_clearup_resource(hw, > ICE_AQC_RES_TYPE_FD_PROF_BLDR_PROFID); > >+ if (ret) { > >+ PMD_INIT_LOG(ERR, "Failed to clearup fdir resource"); > >+ return ret; > >+ } > >+ > >+ ret = ice_clearup_resource(hw, > ICE_AQC_RES_TYPE_HASH_PROF_BLDR_PROFID); > >+ if (ret) { > >+ PMD_INIT_LOG(ERR, "Failed to clearup rss resource"); > >+ return ret; > >+ } > >+ > >+ return 0; > >+} > > s/clearup/cleanup > > Thanks, > Xiaolong > > >+ > > static int > > ice_dev_init(struct rte_eth_dev *dev) > > { > >@@ -1983,6 +2092,12 @@ ice_dev_init(struct rte_eth_dev *dev) > > return ret; > > } > > > >+ ret = ice_reset_fxp_resource(hw); > >+ if (ret) { > >+ PMD_INIT_LOG(ERR, "Failed to reset fxp resource"); > >+ return ret; > >+ } > >+ > > return 0; > > > > err_pf_setup: > >-- > >2.15.1 > > ^ permalink raw reply [flat|nested] 5+ messages in thread
* [dpdk-dev] [PATCH v2] net/ice: cleanup RSS/FDIR profile when device init 2019-09-27 18:32 [dpdk-dev] [PATCH] net/ice: cleanup RSS/FDIR profile when device init Ying Wang 2019-09-30 9:25 ` Ye Xiaolong @ 2019-10-17 18:31 ` Ying Wang 2019-10-18 6:57 ` Ye Xiaolong 1 sibling, 1 reply; 5+ messages in thread From: Ying Wang @ 2019-10-17 18:31 UTC (permalink / raw) To: xiaolong.ye, qi.z.zhang; +Cc: dev, qiming.yang, ying.a.wang, stable When app restart, RSS/FDIR profile ID will increase which may run out of all the profile resources. The patch cleanup RSS/FDIR profile resources when device init to fix this issue. Fixes: d7d150b93070 ("net/ice: enable RSS when device init") Cc: stable@dpdk.org Signed-off-by: Ying Wang <ying.a.wang@intel.com> --- --- v2: Modify patch according to comments. --- drivers/net/ice/ice_ethdev.c | 94 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 94 insertions(+) diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c index dc4d1e6c7..0117e43ae 100644 --- a/drivers/net/ice/ice_ethdev.c +++ b/drivers/net/ice/ice_ethdev.c @@ -13,6 +13,7 @@ #include "base/ice_sched.h" #include "base/ice_flow.h" #include "base/ice_dcb.h" +#include "base/ice_common.h" #include "ice_ethdev.h" #include "ice_rxtx.h" #include "ice_generic_flow.h" @@ -40,6 +41,7 @@ static const char * const ice_valid_args[] = { #define ICE_OS_DEFAULT_PKG_NAME "ICE OS Default Package" #define ICE_COMMS_PKG_NAME "ICE COMMS Package" #define ICE_MAX_PKG_FILENAME_SIZE 256 +#define ICE_MAX_RES_DESC_NUM 1024 int ice_logtype_init; int ice_logtype_driver; @@ -1882,6 +1884,92 @@ ice_vsi_config_sw_lldp(struct ice_vsi *vsi, bool on) return ret; } +static enum ice_status +ice_get_hw_res(struct ice_hw *hw, uint16_t res_type, + uint16_t num, uint16_t desc_id, + uint16_t *prof_buf, uint16_t *num_prof) +{ + struct ice_aqc_get_allocd_res_desc_resp *resp_buf; + int ret; + uint16_t buf_len; + bool res_shared = 1; + struct ice_aq_desc aq_desc; + struct ice_sq_cd *cd = NULL; + struct ice_aqc_get_allocd_res_desc *cmd = + &aq_desc.params.get_res_desc; + + buf_len = sizeof(resp_buf->elem) * num; + resp_buf = ice_malloc(hw, buf_len); + if (!resp_buf) + return -ENOMEM; + + ice_fill_dflt_direct_cmd_desc(&aq_desc, + ice_aqc_opc_get_allocd_res_desc); + + cmd->ops.cmd.res = CPU_TO_LE16(((res_type << ICE_AQC_RES_TYPE_S) & + ICE_AQC_RES_TYPE_M) | (res_shared ? + ICE_AQC_RES_TYPE_FLAG_SHARED : 0)); + cmd->ops.cmd.first_desc = CPU_TO_LE16(desc_id); + + ret = ice_aq_send_cmd(hw, &aq_desc, resp_buf, buf_len, cd); + if (!ret) + *num_prof = LE16_TO_CPU(cmd->ops.resp.num_desc); + else + goto exit; + + ice_memcpy(prof_buf, resp_buf->elem, sizeof(resp_buf->elem) * + (*num_prof), ICE_NONDMA_TO_NONDMA); + +exit: + rte_free(resp_buf); + return ret; +} +static int +ice_cleanup_resource(struct ice_hw *hw, uint16_t res_type) +{ + int ret; + uint16_t prof_id; + uint16_t prof_buf[ICE_MAX_RES_DESC_NUM]; + uint16_t first_desc = 1; + uint16_t num_prof = 0; + + ret = ice_get_hw_res(hw, res_type, ICE_MAX_RES_DESC_NUM, + first_desc, prof_buf, &num_prof); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to get fxp resource"); + return ret; + } + + for (prof_id = 0; prof_id < num_prof; prof_id++) { + ret = ice_free_hw_res(hw, res_type, 1, &prof_buf[prof_id]); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to free fxp resource"); + return ret; + } + } + return 0; +} + +static int +ice_reset_fxp_resource(struct ice_hw *hw) +{ + int ret; + + ret = ice_cleanup_resource(hw, ICE_AQC_RES_TYPE_FD_PROF_BLDR_PROFID); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to clearup fdir resource"); + return ret; + } + + ret = ice_cleanup_resource(hw, ICE_AQC_RES_TYPE_HASH_PROF_BLDR_PROFID); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to clearup rss resource"); + return ret; + } + + return 0; +} + static int ice_dev_init(struct rte_eth_dev *dev) { @@ -2019,6 +2107,12 @@ ice_dev_init(struct rte_eth_dev *dev) return ret; } + ret = ice_reset_fxp_resource(hw); + if (ret) { + PMD_INIT_LOG(ERR, "Failed to reset fxp resource"); + return ret; + } + return 0; err_pf_setup: -- 2.15.1 ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [dpdk-dev] [PATCH v2] net/ice: cleanup RSS/FDIR profile when device init 2019-10-17 18:31 ` [dpdk-dev] [PATCH v2] " Ying Wang @ 2019-10-18 6:57 ` Ye Xiaolong 0 siblings, 0 replies; 5+ messages in thread From: Ye Xiaolong @ 2019-10-18 6:57 UTC (permalink / raw) To: Ying Wang; +Cc: qi.z.zhang, dev, qiming.yang, stable On 10/18, Ying Wang wrote: >When app restart, RSS/FDIR profile ID will increase which may >run out of all the profile resources. > >The patch cleanup RSS/FDIR profile resources when device init >to fix this issue. > >Fixes: d7d150b93070 ("net/ice: enable RSS when device init") >Cc: stable@dpdk.org > >Signed-off-by: Ying Wang <ying.a.wang@intel.com> >--- > >--- >v2: Modify patch according to comments. Minor nit, please be more specific about your changes rather than the general words. >--- > drivers/net/ice/ice_ethdev.c | 94 ++++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 94 insertions(+) > Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> Applied to dpdk-next-net-intel. Thanks, Xiaolong ^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2019-10-18 6:59 UTC | newest] Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2019-09-27 18:32 [dpdk-dev] [PATCH] net/ice: cleanup RSS/FDIR profile when device init Ying Wang 2019-09-30 9:25 ` Ye Xiaolong 2019-10-18 0:42 ` Wang, Ying A 2019-10-17 18:31 ` [dpdk-dev] [PATCH v2] " Ying Wang 2019-10-18 6:57 ` Ye Xiaolong
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).