net/ice: check Rx queue number on RSS init
PATCH 1/1 net/ice: check Rx queue number on RSS init
From: YU DAPENG <[email protected]>
When RSS is initialized, rx queues number is used as denominator to set default value into the RSS lookup table. If it is zero, there will be error of being divided by 0. So add value check to avoid the error.
--- drivers/net/ice/ice_ethdev.c | 6 ++++++ 1 file changed, 6 insertions(+)
drivers/net/ice/ice_ethdev.c +6 -0
@@ -3182,6 +3182,12 @@ static int ice_init_rss(struct ice_pf *pf) vsi->rss_key_size =ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE ; vsi->rss_lut_size = pf->hash_lut_size; +if (nb_q ==0 ) { +PMD_DRV_LOG (WARNING , +"RSS is not supported as rx queues number is zero\n" ); +return 0 ; + } +if (is_safe_mode) {PMD_DRV_LOG (WARNING ,"RSS is not supported in safe mode\n" );return 0 ;
Signature · 1 line
2.27.0
Hi, Dapeng
Regards,
Chen Bo
Quoted 28 lines
-----Original Message-----
From: dev <[email protected]> On Behalf Of [email protected]
Sent: December 23, 2020 13:30
To: Yang, Qiming <[email protected]>; Zhang, Qi Z <[email protected]>
Cc: [email protected]; Yu, DapengX <[email protected]>; [email protected]
Subject: [dpdk-dev] [PATCH] net/ice: check Rx queue number on RSS initFrom: YU DAPENG <[email protected]>
When RSS is initialized, rx queues number is used as denominator to set default value into the RSS lookup table. If it is zero, there will be error of being divided by 0. So add value check to avoid the error.
Fixes 50370662b727 ("net/ice: support device and queue ops")Signed-off-by YU DAPENG <[email protected]>--- drivers/net/ice/ice_ethdev.c | 6 ++++++ 1 file changed, 6 insertions(+)drivers/net/ice/ice_ethdev.c +0
-0@@ -3182,6 +3182,12 @@ static int ice_init_rss(struct ice_pf *pf) vsi->rss_key_size =
drivers/net/ice/ice_ethdev.c static int ice_init_rss(struct ice_pf *pf) · line 3192
Quoted 10 lines
drivers/net/ice/ice_ethdev.c +6
-0ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE ; vsi->rss_lut_size = pf->hash_lut_size; +if (nb_q ==0 ) { +PMD_DRV_LOG (WARNING , +"RSS is not supported as rx queues number is zero\n"); +return 0 ; + } +
Direct return here will introduce subsequent exceptions, the tetpmd will exit.
testpmd> port start all
Configuring Port 0 (socket 1)ice_init_rss(): RSS is not supported as rx queues number is zero
Port 0: 68:20:20:06:01:00
Configuring Port 1 (socket 1)ice_init_rss(): RSS is not supported as rx queues number is zero
Port 1: 68:20:20:06:01:01
Checking link statuses...
Done
testpmd>
Port 1: link state change event
testpmd> show config rxtx
io packet forwarding packets/burst=32
nb forwarding cores=1 - nb forwarding ports=2
port 0: RX queue number: 0 Tx queue number: 1
Rx offloads=0x0 Tx offloads=0x10000
Invalid RX queue_id=0
RX queue: 0
RX desc=0 - RX free threshold=32
RX threshold registers: pthresh=8 hthresh=8 wthresh=0 RX Offloads=0x0
TX queue: 0
TX desc=1024 - TX free threshold=32
TX threshold registers: pthresh=32 hthresh=0 wthresh=0 TX offloads=0x10000 - TX RS bit threshold=32
port 1: RX queue number: 0 Tx queue number: 1
Rx offloads=0x0 Tx offloads=0x10000
Invalid RX queue_id=0
RX queue: 0
RX desc=0 - RX free threshold=32
RX threshold registers: pthresh=8 hthresh=8 wthresh=0 RX Offloads=0x0
TX queue: 0
TX desc=1024 - TX free threshold=32
TX threshold registers: pthresh=32 hthresh=0 wthresh=0 TX offloads=0x10000 - TX RS bit threshold=32
testpmd>
testpmd> start
EAL: Error - exiting with code: 1
Cause: Either rxq or txq are 0, cannot use io fwd mode
Quoted 6 lines
drivers/net/ice/ice_ethdev.c +0
-0if (is_safe_mode) {PMD_DRV_LOG(WARNING, "RSS is not supported in safe
mode\n");
return 0;
--
2.27.0
Hi Chen Bo,
Without this patch, I have reproduced your unexpected test result on XXV710(with i40e pmd driver).
That means the unexpected situation as the below is a new issue different from the original one that this patch try to resolve. The unexpected is not caused by the patch.
The unexpected situation can be reproduced in this way, at the end, testpmd exit with error, and the terminal cannot echo keyboard input.
testpmd> port stop all
testpmd> port config all rxq 0
testpmd> port config all txq 0
testpmd> show config rxtx
testpmd> port start all
testpmd> show config rxtx
testpmd> start
EAL: Error - exiting with code: 1
Cause: Either rxq or txq are 0, cannot use io fwd mode
-----Original Message-----
From: Chen, BoX C
Sent: Wednesday, December 23, 2020 2:23 PM
To: Yu, DapengX <[email protected]>; Yang, Qiming <[email protected]>; Zhang, Qi Z <[email protected]>
Cc: [email protected]; Yu, DapengX <[email protected]>; [email protected]
Subject: RE: [dpdk-dev] [PATCH] net/ice: check Rx queue number on RSS init
Hi, Dapeng
Regards,
Chen Bo
Quoted 27 lines
-----Original Message-----
From: dev <[email protected]> On Behalf Of [email protected]
Sent: December 23, 2020 13:30
To: Yang, Qiming <[email protected]>; Zhang, Qi Z <[email protected]>
Cc: [email protected]; Yu, DapengX <[email protected]>; [email protected]
Subject: [dpdk-dev] [PATCH] net/ice: check Rx queue number on RSS initFrom: YU DAPENG <[email protected]>
When RSS is initialized, rx queues number is used as denominator to set default value into the RSS lookup table. If it is zero, there will be error of being divided by 0. So add value check to avoid the error.
Fixes 50370662b727 ("net/ice: support device and queue ops")Signed-off-by YU DAPENG <[email protected]>--- drivers/net/ice/ice_ethdev.c | 6 ++++++ 1 file changed, 6 insertions(+)(quoted diff) +0
-0
b/drivers/net/ice/ice_ethdev.cindex9a5d6a559..bbb8c1460100644
drivers/net/ice/ice_ethdev.c static int ice_init_rss(struct ice_pf *pf) · line 3191
Quoted 10 lines
drivers/net/ice/ice_ethdev.c +6
-0@@ -3182,6 +3182,12 @@ static int ice_init_rss(struct ice_pf *pf) vsi->rss_key_size =ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE ; vsi->rss_lut_size = pf->hash_lut_size; +if (nb_q ==0 ) { +PMD_DRV_LOG (WARNING , +"RSS is not supported as rx queues number is zero\n"); +return 0 ; +} +
Direct return here will introduce subsequent exceptions, the tetpmd will exit.
testpmd> port start all
Configuring Port 0 (socket 1)ice_init_rss(): RSS is not supported as rx queues number is zero
Port 0: 68:20:20:06:01:00
Configuring Port 1 (socket 1)ice_init_rss(): RSS is not supported as rx queues number is zero
Port 1: 68:20:20:06:01:01
Checking link statuses...
Done
testpmd>
Port 1: link state change event
testpmd> show config rxtx
io packet forwarding packets/burst=32
nb forwarding cores=1 - nb forwarding ports=2
port 0: RX queue number: 0 Tx queue number: 1
Rx offloads=0x0 Tx offloads=0x10000
Invalid RX queue_id=0
RX queue: 0
RX desc=0 - RX free threshold=32
RX threshold registers: pthresh=8 hthresh=8 wthresh=0 RX Offloads=0x0
TX queue: 0
TX desc=1024 - TX free threshold=32
TX threshold registers: pthresh=32 hthresh=0 wthresh=0 TX offloads=0x10000 - TX RS bit threshold=32
port 1: RX queue number: 0 Tx queue number: 1
Rx offloads=0x0 Tx offloads=0x10000
Invalid RX queue_id=0
RX queue: 0
RX desc=0 - RX free threshold=32
RX threshold registers: pthresh=8 hthresh=8 wthresh=0 RX Offloads=0x0
TX queue: 0
TX desc=1024 - TX free threshold=32
TX threshold registers: pthresh=32 hthresh=0 wthresh=0 TX offloads=0x10000 - TX RS bit threshold=32
testpmd>
testpmd> start
EAL: Error - exiting with code: 1
Cause: Either rxq or txq are 0, cannot use io fwd mode
Quoted 5 lines
drivers/net/ice/ice_ethdev.c +0
-0if (is_safe_mode) {PMD_DRV_LOG (WARNING ,"RSS is not supported in safe mode\n" );return 0;
--
2.27.0
Quoted 11 lines
-----Original Message-----
From: [email protected] <[email protected]>
Sent: Wednesday, December 23, 2020 1:30 PM
To: Yang, Qiming <[email protected]>; Zhang, Qi Z <[email protected]>
Cc: [email protected]; Yu, DapengX <[email protected]>; [email protected]
Subject: [PATCH] net/ice: check Rx queue number on RSS initFrom: YU DAPENG <[email protected]>
Quoted 8 lines
When RSS is initialized, rx queues number is used as denominator to set default value into the RSS lookup table. If it is zero, there will be error of being divided by 0. So add value check to avoid the error.
Fixes 50370662b727 ("net/ice: support device and queue ops")Signed-off-by YU DAPENG <[email protected]>
Applied to dpdk-next-net-intel.
Thanks
Qi