Unable to handle kernel NULL pointer dereference at virtual address 0000000000000048

clarke_liu_f — Thu, 18 Apr 2024 09:19:14 GMT

Hi NXP:

I want to use dpdmux in dpaa2 platform.I refer to section 9.2.12.3 Traffic bifurcation using DPDMUX on DPAA2 of LSDK21.08.According to our own requirements, I made some modifications compared to the example.After successfully creating dpdmux, I found a fatal problem:When I ifconfig up/down the Ethernet interface(eth_X2) on another dpmac(dpmac.4) multiple times, the kernel mem abort.

The full log is as follows:

-sh-5.0# dmesg |grep 2160 [ 0.000000] Machine model: NXP Layerscape LX2160ARDB [ 0.643838] Machine: NXP Layerscape LX2160ARDB [ 0.648335] SoC family: QorIQ LX2160A [ 1.990384] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 3.090361] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 4.190360] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 5.290368] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 6.390360] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 7.490360] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 8.590368] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 9.690361] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 10.790360] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 11.894367] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 12.994360] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 14.094360] Hardware name: NXP Layerscape LX2160ARDB (DT) -sh-5.0# restool -v restool v2.3 (commit lf-5.10.52-2.1.0) -sh-5.0# restool -m MC firmware version: 10.29.1 -sh-5.0# uname -a Linux localhost 5.10.35-rt39 #3 SMP PREEMPT_RT Tue Feb 27 12:20:44 UTC 2024 aarch64 GNU/Linux -sh-5.0# -sh-5.0# -sh-5.0# ls-listmac dprc.1/dpmac.4 (end point: dpni.1) dprc.1/dpmac.3 (end point: dpni.0) -sh-5.0# ls-listni dprc.1/dpni.1 (interface: eth_X2, end point: dpmac.4) dprc.1/dpni.0 (interface: eth1, end point: dpmac.3) -sh-5.0# -sh-5.0# echo "dpni.0" > /sys/bus/fsl-mc/drivers/fsl_dpaa2_eth/unbind -sh-5.0# restool dpni destroy "dpni.0" dpni.0 is destroyed -sh-5.0# -sh-5.0# export DPIO_COUNT=8 -sh-5.0# export DPMCP_COUNT=2 -sh-5.0# /usr/local/dpdk/dpaa2/dynamic_dpl.sh dpni parent - dprc.1 Creating Non nested DPRC NEW DPRCs dprc.1 dprc.2 Using board type as 2160 Using High Performance Buffers ##################### Container dprc.2 is created #################### Container dprc.2 have following resources :=> * 2 DPMCP * 16 DPBP * 8 DPCON * 16 DPSECI * 1 DPNI * 8 DPIO * 2 DPCI * 8 DPDMAI * 0 DPRTC ######################### Configured Interfaces ######################### Interface Name Endpoint Mac Address ============== ======== ================== dpni.0 UNCONNECTED 00:00:00:00:5:1 -sh-5.0# -sh-5.0# /usr/local/dpdk/dpaa2/dynamic_dpl.sh dpmac.3 parent - dprc.1 Creating Non nested DPRC NEW DPRCs dprc.1 dprc.3 dprc.2 Using board type as 2160 Using High Performance Buffers ##################### Container dprc.3 is created #################### Container dprc.3 have following resources :=> * 2 DPMCP * 16 DPBP * 8 DPCON * 16 DPSECI * 1 DPNI * 8 DPIO * 2 DPCI * 8 DPDMAI * 0 DPRTC ######################### Configured Interfaces ######################### Interface Name Endpoint Mac Address ============== ======== ================== dpni.2 dpmac.3 -Dynamic- -sh-5.0# -sh-5.0# ls-addni --no-link Created interface: eth0 (object:dpni.3, endpoint: ) -sh-5.0# -sh-5.0#echo dprc.3 > /sys/bus/fsl-mc/drivers/vfio-fsl-mc/unbind -sh-5.0#restool dprc disconnect dprc.3 --endpoint=dpni.2 -sh-5.0#restool dpdmux create --default-if=1 --num-ifs=2 --method DPDMUX_METHOD_CUSTOM --manip=DPDMUX_MANIP_NONE --option=DPDMUX_OPT_CLS_MASK_SUPPORT --container=dprc.1 dpdmux.0 is created under dprc.1 -sh-5.0#restool dprc connect dprc.1 --endpoint1=dpdmux.0.0 --endpoint2=dpmac.3 -sh-5.0#restool dprc connect dprc.1 --endpoint1=dpdmux.0.1 --endpoint2=dpni.3 -sh-5.0#restool dprc connect dprc.1 --endpoint1=dpdmux.0.2 --endpoint2=dpni.2 -sh-5.0#restool dprc assign dprc.1 --object=dpdmux.0 --child=dprc.3 --plugged=1 -sh-5.0#echo dprc.3 > /sys/bus/fsl-mc/drivers/vfio-fsl-mc/bind -sh-5.0# -sh-5.0# -sh-5.0# ls-listmac dprc.1/dpmac.4 (end point: dpni.1) dprc.1/dpmac.3 (end point: dpdmux.0.0) -sh-5.0# ls-listni dprc.1/dpni.3 (interface: eth0, end point: dpdmux.0.1) dprc.1/dpni.1 (interface: eth_X2, end point: dpmac.4) dprc.1/dprc.3/dpni.2 (end point: dpdmux.0.2) dprc.1/dprc.2/dpni.0 -sh-5.0# -sh-5.0# -sh-5.0# -sh-5.0# ifconfig eth_X2 down -sh-5.0# ifconfig eth_X2 up -sh-5.0# [ 693.918781] fsl_dpaa2_eth dpni.1 eth_X2: configuring for inband/sgmii link mode -sh-5.0# -sh-5.0# [ 694.958374] fsl_dpaa2_eth dpni.1 eth_X2: Link is Up - 100Mbps/Full - flow control rx/tx [ 694.958415] IPv6: ADDRCONF(NETDEV_CHANGE): eth_X2: link becomes ready [ 694.998536] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000048 [ 694.998542] Mem abort info: [ 694.998543] ESR = 0x96000004 [ 694.998545] EC = 0x25: DABT (current EL), IL = 32 bits [ 694.998547] SET = 0, FnV = 0 [ 694.998549] EA = 0, S1PTW = 0 [ 694.998550] Data abort info: [ 694.998551] ISV = 0, ISS = 0x00000004 [ 694.998552] CM = 0, WnR = 0 [ 694.998554] [0000000000000048] user address but active_mm is swapper [ 694.998556] Internal error: Oops: 96000004 [#1] PREEMPT_RT SMP [ 694.998559] Modules linked in: sig_sniffk(O) [ 694.998565] CPU: 13 PID: 2773 Comm: irq/138-dpio.0 Tainted: G W O 5.10.35-rt39 #3 [ 694.998570] Hardware name: NXP Layerscape LX2160ARDB (DT) [ 694.998572] pstate: 00000005 (nzcv daif -PAN -UAO -TCO BTYPE=--) [ 694.998575] pc : dpaa2_eth_poll+0xe8/0x380 [ 694.998586] lr : dpaa2_eth_poll+0x108/0x380 [ 694.998589] sp : ffff800016f5bb70 [ 694.998590] x29: ffff800016f5bb70 x28: ffff2fab6337f1c8 [ 694.998594] x27: ffff2fab6337f000 x26: 0000000000000000 [ 694.998597] x25: ffff2fab633c0a00 x24: 0000000000000000 [ 694.998600] x23: 0000000000000000 x22: 0000000000000000 [ 694.998603] x21: ffff2fab633c0a00 x20: 0000000000000001 [ 694.998606] x19: 000000000000000d x18: 000000000000000e [ 694.998608] x17: 0000000000000001 x16: 0000000000000019 [ 694.998611] x15: 0000000000000033 x14: 0000000000000142 [ 694.998614] x13: 0000000000000000 x12: 0000000000000000 [ 694.998617] x11: 00000000000965f1 x10: ffff6a56ce1c0000 [ 694.998620] x9 : ffff800016f5ba10 x8 : 0000000004ffffff [ 694.998623] x7 : ffff800016f80780 x6 : 000000000000000f [ 694.998625] x5 : 0000000000000001 x4 : 0000000000000031 [ 694.998628] x3 : 0000000000000000 x2 : ffff2fab63324020 [ 694.998631] x1 : ffff2fab6337f000 x0 : ffff2fab633c0a00 [ 694.998634] Call trace: [ 694.998636] dpaa2_eth_poll+0xe8/0x380 [ 694.998639] net_rx_action+0x118/0x378 [ 694.998644] __do_softirq+0x10c/0x210 [ 694.998648] __local_bh_enable_ip+0xcc/0x130 [ 694.998652] irq_forced_thread_fn+0x80/0xc0 [ 694.998657] irq_thread+0x150/0x260 [ 694.998660] kthread+0x174/0x198 [ 694.998665] ret_from_fork+0x10/0x34 [ 694.998670] Code: 11000694 aa1b03e1 aa1803e3 aa1503e0 (f9402705) [ 694.998674] ---[ end trace 000000000000000e ]---

Re: Unable to handle kernel NULL pointer dereference at virtual address 0000000000000048

yipingwang — Fri, 26 Apr 2024 09:58:02 GMT

Please use all the images(such as mc and restool) from LSDK 21.08 release.

Please try whether there is problem to use the default configuration descripted in LSDK 21.08 document on your target board.

topic Re: Unable to handle kernel NULL pointer dereference at virtual address 0000000000000048 in Layerscape

Unable to handle kernel NULL pointer dereference at virtual address 0000000000000048

Re: Unable to handle kernel NULL pointer dereference at virtual address 0000000000000048