This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

SK-AM62B-P1: Host timestamp cause CPSW port to lockup

Part Number: SK-AM62B-P1

Tool/software:

We are doing development on mentioned started kit. During that we found issue of CPSW port lock up. logs are as below.

:~# [  186.017319] ------------[ cut here ]------------
[  186.021985] NETDEV WATCHDOG: eth0 (am65-cpsw-nuss): transmit queue 7 timed out 6720 ms
[  186.029954] WARNING: CPU: 0 PID: 0 at /net/sched/sch_generic.c:525 dev_watchdog+0x1f0/0x21c
[  186.038309] Modules linked in: overlay xt_conntrack nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 cfg80211 ipt_REJECT nf_reject_ipv4 xt_tcpudp iptable_filter iptable_mangle iptable_raw ip_tables x_tables rpmsg_ctrl rpmsg_char irq_pruss_intc pru_rproc crct10dif_ce btrfs blake2b_generic libcrc32c rti_wdt xor xor_neon ti_k3_m4_remoteproc ti_k3_r5_remoteproc rtc_ti_k3 k3_j72xx_bandgap sa2ul tps6598x omap_mailbox at24 mcrc64 authenc pruss optee_rng rng_core raid6_pq drm drm_panel_orientation_quirks backlight ipv6
[  186.083113] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.6.44-ti-01377-g617526782e26-dirty #1
[  186.091535] Hardware name: 
[  186.095966] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[  186.102916] pc : dev_watchdog+0x1f0/0x21c
[  186.106916] lr : dev_watchdog+0x1f0/0x21c
[  186.110916] sp : ffff800080003de0
[  186.114218] x29: ffff800080003de0 x28: ffff800080fb7000 x27: 0000000000000101
[  186.121345] x26: ffff800080003e90 x25: 0000000000000007 x24: ffff00000297839c
[  186.128471] x23: 0000000000001a40 x22: 0000000000000007 x21: ffff000002978448
[  186.135597] x20: ffff0000029798c0 x19: ffff000002978000 x18: 000000000000000a
[  186.142723] x17: 6f2064656d697420 x16: 3720657565757120 x15: 74696d736e617274
[  186.149849] x14: 203a297373756e2d x13: 736d203032373620 x12: 74756f2064656d69
[  186.156976] x11: 00000000000001a9 x10: 00000000ffffffea x9 : 000000000000000c
[  186.164101] x8 : ffff000077b49b88 x7 : ffff000077b49b88 x6 : ffff000077b560b0
[  186.171228] x5 : ffff800080949dd8 x4 : ffffffffffffffff x3 : 0000000000000001
[  186.178353] x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffff800080fc3400
[  186.185480] Call trace:
[  186.187918]  dev_watchdog+0x1f0/0x21c
[  186.191572]  call_timer_fn+0x70/0x108
[  186.195233]  __run_timers+0x1ec/0x218
[  186.198891]  run_timer_softirq+0x20/0x40
[  186.202807]  handle_softirqs+0x1fc/0x28c

And then errors like:

[  926.113361] am65-cpsw-nuss 8000000.ethernet eth0: txq:3 DRV_XOFF:0 tmo:478816 dql_avail:-36 free_desc:362
[  931.233359] am65-cpsw-nuss 8000000.ethernet eth0: txq:3 DRV_XOFF:0 tmo:483936 dql_avail:-36 free_desc:362
[  937.121369] am65-cpsw-nuss 8000000.ethernet eth0: txq:3 DRV_XOFF:0 tmo:489824 dql_avail:-36 free_desc:362
[  942.241364] am65-cpsw-nuss 8000000.ethernet eth0: txq:3 DRV_XOFF:0 tmo:494944 dql_avail:-36 free_desc:362

It is just after 1 minute of board get boot. I also found issue in errata as i2401.

But I want to check is this issue resolved in any new silicon revision?

  • What are you trying to do before you get the above issues? Please attach the commands run to replicate this. Also can you clarify what SW are you running, the TI SDK as such or did you build your own?

    But I want to check is this issue resolved in any new silicon revision?

    No there is no planned AM62x version to fix i2401 errata. Linux has a workaround for this errata i2401 (reverts back to fifo for IEEE1588 timestamps, instead of timestamping all packets) for a couple years. Based on your SW version being Linux 6.6 kernel above you should have the workaround.

      Pekka