search

dglo / dor-driver

Test upload of Mantis issues
0 stars 0 forks source link

[jkelley on 2014-01-14 17:32:52] : Unhandled / spurious interrupts #82

Open dglo opened 3 years ago

dglo commented 3 years ago

Running moat14 on dor-driver V02-13-09 or pre-release V02-13-10 results in a few unhandled interrupt messages on PCTS (upgraded SL6.3 x86_64 hub with new SBC). Not sure if this is normal behavior or not.

Additional Information: Jan 14 07:45:25 pcts-hub02 kernel: dh_intr_hdlr: FYI: INTR0 (value at handler entry) 0xfa2157f9 Jan 14 07:45:25 pcts-hub02 kernel: INTR1 (value just after ttsic, rtsic read) 0x200 Jan 14 07:45:25 pcts-hub02 kernel: INTR2 (value before this print statement) 0x200 Jan 14 07:45:25 pcts-hub02 kernel: GSTAT 0x17fd2, DSTAT 0x10f3, TTSIC 0xfefe, RTSIC 0x101fe Jan 14 08:16:49 pcts-hub02 kernel: dh_intr_hdlr: unhandled, spurious/inconsistent interrupt condition on card 1. Proceeding undaunted... call John and/or Kalle if you get this often. Jan 14 08:16:49 pcts-hub02 kernel: dh_intr_hdlr: FYI: INTR0 (value at handler entry) 0x57cbdf8a Jan 14 08:16:49 pcts-hub02 kernel: INTR1 (value just after ttsic, rtsic read) 0x200 Jan 14 08:16:49 pcts-hub02 kernel: INTR2 (value before this print statement) 0x200 Jan 14 08:16:49 pcts-hub02 kernel: GSTAT 0x179d2, DSTAT 0x10f3, TTSIC 0xfefe, RTSIC 0x10000ff Jan 14 08:17:12 pcts-hub02 kernel: dh_intr_hdlr: unhandled, spurious/inconsistent interrupt condition on card 1. Proceeding undaunted... call John and/or Kalle if you get this often. Jan 14 08:17:12 pcts-hub02 kernel: dh_intr_hdlr: FYI: INTR0 (value at handler entry) 0xf66ffa90 Jan 14 08:17:12 pcts-hub02 kernel: INTR1 (value just after ttsic, rtsic read) 0x200 Jan 14 08:17:12 pcts-hub02 kernel: INTR2 (value before this print statement) 0x200 Jan 14 08:17:12 pcts-hub02 kernel: GSTAT 0x179d2, DSTAT 0x10f3, TTSIC 0xfefe, RTSIC 0x10000ff Jan 14 08:35:26 pcts-hub02 kernel: dh_intr_hdlr: unhandled, spurious/inconsistent interrupt condition on card 0. Proceeding undaunted... call John and/or Kalle if you get this often.

dglo commented 3 years ago

[jkelley on 2014-01-14 17:55:02] Wow -- the INTR0 read value is garbage. It should be equal to INTR1 but with GLOBAL_INT enabled.

dglo commented 3 years ago

[jkelley on 2014-01-22 16:36:28] In V02-13-10, this condition is detected and the INTEN register is restored with the shadow version inside the driver. This doesn't address the root cause, however.