Skip to content

Add support for injecting, monitoring, and reporting on tofino's error interrupts#275

Open
Nieuwejaar wants to merge 2 commits into
mainfrom
interrupts
Open

Add support for injecting, monitoring, and reporting on tofino's error interrupts#275
Nieuwejaar wants to merge 2 commits into
mainfrom
interrupts

Conversation

@Nieuwejaar
Copy link
Copy Markdown
Collaborator

We can see dendrite detecting and reporting errors in response to tofino interrupts:

00:11:00.686Z DEBG dpd:   group stat for shadow TCAM ECC pipe: 1 mau: 19: 256
    unit = interrupt monitor
00:11:00.686Z ERRO dpd: single-bit error on TCAM [pipe: 1, mau: 19].  row: 4  addr: 0x16
    unit = interrupt monitor
00:11:00.686Z DEBG dpd: handled interrupt [pipe: 1, mau: 19]
    unit = interrupt monitor
00:11:08.464Z DEBG dpd:   group stat for shadow PCIe Errors: 32
    unit = interrupt monitor
00:11:08.465Z ERRO dpd: CPU request timeout detected
    unit = interrupt monitor
00:11:08.465Z DEBG dpd: handled interrupt PcieCpuTimeout

These two errors were triggered by the new intr tool running in the global zone:

BRM42220007 # ./intr inject tcam ecc -p 1 -m 19 -r 4 -a 22
writing sbe reg: TcamSbeErrlog(BitSet([22, 0, 0, 0])) at 0x59c0e10
writing inject reg IntrInjectMauTcamArray(BitSet([0, 1, 0, 0])) at 0x59c0f20
BRM42220007 # ./intr inject pcie ct
writing pci injection register PcieIntrInj(BitSet([32, 0, 0, 0])) at 0x14c
BRM42220007 #

@Nieuwejaar Nieuwejaar marked this pull request as draft May 22, 2026 21:36
@Nieuwejaar Nieuwejaar marked this pull request as ready for review May 23, 2026 10:56
@Nieuwejaar Nieuwejaar requested a review from rcgoodfellow May 23, 2026 10:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant