Michael Mann
changed
bug 3209
Comment # 3
on bug 3209
from Michael Mann
There are some nasty global variables trying to keep track of request/response
of the LUNs. The LUN is only populated in a FCP_CMND (request) packet, but
both the FCP_DATA and FCP_RSP (response) packets try to use it. This can lead
to odd/wrong behavior based on the parsing order of the packets.
It also appears that fragmentation isn't being done properly (but that's
independent of this bug)
Alexandre,
Would it be possible to keep the trace file private, but make the bug itself
public? It would give more visibility to people more familiar with these
protocols (who would presumably have trace files of their own to use to debug)
I may try to take a shot at a fix, but someone with more familarity with the
protocols may have better luck. It still looks like spaghetti to me at the
moment. I stripped the trace down to only the first 15 packets (which makes
step through debugging easier), but out of respect to Alexandre keeping the
trace private, I will not post the smaller trace file.
You are receiving this mail because:
- You are the assignee for the bug.
- You are watching all bug changes.
