[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Re-strange networking be
- To: "info-mcl" <firstname.lastname@example.org>
- Subject: Re: Re-strange networking be
- From: "Dave Wroblewski" <email@example.com>
- Date: 26 Feb 1993 08:43:51 U
- Cc: firstname.lastname@example.org, email@example.com, "bug-mcl" <firstname.lastname@example.org>
Reply to: RE>Re:strange networking behav
>>The problem occuring with MacTCP.lisp with Quadras from within the event
>>handler can be solved by adding: (FORMAT T ".") (FORCE-OUTPUT) in the body
>>of the while-loop of the function %TCP-CONTROL in MacTCP.lisp.
>>Writing on the standard-output seems to enable the receive of device-
>>interrupts. This seems very odd to me but it works..
>>The problem that I would called "Inference between Device-Interrupt and
>>Virtual Memory" occurs with the builtin Ethernet of Quadras, MacTCP 1.1,
>>within the event-handler, with Virtual Memory Off, when waiting for device-
>>Have you tried to open-tcp-stream twice sequencially? this will also hang
>>your machine. A solution that seems also very odd to me is to (SLEEP 3)
>>between the 2 open-tcp-stream (2 seconds would work only 90% of the case)...
Gosh, talk about serendipity...
I too am having puzzling problems with MacTCP code that works well on a IIfx
but locks up a Quadra. We see two behavours: First, after a few "opens" the
quadra locks up, although never at exactly the same point, but seemingly always
within the tcp code (and I think always within %tcp-control.). Second, there
extreme performance delays when sending & receiving, such that opening a
connection, sending a few bytes, and closing the connection takes 6-7 times
as long on the Quadra as on the IIfx. Since our code depends on quickly opening
closing connections for short messages, tricks like sleeps are a problem.
We've spent the last 2 days trying to isolate the problem, thinking this was
some obscure dependency in our code,init conflict, etc., but if this is a known
bug I'd appreciate some comment on info-mcl from the wonderful folks at apple
cambridge so that I don't spin my wheels too much.
The relevant details are that we are running MCL 2.0, no patches, under
One machine (the happy one) is a IIfx, the other (the unhappy one) is a Quadra
U S WEST Advanced Technologies