From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <2461e98afb958fade826a06a71ae1d7a@centurytel.net> To: 9fans@cse.psu.edu Subject: Re: [9fans] problems with 8l From: "Skip Tavakkolian" MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Date: Mon, 12 Aug 2002 08:04:49 -0400 Topicbox-Message-UUID: dad0653a-eaca-11e9-9e20-41e7f4b1d025 Excellent deduction. I disabled the CPU external cache setting (apparently a default) and the linking problem has gone away. I was also having other stability problems (acme, and rio each crashed during the last week) that I expect this will fix. To paraphrase 'Wayne and Garth', I'm not worthy... Thanks. > actually, based on some of the previous postings, i was starting to wonder > whether you might have something wrong with the underlying system. > the marshal error, for instance, was in the following > d = dirfstat(fd); > if(d == nil) > return 0; > rv = strcmp(d->name, "cons") == 0; > where d->name was nil, which looked unlikely when i checked dirfstat. > > on an IBM Thinkpad T23, i found i needed to set maxmem to exclude > some peculiar NVRAM that's in the top of the address space, or i got > all manner of strange errors (with 3rd edition, and it's still set for 4th > so i'm unsure whether i still need it). as another example, a system with > an AGP card that relied on stealing processor memory had trouble when > the borrowed memory was also allocated by the operating system. > finally, a machine with faulty memory produced the usual assortment of > unpredictable results.