Rafal Boni wrote:
Rafal is right, though I'm not sure exactly what to do about it. After Mr. Kiyohara's last message, I tried to duplicate his set up, but didn't know exactly what version he had downloaded. Based on the date of his diff file, I fetched the 20080821 version of NetBSD, but couldn't build tools because of a problem with cross-compiling libc. Nor would 20080822 or 20080823. Fetching from 20080824, I found that the diff file wouldn't work on all files. The problem with libc was due to a file checked in on 8/20, so I went back to 8/19. But here, again, the diff patch didn't work exactly right, so I manually fetched the three files that were a problem (three dev/pci files), built everything and it all works (or at least seems to). Ahh - life with -current.KIYOHARA Takashi wrote:Hi! Donald, From: Donald T Hayford <don%donhayford.com@localhost> Date: Sun, 07 Sep 2008 07:01:33 -0400There is probably some functionality missing, though I'm not sure what. The system doesn't seem all that stable, and has hung up a few times doing relatively benign things such as:"ls /dev | grep sd0".I think that the cause of hangup is arm. Because I hang up by both KUROBOX_PRO and GUMSTIX.I noticed that at least from the dmesg of Don's machine that his ORION-NAS kernel was based on 4.99.66 TNF sources, which made me think the hangups might be to port-arm/38950. There have also been a pile of pmap changes since then, some of them causing short-term problems. I found the mergeup of TNF changes from ~ 4.99.66 to 4.99.72 (where my tree is currently frozen due to lack of hacking time) in my private evbarm tree helped stability greatly for relatively little effort.--rafal
I suppose if I knew more about CVS than "cvs checkout -D xxxxxxxxxx-UTC src", this would be easier. :-)
Incidentally, dmesg now shows: NetBSD 4.99.72 (KUROBOX_PRO) #0: Sun Sep 7 19:03:23 EDT 2008so I think I'll follow Rafal's advice and freeze where I am for the time being.
Regards, Don