Hi, Thanks for the reply. Some more questions inline ... Alan Cox wrote: > On Wed, Jan 07, 2004 at 04:46:12PM -0800, Yu-Shun Wang wrote: > >>... >>With a bit of digging and luck (not debugging, I am new to >>kernel debugging and don't have the environment setup to >>generate tracebacks), I managed to find the probable cause: >>sys/i386/i386/pmap.c from the commit log included below. >>After reversed the diff between 1.456 and 1.457, the problem >>was fixed. > Please restore this diff and try increasing UMA_BOOT_PAGES. Don't be > overly conservative; try doubling UMA_BOOT_PAGES. Will do. The PCs are in my office, so will have to wait till tomorrow. Just to clarify, you meant changing the value directly in sys/vm/uma_int.h, or is it something I can set as some tunables or hints? >>Weird thing is, it seems like I am the only one hit by >>this problem. I couldn't find any message on CURRENT >>list regarding this. Anyways, dmesg & kernel conf files are >>also attached. Please let me know if you need more info. > I suspect it is related to the number of modules that you loading. Hmm, now that you mentioned, it's a bit much. The previous sys admin (just graduated & left) set those up, they are now under my care. :-) AFAIK, we've been running with these modules and tracking current for quite a while without this problem. I am wondering what changed; or maybe we've been close to the edage and were just lucky? Is there a way to find out? Thanks again, I'll let you know if it works. yushun. > Regards, > Alan -- Yu-Shun Wang <yushunwa_at_isi.edu> USC Information Sciences Institute
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:37 UTC