I've been deploying the 2.6.0 mm-sources in gentoo 1.4 on a small number of laptops that I admin and I've come into a serious problem using the latter test versions.
Up to and including 2.6.0-test9-mm2 things have been running perfectly. However, as of 2.6.0-test9-mm4 up to 2.6.0-test10-mm1 I've got a serious issue with the compiled kernels. It seems the latter versions of the mm-sources aren't mounting the root partitions in read write mode, even though mount shows it as being mounted rw... Quite disturbing. This is turning up on 6 laptops of various brand and architecture. What happens when I boot into one of the offending kernels, the boot process begins as normal, however root isn't mounted in read/write mode. Since the processes being started up can't write to the hard drive, they fail. If I want to get things back to normal, I have to wait till the login prompt, then login and remount root in rw mode. Once done, I can restart those processes that failed to startup and preocede as normal. I can get it all to work ok that way, but something is very amiss in the 2.6.0-test9-mm4 and later sources.. If I take the vanilla (virgin) code off of kernel.org, they will compile and boot as expected sans the mm-patches. No problems with root being mounted in rw... So it's NOT in the kernel source code. I can compile the 2.6.0-mm2 source over and over on the boxes demonstraiting the mount problem and never have a problem with that version of the mm-patch level... And later versions compile ok, but fail to boot as expected. So, is anyone else seeing anything amiss in the later mm-sources? Or am I alone in this one? -- [EMAIL PROTECTED] mailing list
