On Thu, Jul 07, 2005 at 02:30:10AM -0700, Steve Langasek wrote: > AIUI, this bug is the same as bug #234787. Is that correct? If so, > this bug is now sarge-only. Are there any plans to apply that fix for > sarge?
There seems to be at least two problems here: one is that fam did not stop monitoring the mounted CF card before the unmount, and the other is that Gnome desktop incorrectly reported the CF card as being unmounted. The first problem could be the same one as in #234787, but the second sounds really like a Gnome bug. There is no package version number in the bug report, but it does look like the submitter is using testing/unstable, which means his fam program has already been patched for #234787. fam is quite a mess with unresponsive upstream, and it would not surprise me that there are other bugs lurking in the program. Here is my take on what should happen: For the short term, getting the Gnome folks to migrate to gamin (a drop-in replacement of fam) may be the best choice. According to the author of gamin, not supporting FAMSuspendMonitor(), FAMResumeMonitor(), and FAMMonitorCollection() simplifies system design quite a bit, so gamin should be less buggy than fam. For the long term, we need to solve the "monitoring interferes with unmounting" problem at a more fundamental level; otherwise something will always go wrong and prevent unmounting. Unfortunately this interference is the result of limitations of the kernel dnotify API, and the problem can be completely solved only with the new ionotify API, which has not been accepted into the Linux kernel yet. gamin supports ionotify in patched kernels, but fam only understands dnotify. To summarize, we are kind of screwed, because there is not much we can do to help the poor users. In any case, I will start talking to the Gnome team about the migration to gamin. -- Chuan-kai Lin http://www.cs.pdx.edu/~cklin/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]