RFC: Metadata caching and block device "hangup"

From: Simon Richter (Simon.Richter@phobos.fachschaften.tu-muenchen.de)
Date: Sun Jun 18 2000 - 10:24:38 EST


Hi,

In a sub-thread of "Floppy handling", there was the idea that
removing/swaping removable media wasn't a bad thing as long as
 - no metadata is cached by the filesystem
 - disk change events invalidate all buffers associated with that device
 - the user doesn't remove the media while there is I/O in progress.

Items 2 and 3 can be considered given. What is interesting is the current
status on 1. How much data is cached by the various filesystems and not
properly invalidated on a disk change? Would it make sense to add a
"expect disk changes" mount option?

Another thing rms pointed out was the issue of open files and pending
mmap()s. Is there a standards-compliant way to revoke file handles and
mmap()s? (Keep in mind glibc implements fopen() via mmap()) Would it be an
option to invent a new signal for that (default behaviour: kill app)?

   Simon

-- 
PGP public key available from http://phobos.fs.tum.de/pgp/Simon.Richter.asc
 Fingerprint: 10 62 F6 F5 C0 5D 9E D8  47 05 1B 8A 22 E5 4E C1
Hi! I'm a .signature virus! Copy me into your ~/.signature to help me spread!

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Fri Jun 23 2000 - 21:00:15 EST