Q: raw disk devices?

Harald Milz (hm@seneca.muc.de)
Sat, 19 Apr 1997 18:29:07 +0200

From: hm@seneca.muc.de (Harald Milz)
Subject: Q: raw disk devices?
To: linux-kernel@vger.rutgers.edu
Organization: Linux.DE
X-PGP-Fingerprint: 44 5B 4E 65 50 F5 99 2B 20 55 C1 B0 56 FC C4 00
X-Nospam: Unsolicited ads are unwanted and will be proofread for $199/hr.


I have a question concerning raw (i.e. unbuffered) disk devices. An
iX magazine article (issue 5/97) says there are no raw devices which
make sure data is sychronously written to disk. However when looking
at the kernel code, O_SYNC is commented as not yet being implemented
(/usr/src/linux/fs/ioctl.c) but deeper in the code data seems to actually
be sync'ly written (fs/block_dev.c). Martin v. Loewis suggested this.
What exactly is true?

The question is whether databases requiring raw devices can really be
accomodated by the sync blockwrite or not.

Linus, can you please shed some light on this? Thanks!


Harald Milz ----------- PGP Public Key: finger hm[AT]muc.de (Key ID 7ADC4839)

"Really ?? What a coincidence, I'm shallow too!!"