public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* Flushing buffer and page cache
@ 2001-02-17 10:39 Florian Weimer
  0 siblings, 0 replies; 3+ messages in thread
From: Florian Weimer @ 2001-02-17 10:39 UTC (permalink / raw)
  To: linux-kernel

Is it possible to flush all entries in the buffer cache corresponding
to a single block device (i.e. simply drop them if they aren't dirty,
or write them to disk and drop them after this if they are dirty)?

I've got another device in my SCSI chain which writes to the disk, and
if the caches are not flushed, the computer won't see the updates.
(Synchronization is done manually, so it's not an issue---trust me, I
know what I'm doing. ;-)

Kernel version doesn't matter. ;-)

-- 
Florian Weimer 	                  Florian.Weimer@RUS.Uni-Stuttgart.DE
University of Stuttgart           http://cert.uni-stuttgart.de/
RUS-CERT                          +49-711-685-5973/fax +49-711-685-5898

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Flushing buffer and page cache
@ 2001-02-17 16:12 James Bottomley
  0 siblings, 0 replies; 3+ messages in thread
From: James Bottomley @ 2001-02-17 16:12 UTC (permalink / raw)
  To: Florian Weimer; +Cc: linux-kernel

> Is it possible to flush all entries in the buffer cache corresponding
> to a single block device (i.e. simply drop them if they aren't dirty,
> or write them to disk and drop them after this if they are dirty)?

Yes, just send the BLKFLSBUF ioctl to the device this syncs the device then 
removes all the buffers from the cache.  We use it as a tool to move a SAN 
device around a cluster, which is similar to what you want to do.

James Bottomley




^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Flushing buffer and page cache
@ 2001-02-17 18:24 Douglas Gilbert
  0 siblings, 0 replies; 3+ messages in thread
From: Douglas Gilbert @ 2001-02-17 18:24 UTC (permalink / raw)
  To: linux-kernel

James Bottomley wrote:

> > Is it possible to flush all entries in the buffer cache corresponding
> > to a single block device (i.e. simply drop them if they aren't dirty,
> > or write them to disk and drop them after this if they are dirty)?
> 
> Yes, just send the BLKFLSBUF ioctl to the device this syncs the device then 
> removes all the buffers from the cache.  We use it as a tool to move a SAN 
> device around a cluster, which is similar to what you want to do.

Last time this question was raised, someone mentioned
a little utility called flushb . Here is its source:

/*
 * flushb.c --- This routine flushes the disk buffers for a disk
 */

/*
 * modified August 2000 by Juri Haberland
 * juri.haberland@innominate.de
 */

#include <stdio.h>
#include <fcntl.h>
#include <linux/fs.h>

#define NOARGS void

const char *progname;

static void usage(NOARGS)
{
        fprintf(stderr, "Usage: %s disk\n", progname);
        exit(1);
}      

int main(int argc, char **argv)
{
        int     fd;

        progname = argv[0];
        if (argc != 2)
                usage();

        fd = open(argv[1], O_RDONLY, 0);
        if (fd < 0) {
                perror("open");
                exit(1);
        }
        /*
         * Note: to reread the partition table, use the ioctl
         * BLKRRPART instead of BLKFSLBUF.
         */
        if (ioctl(fd, BLKFLSBUF, 0) < 0) {
                perror("ioctl BLKFLSBUF");
                exit(1);
        }
        return 0;
}

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2001-02-17 18:30 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2001-02-17 18:24 Flushing buffer and page cache Douglas Gilbert
  -- strict thread matches above, loose matches on Subject: below --
2001-02-17 16:12 James Bottomley
2001-02-17 10:39 Florian Weimer

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox