Re: [patch resend v4] update ctime and mtime for mmaped write

[Andrew Morton]

On 27 Mar 2007 16:09:33 -0400
linux@xxxxxxxxxxx wrote:

> What part of "The msync() function writes all modified data to
> permanent storage locations [...] For mappings to files, the msync()
> function ensures that all write operations are completed as defined
> for synchronised I/O data integrity completion." suggests that it's not
> supposed to do disk I/O?  How is that uselessly vague?
> 

Because for MS_ASYNC, "msync() shall return immediately once all the write
operations are initiated or queued for servicing".

ie: the writes can complete one millisecond or one week later.  We chose 30
seconds.

And this is not completely fatuous - before 2.6.17, MAP_SHARED pages could
float about in memory in a dirty state for arbitrarily long periods -
potentially for the entire application lifetime.  It was quite reasonable
for our MS_ASYNC implementation to do what it did: tell the VM about the
dirtiness of these pages so they get written back soon.

Post-2.6.17 we preserved that behaviour.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/