Re: Searching for maintainers: dual monitor patches, bdflush

Eric W. Biederman (ebiederm+eric@npwt.net)
19 Jun 1998 16:00:09 -0500


>>>>> "RR" == Rik van Riel <H.H.vanRiel@phys.uu.nl> writes:

RR> On Fri, 19 Jun 1998, Pavel Machek wrote:
>> > Bdflush is obsolete. You should adapt kflushd (the
>> > in-kernel bdflush) instead...
>>
>> Bdflush is not obsolete. Kflushd will only flush buffers to disk if
>> you are very low on memory. You still need something that flushes old
>> buffers to disk.

RR> Hmm, you seem to be right indeed :( Upon reading the
RR> code I saw, much to my dismay, that bdflush doesn't
RR> write out dirty buffers all by itself (unless I skipped
RR> a vital part of code :)...

update/bdflush whatever it is called should call
sys_bdflush(func = 1)
Which eventually calls sync_old_buffers.
Which could use a little bit of work to be jiffie wrap safe, and
int/long consistent, but is otherwise fine.

The process on my machine that does this is update and it is an old
trusted a.out process..

Why do you need to rewrite it?

Eric

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu