[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: WARNING: hdparm weirdness.



Hi,


> On Mon 04 Dec 2000, T. Weyergraf wrote:
> > 
> > I have found at least one other file in one of my FS, that has been corrupted as
> > well, but was totally unrelated to hdparm. I have some reason to believe, that the
> > hdparm-file is not the only one being modified. 
> 
> Have you verified that the corruption is still there after a reboot?
> Perhaps it's just the buffer pool for the disk getting nuked.

That's it. I just nukes the buffer cache. Well, *just* might be a bit weak here.
IIRC, there are some sublte bugs in the buffer cache as of 2.4.0-test11, which
are appearently triggered by hdparm.
A reboot *cleared* the problem - at least as far as i know. Can't yet reliably tell for files,
which were written to the buffer-cache, but not yet flushed - *sic*.

The lastest test12-pre seems to address that problem. 

Oh boy, I wish I wouldn't need 2.4.0-test<something> ;-)

Anyway, thanks.

> 
> 
> Paul Slootman
> -- 

Regards,
T. Weyergraf


-- 
Thomas Weyergraf                                                kirk@colinet.de
My Favorite IA64 Opcode-guess ( see arch/ia64/lib/memset.S )
"br.ret.spnt.few" - got back from getting beer, did not spend a lot.




Reply to: