On Tue, 2006-02-14 at 18:26 +0100, Arne Redlich wrote:
> Am Dienstag, den 14.02.2006, 11:53 -0500 schrieb Ming Zhang:
> > On Tue, 2006-02-14 at 13:54 +0100, Arne Redlich wrote:
> > > Am Dienstag, den 14.02.2006, 12:05 +0100 schrieb Damian Pietras:
> > > > This is the ReadOnly patch recently sent by Ming Zhang ported
> > > > to 0.4.12 (and 0.4.13). I've also added information about IOMode
> > > > to /proc/net/iet/volume.
> > > >
> > > > It was tested with Windows and Linux. Linux does not recognize the
> > > > device as write protected, this is a bug in the kernel and will be fixed
> > > > in the next version.
> > > >
> > > Just had a glimpse at it - "IOMode" shouldn't be specific to fileio, it
> > > should rather be independent of the iotype.
> >
> > agree. but where u have this impression?
>
> Because the IOMode token is only evaluated in parse_fileio_params().
i see. i did not read the patch carefully. the patch i gave Damian for
reference is to do this parsing this volume.c
>
> > also this patch still has that parameter parsing issue before it can be
> > accepted.
> >
>
> IIRC, the point I made above was the parameter parsing issue (or a
> variation of it), wasn't it? :P Need to dig it out and verify it,
> though.
yes, i think so, i forgot the detail. :P
>
> Anyway, besides some minor coding style issues, there's also this
> needless renaming of set_iotype() -> parse_volume_args() without actual
> changes to the function (or did I miss 'em?)
yes, i put the iomode parsing in volume.c set_iotype. so that is why i
changed the name to it. i do not know why Damian moved that parsing into
fileio.c.
attached is my original patch.
>
> Arne
r1184.ro.patch
Description: Text Data
|
