tech-kern archive

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

Re: Autoloading of pseudo-device modules

Reinoud Zandijk writes:
> Hi folks,
> On Wed, Jan 14, 2009 at 09:19:08AM -0600, Greg Oster wrote:
> > Adam Hamsik writes:
> > > * I'm really not sure if raid diff is ok because radiframe code is  
> > > hard to understand.
> > 
> > The RAIDframe code has a number of issues which will require fixing 
> > before it can become a module... For example, raidattach() is only 
> > designed to be called exactly once, and then only very early on in 
> > the boot process...  As well, this code is expected to have been run 
> > before any particular RAID device is configured...  
> well, won't/can't the RAID bootloader autoload this module if its detected?

I'm sure the bootloader will do the right things... but I know right 
now that the RAIDframe code will have problems.  It may be the case 
that RAIDframe will work if / is a RAID set and the module is never 
unloaded.  In short, however, the RAIDframe code was never designed to 
work as a module, and so it will need fixing in order to make it work in 
all cases.. (e.g. after module unloading and reloading, a situation 
which is not going to work at all in RAIDframe right now..)


Greg Oster

Home | Main Index | Thread Index | Old Index