Source-Changes archive

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

Re: CVS commit: src/sys/miscfs/genfs



On Wed, Nov 30, 2005 at 04:51:41PM +0100, Reinoud Zandijk wrote:
> On Wed, Nov 30, 2005 at 03:45:16AM +0000, YAMAMOTO Takashi wrote:
> > Modified Files:
> >     src/sys/miscfs/genfs: genfs_vnops.c
> > 
> > Log Message:
> > revert rev.1.111 as it isn't necessary or correct.
> > - currently no one in tree has a problem with zero b_lblkno, afaik.
> > - this buf is used for "devvp", so it doesn't make sense to
> >   use lbn in the "vp".
> 
> i'll have to disagree. Some filingsystems like ntfs, arla and others that 
> do 1:1 translation in their vop_bmap function DONT pass their "devvp" but 
> their "vp" and thus it makes sense for them. As a compromise we could 
> adjust the value of b_lblkno the same way as b_blkno is adjusted. In that 
> way it would allways point to the correct lbn whatever scheme or method is 
> used.

As I understand it, these file systems work fine now. Thus they do NOT 
NEED such a change.

The only reason I see for this change was that it makes more sense for 
you. I don't think that's a good reason for this change given that the 
b_blkno == and != b_lblkno cases have different semantic meanings.

Feel free to change b_lblkno as you see fit once you are in your own 
strategy routine. When you get into your routine, b_lblkno is all yours.

Take care,

Bill

Attachment: pgp8mip6eqAy_.pgp
Description: PGP signature



Home | Main Index | Thread Index | Old Index