Re: [Ext2-devel] [RFC] [PATCH 2/4]delayed allocation for ext3

[Mingming Cao <cmm@us.ibm.com>]

On Sun, 2005-07-17 at 19:47 -0600, Andreas Dilger wrote:
> On Jul 17, 2005  10:40 -0700, Mingming Cao wrote:
> > @@ -373,6 +373,7 @@ struct ext3_inode {
> >  #define EXT3_MOUNT_BARRIER		0x20000 /* Use block barriers */
> >  #define EXT3_MOUNT_NOBH			0x40000 /* No bufferheads */
> >  #define EXT3_MOUNT_QUOTA		0x80000 /* Some quota option set */
> > + #define EXT3_MOUNT_DELAYED_ALLOC	0xC0000 /* Delayed Allocation */
> 
> This doesn't make sense.  DELAYED_ALLOC == QUOTA | NOBH?
> 


Ah...:-)  Badari used 0x80000 for DELAYED_ALLOC in the previous patch
(2.6.11 based).When moving those patches forward to 2.6.13-rc3, we found
the conflict with QUOTA, and obviously picked up a wrong value.

> > +     {Opt_delayed_alloc, "delalloc"},
> 
> Is this a replacement for Alex's delalloc code?  We also use delalloc for
> that code and if they are not interchangeable it will cause confusion
> about which one is in use.
> 

Okey, will think a new name for this feature to avoid confusion.  Alex's
delalloc is bond to extent tree structure so it's hard to be adopted
directly to the current ext3 layout, so, I'd say this work done by
Badari(inspired by Alex's work) is a different implementation.

> > +     if (test_opt(sb, DELAYED_ALLOC)) {
> > +             if (!(test_opt(sb, DATA_FLAGS) == EXT3_MOUNT_WRITEBACK_DATA)) {
> > +                     printk(KERN_WARNING "EXT3-fs: Ignoring delall option - "
> > +                             "its supported only with writeback mode\n");
> 
> Should be "ignoring delalloc option".
>  
Fixed. 


Thanks for looking at this.

Mingming


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/