Re: [PATCH] nilfs2: depending on flags, update segment usage instead of cleaning

[Andreas Rohner <andreas.rohner-hi6Y0CQ0nG0@public.gmane.org>]

Hi Ryusuke,

On 2014-01-19 17:49, Ryusuke Konishi wrote:
> Could you consider adding NILFS_IOCTL_SET_SUINFO instead of extending
> v_flags of NILFS_IOCTL_CLEAN_SEGMENTS ?

Yes sure. I actually considered that writing the patch, but then shyed
away from adding a new ioctl.

> It is hacky to extend NILFS_IOCTL_CLEAN_SEGMENTS like this, and,
> unfortunately, argv[]->v_flags of NILFS_IOCTL_CLEAN_SEGMENTS is not
> zero-filled in the current library implementation.  This is our
> mistake (so I will fix it soon), but we cannot use these flags for
> some time.  Otherwise, existing cleanerds will go wrong when this is
> merged into kernel.

Ah yes I didn't think of that.

> Presence of ioctls can be tested with ENOTTY error, so libnilfs
> can know whether nilfs in underlying kernel has NILFS_IOCTL_SET_SUINFO
> ioctl or not, and we can extend the library keeping compatibility
> by using this nature.
> 
> A good example of code updating metadata file is
> nilfs_ioctl_change_cpmode() even though NILFS_IOCTL_SET_SUINFO will
> need nilfs_ioctl_wrap_copy().  It would be helpful for you.
> 
> Additional comments are as follows:
> 
> - For NILFS_IOCTL_SET_SUINFO, v_flags should be used to define which
>   fields (lastmod, nblocks, flags) are modified.  These flags should
>   be defined with bit masks.

Thank you for your comments. I will try and implement it and come back
with a new version of my patch.

Best regards,
Andreas Rohner

--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html