Re: [PATCH v6 0/4] hide ->i_state behind accessors

[Dave Chinner <david@fromorbit.com>]

On Tue, Sep 23, 2025 at 12:47:06PM +0200, Mateusz Guzik wrote:
> First commit message quoted verbatim with rationable + API:
> 
> [quote]
> Open-coded accesses prevent asserting they are done correctly. One
> obvious aspect is locking, but significantly more can checked. For
> example it can be detected when the code is clearing flags which are
> already missing, or is setting flags when it is illegal (e.g., I_FREEING
> when ->i_count > 0).
> 
> Given the late stage of the release cycle this patchset only aims to
> hide access, it does not provide any of the checks.
> 
> Consumers can be trivially converted. Suppose flags I_A and I_B are to
> be handled, then:
> 
> state = inode->i_state          => state = inode_state_read(inode)
> inode->i_state |= (I_A | I_B)   => inode_state_set(inode, I_A | I_B)
> inode->i_state &= ~(I_A | I_B)  => inode_state_clear(inode, I_A | I_B)
> inode->i_state = I_A | I_B      => inode_state_assign(inode, I_A | I_B)
> [/quote]
> 
> Right now this is one big NOP, except for READ_ONCE/WRITE_ONCE for every access.
> 
> Given this, I decided to not submit any per-fs patches. Instead, the
> conversion is done in 2 parts: coccinelle and whatever which was missed.
> 
> Generated against:
> https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git/commit/?h=vfs-6.18.inode.refcount.preliminaries

Much simpler and nicer than the earlier versions. Looks good.

Reviewed-by: Dave Chinner <dchinner@redhat.com>

-- 
Dave Chinner
david@fromorbit.com