Re: [PATCH v6 3/8] btrfs: add a bio argument to btrfs_csum_one_bio
From: Daniel Vacek
Date: Tue Nov 18 2025 - 10:51:43 EST
On Tue, 18 Nov 2025 at 16:08, Christoph Hellwig <hch@xxxxxxxxxxxxx> wrote:
> On Tue, Nov 18, 2025 at 03:05:25PM +0100, Daniel Vacek wrote:
> > But giving it another thought and checking the related fscrypt code,
> > the encrypted bio is allocated in blk_crypto_fallback_encrypt_bio()
> > and freed in blk_crypto_fallback_encrypt_endio() before calling
> > bio_endio() on our original plaintext bio.
>
> That code is getting major refactoring right now, and allowing the
> file system to hook into the submission is a possibility.
>
> The problem is that I have no idea what you're trying to do as the
> context is missing.
>
> In general prep series should be self contained and at least borderline
> useful by themselves. Adding random dead code checks or weird arguments
> as done here are not useful in a prep series without context, they
> should be close to the code making use of them to be understandable.
Yeah, agreed. It would be better to send this one together with the
fscrypt changes later. I was suggested by Dave to pick this one ahead
and I did not argue while I probably should have.
--nX
On Tue, 18 Nov 2025 at 16:08, Christoph Hellwig <hch@xxxxxxxxxxxxx> wrote:
>
> On Tue, Nov 18, 2025 at 03:05:25PM +0100, Daniel Vacek wrote:
> > But giving it another thought and checking the related fscrypt code,
> > the encrypted bio is allocated in blk_crypto_fallback_encrypt_bio()
> > and freed in blk_crypto_fallback_encrypt_endio() before calling
> > bio_endio() on our original plaintext bio.
>
> That code is getting major refactoring right now, and allowing the
> file system to hook into the submission is a possibility.
>
> The problem is that I have no idea what you're trying to do as the
> context is missing.
>
> In general prep serious should be self contained and at least borderline
> useful by themselves. Adding random dead code checks or weird arguments
> as done here are not useful in a prep series without context, they
> should be close to the code making use of them to be understandable.