Re: [GIT PULL] bcachefs fixes for 6.12-rc5

[Kees Cook <kees@kernel.org>]

On Tue, Oct 22, 2024 at 03:06:38PM -0400, Sasha Levin wrote:
> On Tue, Oct 22, 2024 at 01:39:10PM -0400, Kent Overstreet wrote:
> > 
> > The following changes since commit 5e3b72324d32629fa013f86657308f3dbc1115e1:
> > 
> >  bcachefs: Fix sysfs warning in fstests generic/730,731 (2024-10-14 05:43:01 -0400)
> > 
> > are available in the Git repository at:
> > 
> >  https://github.com/koverstreet/bcachefs tags/bcachefs-2024-10-22
> 
> Hi Linus,
> 
> There was a sub-thread on the linus-next discussion around improving
> telemetry around -next/lore w.r.t soaking time and mailing list reviews
> (https://lore.kernel.org/all/792F4759-EA33-48B8-9AD0-FA14FA69E86E@kernel.org/).
> 
> I've prototyped a set of scripts based on suggestions in the thread, and
> wanted to see if you'd find it useful. A great way to test it out is with
> a random pull request you'd review anyway :)

This looks really nice to me! Maybe add top-level grade ("B-") based on
the stats available. You can call it the Nominal Acceptance Grade bot
(NAGbot). Can so that everyone will pay attention, declare that it is an
LLM (Large Linus Model).

> Is the below useful in any way? Or do you already do something like this
> locally and I'm just wasting your time?
> 
> If it's useful, is bot reply to PRs the best way to share this? Any
> other information that would be useful?
> 
> Here it goes:
> 
> 
> Days in -next:
> ----------------------------------------
>  0  | ███████████ (5)
>  1  |
>  2  | █████████████████████████████████████████████████ (21)
>  3  |
>  4  |
>  5  |
>  6  |
>  7  |
>  8  |
>  9  |
> 10  |
> 11  |
> 12  |
> 13  |
> 14+ |
> 
> Commits that didn't spend time in -next:

I'd include a count summary "(5 of 26: 19%)"

> --------------------
> a069f014797fd bcachefs: Set bch_inode_unpacked.bi_snapshot in old inode path
> e04ee8608914d bcachefs: Mark more errors as AUTOFIX
> f0d3302073e60 bcachefs: Workaround for kvmalloc() not supporting > INT_MAX allocations
> 3956ff8bc2f39 bcachefs: Don't use wait_event_interruptible() in recovery
> eb5db64c45709 bcachefs: Fix __bch2_fsck_err() warning

And then maybe limit this to 5 or 10 (imagine a huge PR like netdev or
drm).

> 
> 
> Commits that weren't found on lore.kernel.org/all:

"(9 of 26: 35%)"

> --------------------
> e04ee8608914d bcachefs: Mark more errors as AUTOFIX
> f0d3302073e60 bcachefs: Workaround for kvmalloc() not supporting > INT_MAX allocations
> bc6d2d10418e1 bcachefs: fsck: Improve hash_check_key()
> dc96656b20eb6 bcachefs: bch2_hash_set_or_get_in_snapshot()
> 15a3836c8ed7b bcachefs: Repair mismatches in inode hash seed, type
> d8e879377ffb3 bcachefs: Add hash seed, type to inode_to_text()
> 78cf0ae636a55 bcachefs: INODE_STR_HASH() for bch_inode_unpacked
> b96f8cd3870a1 bcachefs: Run in-kernel offline fsck without ratelimit errors
> 4007bbb203a0c bcachefS: ec: fix data type on stripe deletion

Nice work!

-- 
Kees Cook