Re: [GIT PULL] bcachefs fixes for 6.16-rc4

[Josef Bacik <josef@toxicpanda.com>]

On Thu, Jun 26, 2025 at 10:22:52PM -0400, Kent Overstreet wrote:
> per the maintainer thread discussion and precedent in xfs and btrfs
> for repair code in RCs, journal_rewind is again included
>

I'm replying to set the record straight. This is not the start of the
discussion. I am not going to let false statements stand by unchallenged
however.

Sterba has never sent large pull requests in RCs, certainly not with features in
them.  Even when Chris was the maintainer and we were a little faster and looser
and were pushing the envelope to see what Linus would accept we didn't ship
anything near this volume of patches past rc1.

And the numbers don't lie.

josef@fedora:~/linux$ git tag --contains 1c6fdbd8f2465ddfb73a01ec620cbf3d14044e1a | grep -v rc > bcachefs-tags
josef@fedora:~/linux$ git tag --contains be0e5c097fc206b863ce9fe6b3cfd6974b0110f4 | grep -v rc > tags
josef@fedora:~/linux$ for i in $(cat tags); do git log --no-merges --oneline $i-rc2..$i fs/btrfs | wc -l; done > btrfs-counts.txt
josef@fedora:~/linux$ for i in $(cat bcachefs-tags); do git log --no-merges --oneline $i-rc2..$i fs/bcachefs | wc -l; done > bcachefs-counts.txt

josef@fedora:~/linux$ R -q -e "x <- read.csv('btrfs-counts.txt', header = F); summary(x); sd(x[ , 1 ])"
> x <- read.csv('btrfs-counts.txt', header = F); summary(x); sd(x[ , 1 ])
       V1       
 Min.   : 0.00  
 1st Qu.:10.25  
 Median :19.00  
 Mean   :20.48  
 3rd Qu.:27.50  
 Max.   :55.00  
[1] 11.77108
> 
josef@fedora:~/linux$ R -q -e "x <- read.csv('bcachefs-counts.txt', header = F); summary(x); sd(x[ , 1 ])"
> x <- read.csv('bcachefs-counts.txt', header = F); summary(x); sd(x[ , 1 ])
       V1        
 Min.   :  0.00  
 1st Qu.: 38.50  
 Median : 70.00  
 Mean   : 63.86  
 3rd Qu.: 81.50  
 Max.   :137.00  
[1] 45.28218
> 

So even including the wilder times of kernel development in general and btrfs's
specifically, our worst window was 55 patches, less than your mean.

These are not the same thing. Do not equivicate the two.  Sterba is a phenomenal
maintainer who does his job well, manages to work with Linus just fine. We are
not the same, we do not work the same, and we absolutely do follow the rules, as
do 99.99% of the kernel community.

If xfs has done this then good for them. Those developers have a track record of
doing the right thing over a long period of time. Btrfs for sure hasn't. Thanks,

Josef