Re: [PATCH v6 00/15] Fast kernel headers: split linux/mm.h

[Max Kellermann]

On Tue, Jun 4, 2024 at 3:02 PM Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
> I am not a fan of these patches.  They will make it harder to work on
> the MM system.  We briefly discussed them at LSFMM and nobody was in
> favour of them.  I'm afraid you're shouting into the wind.

Thanks for letting me know, but I'm confused because similar patches
have been merged pretty often.

For a very weird example, look at commit adeb04362d74 ("kernel.h: Move
upper_*_bits() and lower_*_bits() to wordpart.h") which was submitted
on Feb 14th; and then look at
https://lore.kernel.org/lkml/20240209164027.2582906-34-max.kellermann@xxxxxxxxx/
- it's exactly the same patch, but I submitted it 5 days earlier, yet
the other one was merged.

Other recent examples (though without offending earlier patches from me):
- d186eb1ee885 ("cpumask: split out include/linux/cpumask_types.h")
- f7515d9fe8fc4 ("objtool: Add objtool_types.h")
- cb5a065b4ea9 ("headers/deps: mm: Split <linux/gfp_types.h> out of
<linux/gfp.h>")
- 50d91c765825 ("hrtimers: Split out hrtimer_types.h")
- 9983deb26d90 ("Split out irqflags_types.h")
- 6d5e9d636830 ("pid: Split out pid_types.h")
.... and so on ...

If there is a general agreement that nobody is in favor of these
patches, why are they merged when other people submit them?

Max