On 23/09/2024 04:33, Christoph Hellwig wrote:
On Mon, Sep 23, 2024 at 12:57:32PM +1000, Dave Chinner wrote:
Ok, but that's not going to be widespread. Very little storage
hardware out there supports atomic writes - the vast majority of
deployments will be new hardware that will have mkfs run on it.
Just about every enterprise NVMe SSD supports atomic write size
larger than a single LBA, because it is completely natural fallout
from FTL deѕign. That beeing said to support those SSDs a block
size of 16 or 32k would be a lot more natural than all the forcealign
madness.
Outside the block allocator changes, most changes for forcealign are
just refactoring the RT big alloc unit checks. So - as you have said
previously - this so-called madness is already there. How can the sanity
be improved?
To me, yes, there are so many "if (RT)" checks and special cases in the
code, which makes a maintenance headache.
