On Thu, Dec 5, 2024 at 2:08 PM Max Kellermann max.kellermann@ionos.com wrote:
On Thu, Dec 5, 2024 at 1:57 PM Alex Markuze amarkuze@redhat.com wrote:
I will explain the process for ceph client patches. It's important to note: The process itself and part of the automation is still evolving and so many things have to be done manually.
None of this answers any of my questions on your negative review comments.
I will break it up into three separate patches, as it addresses three different issues.
... one of which will be my patch.
It looks like Alex's preference is to address these memory leaks by passing true for own_pages to osd_req_op_extent_osd_data_pages() and adjusting where the OSD request is put accordingly -- this is what he folded in his WIP patch which was posted in Luis' thread [1] the day after this patch and pushed to the testing branch earlier today [2]. Unfortunately an update to this thread was missed, leaving everyone including myself confused.
Max, would you be willing to redo this patch to pass true for own_pages and post a v2? There is nothing "bad", "partial" or otherwise wrong with this version, but having the pages be taken care of automatically is a bit nicer and a conflict with Alex's ongoing work would be avoided.
[1] https://lore.kernel.org/ceph-devel/CAO8a2ShzHuTizjY+T+RNr28XLGOJPNoXJf1rQUbu... [2] https://github.com/ceph/ceph-client/commit/2a802a906f9c89f8ae492dbfcd82ff412...
Thanks,
Ilya