On Thu, Apr 16, 2020 at 5:45 PM Justin Gottula <justin@xxxxxxxxxxxx> wrote: > > 2. There is a bit of a predicament with the keep-going-even-if-an-error- > happened approach. [...] Oops. I meant to also include this in my reply: I suppose a reasonable way to address #2 might be to condense any error(s) arising from submit_bio_wait into one writeback_store error return value whose purpose is to indicate that the backing device itself had some sort of write problem(s).
