On Mon, Feb 6, 2023 at 4:51 PM Uwe Kleine-König <u.kleine-koenig@xxxxxxxxxxxxxx> wrote: > > Hello, > > ah, this is the mail I missed before. > > On Wed, Jan 25, 2023 at 11:11:59PM +0100, Wolfram Sang wrote: > > > > > So, this code handled all my stress-testing well so far. I'll try to > > > think of some more ideas until this evening, but likely I will apply it > > > later. Nonetheless, more review eyes are still welcome! > > > > Ah yes, I now recalled why I had the gut feeling that this solution is > > not complete. See this mail thread from 2015: > > > > https://lkml.iu.edu/hypermail/linux/kernel/1501.2/01700.html > > > > There are still drivers using i2c_del_adapter()+kfree(), so removing the > > completion could cause use-after-free there, or? > > There is also a strange construct in spi that I understand at one point > in time, but I failed to swap it in quickly. It's about commit > 794aaf01444d4e765e2b067cba01cc69c1c68ed9. I think there should be a > nicer solution than to track if the controller was allocated using devm, > but I don't remember the details. But before addressing the i2c problem > it might be worth to invest some time into that spi issue to not make > the same mistake for i2c. > Yeah, I've seen these constructs before elsewhere... Sadly, we have workarounds upon workarounds within workarounds chased by other workarounds due to this issue. Bart
