On Fri Aug 11, 2023 at 9:55 PM EEST, Linus Torvalds wrote: > On Fri, 11 Aug 2023 at 10:22, Jarkko Sakkinen <jarkko@xxxxxxxxxx> wrote: > > > > I was planning to send a PR to Linus with a quirk for MSI GS66 Stealth > > 11UG, and apparently this bug report would add two additional MSI > > entries. This is becoming quickly a maintenance hell. > > Honestly, what would be the immediate effects of just not enabling the > TPM irq by default at all, and making it an explicit opt-in? > > When a common solution is to just disable the TPM in the BIOS > entirely, and the end result is a working system, I really get the > teeling that this is all pain for very very little gain. > > Would anybody even notice if we just disabled it by default and added > a "if you really want it, use 'tpm=irq' kernel command line"? > > Hmm? > > Linus It would be in line with what I proposed in [1] (option A), and also what Grundik said in this thread. Right now IRQs should be only enabled by a command-line option, and what "allow list" would mean in practice should really be something that would be advertised e.g. Device Tree or ACPI, not really part of the solution implemented right now. I'll send a patch ASAP, which makes this feature opt-in. [1] https://lore.kernel.org/linux-integrity/CUPW0XP1RFXI.162GZ78E46TBJ@suppilovahvero/ BR, Jarkko
