On 7/24/24 10:36, Bjorn Helgaas wrote: > On Wed, Jul 24, 2024 at 03:45:34PM +0100, Alejandro Lucero Palau wrote: >> On 7/17/24 21:55, Wei Huang wrote: >>> TLP headers with incorrect steering tags (e.g. caused by buggy driver) >>> can potentially cause issues when the system hardware consumes the tags. >>> Provide a kernel option, with related helper functions, to completely >>> prevent TPH from being enabled. >> >> Maybe rephrase it for including a potential buggy device, including the cpu. >> >> Also, what about handling this with a no-tph-allow device list instead of a >> generic binary option for the whole system? >> >> Foreseeing some buggy or poor-performance implementations, or specific use >> cases where it could be counterproductive, maybe supporting both options. > > Makes sense if/when we need it. IMO no point in adding an empty list > of known-broken devices. We can add quirks.c, or something similar later, after TPH support is enabled.