Hi Ulf, On Tue, Apr 19, 2022 at 8:15 PM Ulf Hansson <ulf.hansson@xxxxxxxxxx> wrote: > We can move this into mmc_decode_cid() instead, to avoid open coding. > Moreover, this would make it work for SDIO/SD combo cards too, > whatever that matters. Making it work in more places sounds good. > One thing though, what is the typical execution time to do this? > Probably negligible, but as this may be a card that holds the rootfs, > it could delay the boot to be completed. Extremely negligible. Basically the only place you shouldn't call add_device_randomness() is from a hard IRQ handler, but other than that, go wild. It's actually used all over the boot sequence pretty liberally. Grep around a bit and you'll see all the odd drivers it's called from. And perhaps in the process you'll notice a few other places that might benefit from calling it. In general, if you have a strong sense of niche hardware things where you notice that a certain identifier or field is "probably quasi-unique-ish", an initiative like the one Linus took with this patch seems like a positive thing to me. Jason
