On Sat, Oct 1, 2022 at 9:32 PM Bernhard Reutner-Fischer <rep.dot....@gmail.com> wrote: > > On Sat, 1 Oct 2022 21:06:48 +0200 > "Jason A. Donenfeld" <ja...@zx2c4.com> wrote: > > > On Fri, Sep 30, 2022 at 04:05:20PM +0200, Jason A. Donenfeld wrote: > > > With the kernel patch linked below, Linux ingests a RNG seed > > > passed from the hypervisor. So, pass this for the Malta platform, and > > > reinitialize it on reboot too, so that it's always fresh. > > > > > > Link: > > > https://lore.kernel.org/linux-mips/20220930140138.575751-1-ja...@zx2c4.com/ > > > > The kernel side of this has now landed, so we can move ahead on the QEMU > > side: > > https://git.kernel.org/pub/scm/linux/kernel/git/mips/linux.git/commit/?id=056a68cea01edfa78b3474af1bfa39cc6bcc7bee > > > > s/a RNG/an RNG/ > > What about rngseed=""? > len=min(4711,0) > hex2bin(..0) will return false so ok. > rndseed="0" is problably fine, but is it worthy and desired? 00, 0x0. > Other than that, sounds plausible. IMHO. > thanks,
Not sure I understand the substantive part of your message. You're wondering whether it's okay to ingest length 1 (or one half?) inputs? The kernel will take whatever the firmware gives it; that's by design. The firmware in turn provides whatever it can, optimally 32 bytes as QEMU does with this patch. Maybe you could use some more words to describe what your thoughts are? Jason
