On Tue, Jul 07, 2020 at 04:36:32PM -0600, Alan Somers wrote: > Yeah, just apply the patch and run the geli tests. The patch add one > additional geli test, too. IIUC, geli doesn't use the thread pool when > it's using hardware crypto, which is why I think it's important that change > be tested both ways. >From my reading of the code that isn't the case, but I might be missing something. r226840 looks like it might be relevant. > The change is here: > https://reviews.freebsd.org/D25587 Works fine using the hardware driver. I had to patch the new test per my comments in the review. > -Alan > > On Tue, Jul 7, 2020 at 4:24 PM Mark Johnston <markj_at_freebsd.org> wrote: > > > On Tue, Jul 07, 2020 at 03:18:02PM -0600, alan somers wrote: > > > I'm working on a change to geli, but I need to test it on a system with a > > > hardware crypto card. Not merely one with AES-NI, but one of the old > > > fashioned PCI kind that works with crypto(9). Would anybody who runs > > > CURRENT on such a machine be willing to test my patch for me? > > > > I'm finishing work on a new hardware crypto driver and can test if you > > like. Do you want to run the existing GELI tests with the patch > > applied, or something else? > > > > I don't quite understand why GELI would care about the distinction, > > though. Doesn't GELI use AES-NI via crypto(9) as well? > >Received on Tue Jul 07 2020 - 21:25:06 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:24 UTC