Message ID | 20190517023924.1686-1-richard.henderson@linaro.org |
---|---|
Headers | show |
Series | Add qemu_getrandom and ARMv8.5-RNG etc | expand |
On Thu, May 16, 2019 at 07:38:59PM -0700, Richard Henderson wrote: > The change in v8 is to which objects are linked in to user-only > from the crypto directory. Daniel asked for all of crypto-obj-y > to be added, but I have been unable to find a solution that works. > > If I add crypto-obj-y, then authz-obj-y must be included to resolve > dependencies. Daniel suggested splitting authz-obj-y into two, so > that linux-user plus some of the tools need not link against libpam. > > However, I tried that, and in the process managed to break testing. > I'm not really sure what I did wrong: > > TEST check-speed: tests/benchmark-crypto-cipher > ERROR - too few tests run (expected 32, got 0) > > Or maybe it was broken before, but at least this way I'm not touching > any of the variables that affect tests/Makefile.include. > > Given that user-only *is* being linked against the enabled crypto libs, > the behaviour between static and non-static is identical, which is I > believe the major portion of Daniel's request. I think further cleanup > to the makefiles can be done separately. > For avoidance of any doubt Acked-by: Daniel P. Berrangé <berrange@redhat.com> for Richard to send a pull request with any of the crypto related patches I'm tagged as maintainer for. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
On 22/05/2019 12:58, Daniel P. Berrangé wrote: > On Thu, May 16, 2019 at 07:38:59PM -0700, Richard Henderson wrote: >> The change in v8 is to which objects are linked in to user-only >> from the crypto directory. Daniel asked for all of crypto-obj-y >> to be added, but I have been unable to find a solution that works. >> >> If I add crypto-obj-y, then authz-obj-y must be included to resolve >> dependencies. Daniel suggested splitting authz-obj-y into two, so >> that linux-user plus some of the tools need not link against libpam. >> >> However, I tried that, and in the process managed to break testing. >> I'm not really sure what I did wrong: >> >> TEST check-speed: tests/benchmark-crypto-cipher >> ERROR - too few tests run (expected 32, got 0) >> >> Or maybe it was broken before, but at least this way I'm not touching >> any of the variables that affect tests/Makefile.include. >> >> Given that user-only *is* being linked against the enabled crypto libs, >> the behaviour between static and non-static is identical, which is I >> believe the major portion of Daniel's request. I think further cleanup >> to the makefiles can be done separately. >> > > For avoidance of any doubt > > Acked-by: Daniel P. Berrangé <berrange@redhat.com> > > for Richard to send a pull request with any of the crypto related patches > I'm tagged as maintainer for. The same for linux-user part: Acked-by: Laurent Vivier <laurent@vivier.eu> Thanks, Laurent