Dmitrii Kuvaiskii
Dmitrii Kuvaiskii
@lejunzhu Thank you! I finally understand the usefulness of KSS. The missing ingridient for me was Step 2 above -- the remote user verifies `CONFIGID` to be a known golden...
5. LibOS/shim/test/regression should be moved to LibOS/regression (similar to Pal/regression)
> 2. Instead of using macros, leverage `call_xxx` on link-time for backend-specific code; I vote for `hostcall_xxx` prefix.
Ah, so `DkDebugDescribeLocation()` also works on "base diffs", not actual addresses? I read the code wrong then. Anyway, this only shows how confusing this `addr` word is. Btw, is this...
I confirm that the newer LTP fails in these tests. Is it time for Gramine to move to a newer LTP, and fix such problems along the way? @mkow @boryspoplawski...
@woju Could you please fix `Multiple versions installed together, for signing` section -- it ends abruptly. (I will remove this comment once you fixed it.)
Looks good to me, but I'm very bad at packaging. Couple points: - We definitely need a guidance for tools (I'm thinking of GSC and the smaller ones like `argv_serializer`)...
> What about finally switching to static linking of PAL with LibOS? :) It would solve all such issues and many more. I would love that, and I want it...
@yamahata raised a very good point. Graphene creates named UNIX domain sockets (to communicate between Graphene processes) under `/tmp`.
This looks like a metaissue, and we already implement most of the recommendations here during Meson build and install. Should we still keep this issue open?