On Wednesday, March 6, 2019 at 4:27:31 AM UTC-5, Nadav Har'El wrote: > > On Wed, Mar 6, 2019 at 6:24 AM Waldek Kozaczuk <[email protected] > <javascript:>> wrote: > >> With couple of last patches applied yesterday we finally made OSv >> properly and fully run on Firecracker. This includes ability to run >> arbitrary images (zfs, rofs or ramfs) with networking and SMP supported as >> well. >> > > Very nice! I tried a couple of things, including interactive applications > which didn't work correctly yesterday, but now seems to work correctly. > Great work! > > A few final points on my wishlist for this subject (none of this is > urgent): > 1. Avoid the conversion to the (very large...) "raw image" in > scripts/firecracker.py. According to an example > https://medium.com/@s8sg/quick-start-with-firecracker-and-firectl- > 2. Find the places in the documentation (?) where we list the supported > hypervisors, and add Firecracker to the list. > in-ubuntu-f58aeedae04b, it seems firecracker is supposed to support qcow2, > not just raw images? > 3. scripts/run.py has a "-p" option to choose hypervisor (see "launchers = > ..." list in run.py), we could add firecracker to this list and call > scripts/firecracker.py. > 4. Now that you fixed the vmlinuz support, does it mean we can also use > Qemu's "-kernel" option? Does it make boot any faster? >
>From what I understand newly enhanced loader.elf can be booted as vmlinux (uncompressed 64-bit ELF aka direct kernel boot) which neither QEMU or nemu supports yet. Here is an issue somebody opened - https://github.com/intel/nemu/issues/205 - where they say it should be supported in version 4 of qemu. In order to support vmlinuz (aka bzImage I think) - 32 bit compressed kernel kind of equivalent to our lzloader.elf we would need to build extra artifact that would resemble it or make lzloader.elf compatible with it. -- You received this message because you are subscribed to the Google Groups "OSv Development" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
