| *** mcastelinoo has joined #kata-dev | 00:29 | |
| *** mcastelino has quit IRC | 00:33 | |
| *** mcastelinoo has quit IRC | 00:40 | |
| kata-dev-irc-bot | <jonolson> @bergwolf I'm definitely interested in startup time here -- if there is any impact, I'd expect it to depend on the size of your initrd whether it's compressed -- I _think_ decompressing and mounting it is effectively serialized during boot, so anything that could be saved there would contribute directly -- I'm not expecting anything magical here, question for me is really whether it's microseconds or milliseconds | 01:01 |
|---|---|---|
| kata-dev-irc-bot | :slightly_smiling_face: | 01:01 |
| kata-dev-irc-bot | <jonolson> @bergwolf in terms of protection, if you mount the NVDIMM ro the host shouldn't try writing to it and mmap() it with PROT_READ -- or, if you want to be fancy, mmap() it MAP_PRIVATE and get CoW semantics. | 01:03 |
| kata-dev-irc-bot | <jonolson> I don't know whether qemu directly allows for either of the above without patching, mind you :slightly_smiling_face: | 01:05 |
| *** zerocoolback has joined #kata-dev | 01:20 | |
| *** zerocoolback has quit IRC | 01:29 | |
| *** zerocoolback has joined #kata-dev | 01:40 | |
| *** zerocoolback has quit IRC | 01:41 | |
| *** zerocoolback has joined #kata-dev | 01:42 | |
| *** zerocoolback has joined #kata-dev | 01:44 | |
| *** zerocoolback has quit IRC | 02:08 | |
| *** eernst has quit IRC | 02:26 | |
| *** zerocoolback has joined #kata-dev | 02:37 | |
| *** zerocoolback has quit IRC | 02:45 | |
| *** zerocoolback has joined #kata-dev | 02:49 | |
| *** zerocoolback has quit IRC | 02:56 | |
| *** zerocoolback has joined #kata-dev | 03:08 | |
| *** zerocoolback has quit IRC | 03:43 | |
| *** zerocoolback has joined #kata-dev | 03:48 | |
| *** zerocoolback has quit IRC | 03:55 | |
| *** zerocoolback has joined #kata-dev | 03:59 | |
| *** zerocoolback has quit IRC | 04:03 | |
| *** zerocoolback has joined #kata-dev | 04:21 | |
| *** sjas_ has joined #kata-dev | 04:25 | |
| *** sjas has quit IRC | 04:29 | |
| *** zerocoolback has quit IRC | 04:32 | |
| *** zerocoolback has joined #kata-dev | 04:33 | |
| *** zerocoolback has quit IRC | 04:37 | |
| *** zerocoolback has joined #kata-dev | 04:55 | |
| *** zerocoolback has quit IRC | 05:01 | |
| *** zerocoolback has joined #kata-dev | 05:05 | |
| *** zerocoolback has quit IRC | 05:16 | |
| *** zerocoolback has joined #kata-dev | 05:53 | |
| *** zerocoolback has quit IRC | 06:20 | |
| *** zerocoolback has joined #kata-dev | 07:05 | |
| *** zerocoolback has quit IRC | 07:37 | |
| *** zerocoolback has joined #kata-dev | 07:51 | |
| *** zerocoolback has quit IRC | 07:56 | |
| kata-dev-irc-bot | <bergwolf> @jonolson Yes, you are right. Decompressing can make a difference. By `initrd`, we are actually using `initramfs` that is just a gzipped cpio archive so no _mount_ action is involved. | 07:58 |
| kata-dev-irc-bot | <bergwolf> @jonolson Yes, we can mount the NVDIMM device read only, or let qemu pass the nvdimm device readonly. But I do not see kata runtime handling it in any place. Therefore I asked @sebastien.boeuf about it. | 08:01 |
| *** zerocoolback has joined #kata-dev | 08:03 | |
| *** zerocoolback has quit IRC | 08:11 | |
| *** fiddletwix has quit IRC | 08:13 | |
| *** sjas_ is now known as sjas | 08:48 | |
| *** yingjun has joined #kata-dev | 09:59 | |
| *** yingjun has quit IRC | 10:38 | |
| *** zerocoolback has joined #kata-dev | 10:38 | |
| *** zerocoolback has quit IRC | 12:08 | |
| *** zerocoolback has joined #kata-dev | 12:11 | |
| *** zerocoolback has quit IRC | 12:21 | |
| *** zerocoolback has joined #kata-dev | 13:07 | |
| *** zerocoolback has quit IRC | 13:12 | |
| *** yingjun has joined #kata-dev | 13:37 | |
| *** yingjun has quit IRC | 13:38 | |
| *** yingjun has joined #kata-dev | 13:39 | |
| *** zerocoolback has joined #kata-dev | 13:40 | |
| *** yingjun has quit IRC | 13:54 | |
| *** zerocoolback has quit IRC | 13:56 | |
| *** zerocoolback has joined #kata-dev | 14:04 | |
| *** zerocoolback has quit IRC | 14:08 | |
| *** jlbutler has quit IRC | 14:26 | |
| *** zerocoolback has joined #kata-dev | 14:38 | |
| *** zerocoolback has quit IRC | 14:43 | |
| *** zerocoolback has joined #kata-dev | 14:45 | |
| *** sjas has quit IRC | 14:56 | |
| *** zerocoolback has quit IRC | 15:29 | |
| *** eernst has joined #kata-dev | 15:47 | |
| *** eernst has quit IRC | 16:43 | |
| *** jlbutler has joined #kata-dev | 16:58 | |
| *** sjas has joined #kata-dev | 19:00 | |
| *** spotz has quit IRC | 19:16 | |
| *** mudpuppy has quit IRC | 19:21 | |
| *** eernst has joined #kata-dev | 19:30 | |
| *** eernst has quit IRC | 19:30 | |
| *** jlbutler has quit IRC | 19:51 | |
| *** jlbutler has joined #kata-dev | 20:19 | |
| *** justJanne has quit IRC | 20:41 | |
| *** justJanne has joined #kata-dev | 20:42 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!