> Subject: Re: [PATCH 06/10] Add libbaremetal > > On 18.08.20 08:22, Peng Fan wrote: > >> Subject: Re: [PATCH 06/10] Add libbaremetal > >> > >> On 17.08.20 11:41, Peng Fan wrote: > >>>> Subject: Re: [PATCH 06/10] Add libbaremetal > >>> > >>> Since we come to discuss in a bigger picture, how do you suggest to > >>> proceed about the bootloader part? > >>> > >> > >> The key is to understand to use cases. What are your primary drivers > >> to push the loader out of Linux? > > > > We have product using aarch64 jailhouse with aarch32 root cell Linux. > > With Linux as the loader, we have to add lots of hacks. > > > > Alought we are trying to upstream aarch64 jailhouse with aarch32 > > inmate cell, we consider to make root cell in aarch32 mode finally. > > > > Interesting. aarch32 with pae? What kind of hacks would be needed? To jump > from aarch32, likely EL2 via the hyp stub, to aarch64?
Yes. We have LPAE enabled. Just part of hacks: a stub code to do EL2 initialization, built out using uboot baremtal, just like the bootloader did. a magic code in the head of kernel head.S and secondary_start to handle switching to aarch32, the magic code will jump to the aarch64 stub code, then switch to aarch32 mode and back to Linux entry. When aarch32 linux jailhouse driver execute entry in jailhouse, we also patch jailhouse with a magic code to handle switch and other stuff. Since that code is not public available, I could not provide more details. Regards, Peng. > > > Is it "just" boot speed, ie. do you need a full-featured > >> Jailhouse later on? > > > > Need a full-featured jailhouse. > > But then how does the early partitioning help? You still need to make every > runtime feature work from aarch32 root cell against some aarch64 Jailhouse. > Loading, specifically. > > > > > Or how close would a static boot-time partitioning come to > >> your use cases? > > > > Not evaluated this. But if it could support communication between > > cells, it should be fine to use it. > > > > Sure, ivshmem will remain a must. We would not support cell destruction and > creation, possibly cell re-loading if there is a good use case. > > Jan > > -- > Siemens AG, Corporate Technology, CT RDA IOT SES-DE Corporate > Competence Center Embedded Linux -- You received this message because you are subscribed to the Google Groups "Jailhouse" group. To unsubscribe from this group and stop receiving emails from it, send an email to jailhouse-dev+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/jailhouse-dev/DB6PR0402MB2760E6327F6376DC675A066B885C0%40DB6PR0402MB2760.eurprd04.prod.outlook.com.
