Am Thu, 4 Feb 2021 02:33:20 +0000 schrieb Hongbo Wang <hongbo.w...@nxp.com>:
> > > .../dts/inmate-ls1043a-rdb-fman-ucode.dtsi | 1030 > > +++++++++++++++++ > > > configs/arm64/dts/inmate-ls1043a-rdb.dts | 767 +++++++++++- > > > configs/arm64/ls1043a-rdb-linux-demo.c | 57 +- > > > 3 files changed, 1843 insertions(+), 11 deletions(-) create mode > > > 100644 configs/arm64/dts/inmate-ls1043a-rdb-fman-ucode.dtsi > > > > > > > What exactly will that allow the non-root cell to do? Own the DPAA > > completely (taking it from the root cell)? Or does this enable > > sharing (and then only in a cooperative way, due to architectural > > limitations of the DPAAv1)? > > there are some case that user want to non-root cell can connect > ethernet and cloud via DPAA1 port, and test their performance, so we > plan to add DPAAv1 support in jailhouse. > > in this patch set, all DPAA ports are owned by non-root cell, root > cell don't connect ethernet directly via DPAA. Radu-andrei Bulie already has a setup where root and non-root can share DPAA1 in a cooperative way. That depends on driver changes in Linux but in the end is much more powerful than what is proposed here. But it also has nasty implications on isolation between the cells. Handing all of DPAA1 to non-root is an extreme case, next to leaving all in root. While "something in the middle" seems way more useful/flexible and it would be a shame to just go for the extremes. I assume that people will want that sharing eventually, so it should be considered already to see how it fits on the second "extreme" that is proposed here. Henning > > thanks, > hongbo > -- 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/20210206122300.09b3711c%40md1za8fc.ad001.siemens.net.