[15:16] Welcome to the Linux Plumbers Conference 2020 BOF1/Virtual-Room!

Please remember that the LPC anti-harassment policy applies to interaction in this room, and that this room is being recorded.

Please keep your microphone muted and your video off except when participating in the discussion.

This server is running BigBlueButton. [15:18] Drew Fustini : Hello all :) [15:19] Pavel Pisa : Hello Drew, I can hear you well. [15:38] Pavel Pisa : Please, can you send URLs to the GITs with components you are working{talking about. [15:39] Martin Peres : Pavel: https://linuxplumbersconf.org/event/7/contributions/705/ has most of the links already, but we'll try our best! [15:43] Pavel Pisa : Thanks. [15:56] Pavel Pisa : You can enable download arrow directly on the bottom of the presentation in BigBluButton upload presebtation menu. [15:57] Pavel Pisa : How it is with BoFs, it seems that they have no stream. [16:02] Jamie Lokier : Audio send works, but not video: "Media could not reach the server (error 1020)". [16:02] Martin Peres : I had this issue when connected to my company's VPN. [16:03] Danielle Costantino : Can you enable downloads of the presentation? (click on plus, add slides, click on the download symbol) [16:03] Martin Peres : Although even the Audio wasn't working [16:03] Jason Sidebottom : You were cutting out a bit Jamie... maybe your link, or maybe the server is struggling. [16:03] Jamie Lokier : My link's generally ok with other conferencing tools. I guess it could be having a problem today and yesterday... [16:04] Jamie Lokier : I'll be trying a different link later today, see how that goes. [16:23] Rohit Athavale : +1 for regmap [16:23] Jason Sidebottom : As far as the device-tree vs hotplug discussion goes - these devices aren't *actually* hot plug. They all exist at boot time, so building a DTB when building your soft core isn't really that big a deal [16:23] Jason Sidebottom : Bonus points for being supported in uboot, zephyr, etc [16:24] Christopher Friedt : Hehehe... [16:24] Jason Sidebottom : Oh hey, Jamies webcam is working :D [16:26] Jason Sidebottom : USB is already handled in-kernel though, right? [16:26] Jamie Lokier : Jason, webcam works when using my mobile's wifi :-) [16:27] Jason Sidebottom : Nice [16:29] Geert Uytterhoeven : Using UUIDs? To avoid the need for a central registry? [16:30] Geert Uytterhoeven : Or (truncated) $(sha1sum ) [16:31] trix@redhat.com trix@redhat.com : Could this fit into DFL ? [16:32] karthik poduval : could driver get product id as module parameter (in case its dynamic)? [16:34] Rich Persaud : Is there a requirement for such devices to be accessible to virtual machines, e.g. via virtio? [16:37] Martin Peres : Rich: that would be a good reason to expose the blocks through the normal PCI / USB way [16:38] Rich Persaud : This may be of interest: https://www.intel.com/content/dam/www/public/us/en/documents/reference-guides/pcie-device-security-enhancements.pdf [16:38] Rich Persaud : "PCIe Device Authentication adapts the existing USB Type-C Authentication mechanisms to PCIe---the new elements are the specific PCIe register interface and the associated mechanisms, plus some details that are necessarily specific to PCIe." [16:43] Rich Persaud : There is also an in-progress DMTF standard (SPDM) for attestation/measurement of device and firmware state, https://www.platformsecuritysummit.com/2019/speaker/plank/ [16:44] Rohit Athavale : Don't know a whole lot of LiteX, how applicable is this concept applicable to AXI or AXI-Lite ? [16:47] Drew Fustini : rohit: LiteX can use different buses [16:47] Drew Fustini : Wishbone is common but there are others [16:47] Drew Fustini : Rich: there was discussion on the linux-litex mailing list about have the LiteX modules use a virtio interface [16:48] Jamie Lokier : +1 to classes. Otherwise you get vendors using fake vendor ids for compatibility. [16:49] Jamie Lokier : And then you have to do weird register tests etc to find out which device is really attached and which bugs to work around. [16:49] Martin Peres : Any link I should check out, Jamie? [16:50] Drew Fustini : https://groups.google.com/g/linux-litex/ [16:50] Rich Persaud : The device auth/attestation proposals above involve strong authentication of device and vendor identity [16:50] Jamie Lokier : Martin, none come to mind. [16:51] Martin Peres : ok, thanks :) I'll make sure to read more about the device tree. I have only a light understanding as my background is almost-fully on graphics [16:52] Jamie Lokier : Thanks everyone! [16:53] Jason Sidebottom : Bye! [16:55] Pavel Pisa : Please, if you have time, what is a format of register maps specifications for Litex? Do you generate header files automatically? By which tool? [16:56] Drew Fustini : Pavel: let me get the link [16:56] Pavel Pisa : I see Migen [16:56] Drew Fustini : https://github.com/enjoy-digital/litex/wiki/LiteX-for-Hardware-Engineers [16:57] Drew Fustini : and https://github.com/enjoy-digital/litex/wiki/CSR-Bus [16:57] Pavel Pisa : Thanks [16:58] Drew Fustini : https://lore.kernel.org/patchwork/project/lkml/list/?series=457681 [16:58] Pavel Pisa : We use IPXACT XML format [16:58] Drew Fustini : this is the current patch series for the LiteX SoC and UART drivers if you want to see what it looks like to access the CSR from the software side [16:59] Drew Fustini : ah, I don't think there is a format like that. I believe it all defined in the LiteX python code [16:59] Drew Fustini : https://github.com/enjoy-digital/litex/wiki/CSR-Bus#22-csr-bank-id [17:00] Drew Fustini : i maybe the best explanation [17:00] Pavel Pisa : But you probably want to generate HW which matches headers and thengenerate datasheets documentation. [17:00] Drew Fustini : but also #litex on freenode [17:00] Drew Fustini : the creator Flourent (enjoy-digital) is one there and is very helpful person