<digetx> it shouldn't be difficult to extract registers info from existing drivers and implement a minimal support that is actually needed by devices
<DavidHeidelberg> digetx yeah, it's done and driver works, but datasheet would be nice.
<DavidHeidelberg> digetx have you ever tried mail qualcomm about datasheet? I tried now for smb136 & smb347, I wonder what reply do I get
<digetx> I doubt that you will get anything since qualcomm has nothing to do with the summit
<digetx> or have?
<DavidHeidelberg> digetx they bought them :P
<DavidHeidelberg> at least it seems
<DavidHeidelberg> (if you open summit microsystems linkedin page, it'll redirect it to qcom, and later smb are id "qcom,smb...")
<digetx> ok
<digetx> cyndis: will those virtualized clients use fixed syncpoint IDs? if yes, how kernel will know those fixed IDs?
crabbedhaloablut has quit [Remote host closed the connection]
crabbedhaloablut has joined #tegra
<cyndis> digetx: device tree will have the range of syncpoints owned by the vm
gouchi has joined #tegra
<digetx> cyndis: how many syncpoints will have client, will it have spare syncpoints? what about the shared cross-vm syncpoints, won't they need a fixed pre-defined IDs?
<digetx> perhaps I will keep the host1x always-on in upstream driver for now, it's not only the waiter that needs rpm, but also IOCTL and sysfs paths, doesn't make sense to enable rpm until we'll reorganize the code
<digetx> cyndis: started https://docs.google.com/document/d/1yRYJTqqwwsaAv_EdN2MzxkGpRMuTOxxQMj6JGxa4jyE please comment and amend it, will see where it will go; I made doc editable for everyone
gouchi has quit [Remote host closed the connection]