lanefu changed the topic of #armbian-rockchip to: Armbian - Linux for ARM development boards | Rockchip SoC | www.armbian.com | This channel is relayed to the equivalent Discord channel | this channel is logged
<Armbian-Discord>
<ManoftheSea> Rk3399 sata driver? Soc specific?
<Armbian-Discord>
<Tonymac32> wut
<Armbian-Discord>
<ManoftheSea> I was, very poorly, asking if the SATA driver for RK3399 is a SOC-specific driver. Then I forgot about it.
<Armbian-Discord>
<ManoftheSea> CONFIG_PHY_ROCKCHIP_PCIE=m hmm, I wonder if that's loaded
<Armbian-Discord>
<ManoftheSea> okay, it is, and I can see the SATA controller.
<Armbian-Discord>
<ManoftheSea> JMB58x AHCI SATA controller
<Armbian-Discord>
<ManoftheSea> which is just the ahci driver. lspci -vv says it's in use. Yet, I can't see disks.
<Armbian-Discord>
<ManoftheSea> I... wonder if they aren't spinning up, because there's something I'm not toggling.
<Armbian-Discord>
<Tenkawa> @ManoftheSea I assume you had SATA_HOST etc enabled too right?
<Armbian-Discord>
<ManoftheSea> I think so
<Armbian-Discord>
<Tenkawa> there's also AHCI drivers too that need to be inserted/compiled
<Armbian-Discord>
<ManoftheSea> I'll be getting back into it today
<Armbian-Discord>
<ManoftheSea> Ahci is "y"
<Armbian-Discord>
<Tenkawa> there's multiple.. not just one
<Armbian-Discord>
<ManoftheSea> Ok. I'll look there
<Armbian-Discord>
<Tenkawa> Yeah AHCI is a pain
<Armbian-Discord>
<Tenkawa> too much overlapping/interweaving
<Armbian-Discord>
<ManoftheSea> hmm... apparently I've been trying to use a mainline 5.15.87 kernel on the helios64. I wonder if that's part of my problem.
<Armbian-Discord>
<Tenkawa> bummer
<Armbian-Discord>
<Tenkawa> especially if the helios64 has one of these vendor heavily modified non-mainline ones
<Armbian-Discord>
<Tenkawa> (ie most rockchip ones)
<Armbian-Discord>
<ManoftheSea> no, but there are patches on current for armbian, so I presume they're not upstream in 5.15
<Armbian-Discord>
<ManoftheSea> I'mma try 6.1.2
<Armbian-Discord>
<Tenkawa> Well even the Khadas 5.10 tree vs the Radxa one was odd...
<Armbian-Discord>
<Tenkawa> I started at the same .66 point and could get much better results on the Khadas hardware using Radxa's tree
<Armbian-Discord>
<Tenkawa> (odd as that sounds)
<Armbian-Discord>
<Tenkawa> At some point though the MCU and the RNG starts diverging though and thats problematic
<Armbian-Discord>
<ManoftheSea> itif you like it than you shoulda put a RNG on
<Armbian-Discord>
<Tenkawa> It has one
<Armbian-Discord>
<Tenkawa> the kernel access to it was the problem
<Armbian-Discord>
<Tenkawa> I need to find the entry point and add it to the other source driver
<Armbian-Discord>
<Tenkawa> The MCU though is a bear of a undertaking
<Armbian-Discord>
<rpardini> it's insane how much time/$ Khadas put into that MCU code. Just terrible they did on a 2.6.x vendor kernel
<Armbian-Discord>
<rpardini> those vendors never put any money into software, and when they do, they throw it away. a sad world
<Armbian-Discord>
<rpardini> (they're probably laughing at me right now, sitting on a pile of Android-originated money.)
<lanefu>
yeah my buddy is running a helios64 on a random image i built iwth kernel 6.0.6 and its held up