vstehle has quit [Read error: Connection reset by peer]
vstehle has joined #linux-rockchip
matthias_bgg has joined #linux-rockchip
midnight has quit [Ping timeout: 244 seconds]
midnight has joined #linux-rockchip
Net147_ has quit [Read error: Connection reset by peer]
Net147 has joined #linux-rockchip
Net147 has quit [Read error: Connection reset by peer]
Net147 has joined #linux-rockchip
matthias_bgg has quit [Ping timeout: 264 seconds]
chewitt has joined #linux-rockchip
alpernebbi has joined #linux-rockchip
s_frit has joined #linux-rockchip
hexdump0815 has joined #linux-rockchip
swiftgeek has joined #linux-rockchip
matthias_bgg has joined #linux-rockchip
kaspter has quit [Remote host closed the connection]
kaspter has joined #linux-rockchip
Net147 has quit [Read error: Connection reset by peer]
Net147 has joined #linux-rockchip
Net147 has quit [Read error: Connection reset by peer]
Net147 has joined #linux-rockchip
damex has quit [Ping timeout: 240 seconds]
lkcl has quit [Ping timeout: 264 seconds]
damex has joined #linux-rockchip
lkcl has joined #linux-rockchip
damex has quit [Read error: Connection reset by peer]
damex has joined #linux-rockchip
tuxd3v has quit [Ping timeout: 265 seconds]
tuxd3v has joined #linux-rockchip
WoC has quit [Remote host closed the connection]
WoC has joined #linux-rockchip
tuxd3v has quit [Read error: Connection reset by peer]
tuxd3v has joined #linux-rockchip
<uis>
What's the point of CONFIG_TPL_ROCKCHIP_BACK_TO_BROM?
WoC has quit [Ping timeout: 264 seconds]
alyssa has joined #linux-rockchip
vstehle has quit [Quit: WeeChat 3.0]
<alyssa>
daniels: mmind00: quesiton #1 with rk3399 afbc ytr vop stuff is whether the hw actually supports both orders
<alyssa>
which isn't totally obvious to me
<alyssa>
certainly appears so from the kernel impl
<alyssa>
Amlogic is the more interesting case, only advertising DRM_FORMAT_*BGR
<alyssa>
but iirc that's RGB* in non-kernel speak?
<daniels>
depends what you mean by non-kernel speak, since Wayland at least speaks kernel-endian for formats
<daniels>
XRGB is (x << 24) | (r << 16) | (g << 8) | b
<tlwoerner>
uis: normally an SoC's bootROM loads the next stage bootloader, then that stage loads the next stage, and so on and so on
<tlwoerner>
uis: but with some rockchip devices a TPL stage is loaded by the bootROM, but instead of the TPL loading the SPL, the code has to jump back to the bootROM
<tlwoerner>
uis: the friendly people in #u-boot might be able to help
<apritzel>
tlwoerner: wasn't it that the TPL is (ab)using the (MMC) loading routines from the BROM?
<uis>
And what purpose of TPL? Why not to use SPL only?
<tlwoerner>
apritzel: probably :-) some sort of interaction between the TPL and bootROM
<uis>
Or TPL only. Sounds strange...
<apritzel>
uis: I think because SRAM is so small that not both DRAM init *and* all the loading code fits into there
<tlwoerner>
u-boot grows over time and doesn't fit so we need SPL
<tlwoerner>
SPL grows over time so we need TPL
<apritzel>
tlwoerner: something like this, yes
<apritzel>
the whole FIT loading, verification and Falcon support (direct kernel boot) needs more code these days
<tlwoerner>
not to mention finishing the conversion from #include files to Kconfig
<apritzel>
tlwoerner: well, if done correctly this doesn't need to create bigger code, modern toolchains are quite clever in dropping unneeded bits
<tlwoerner>
apritzel: ahhh, sorry i misunderstood what you were saying by "more code" i thought you meant "that work could use more coding" not "that work expands the size of the resulting object file"
<tlwoerner>
apritzel: in that case… not to mention driver model and devicetree support ;-)
<apritzel>
yeah, sorry, sloppy wording, increase the .text and .rodata segment ;-)
alpernebbi has quit [Quit: alpernebbi]
hexdump0815 has quit [Quit: Connection closed]
swiftgeek has quit [Ping timeout: 264 seconds]
<uis>
But sun4i can boot without tpl
swiftgeek has joined #linux-rockchip
eballetbo has quit [*.net *.split]
eballetbo has joined #linux-rockchip
<apritzel>
uis: because the A10 is ARMv7, so can use Thumb2, has 48KB of SRAM, and just needs to load a legacy U-Boot image (no FIT, no signatures)
<uis>
FIT is embedding kernel to u-boot?
<apritzel>
uis: no, it's just a more sophisticated image format, with multiple images, and stuff like hashes, signatures, and a selection of images to choose from: but that comes at a price, namely code and data size
<uis>
But it should be not in SPL
<apritzel>
FIT was *originally* meant to combine kernel, initrd and DTB, but then later got applied to the SPL loading U-Boot proper, a DTB, and other firmware blobs
<apritzel>
(like Trusted Firmware, management controller firmware, you name it)
alyssa has left #linux-rockchip [#linux-rockchip]
mraynal has quit [Remote host closed the connection]