guillaume_g has quit [Quit: Konversation terminated!]
gsz has joined #u-boot
<apalos>
sjg1: Doesn't the same assumption hold for OF_HOSTFILE?
<apalos>
We might as well define a board_fdt_blob_setup() for sandbox which just calls sandbox_read_fdt_from_file() instead of having yet another Kconfig option
mckoan|away is now known as mckoan
guillaume_g has joined #u-boot
rtjure has quit [Ping timeout: 240 seconds]
<apalos>
sjg1: and I missed something yesterday on risc-v OF_PRIOR_STAGE can also fixup the u-boot dtb apparently
<apalos>
as long as CONFIG_OF_BOARD_FIXUP is defined in the Kconfig
<apalos>
so it's the Kconfig is just defining 2 different ways of providing the dtb. If I undestarnd start.S correctly, the dtb can either be passed on r1 or a 'prior_stage_fdt_address' symbol
<apalos>
the later overwrites the a1 value, if selected
<wooosaiii>
Does anyone know if FIT image hash (crc32) fails in SPL, does it automatically load the next configuration ?
<apalos>
wrong, it stores in in prior_stage_fdt_address *as well* if OF_PRIOR_STAGE is selected, but why. It's already stored in gd ...
sszy has joined #u-boot
tnovotny has joined #u-boot
monstr has joined #u-boot
GNUtoo has quit [Remote host closed the connection]
GNUtoo has joined #u-boot
PatDelaunay has joined #u-boot
mmu_man has joined #u-boot
gsz has quit [Ping timeout: 252 seconds]
mckoan is now known as mckoan|away
mthall has quit [Ping timeout: 252 seconds]
joaoSchim_ has joined #u-boot
JoaoSchim has quit [Ping timeout: 246 seconds]
joaoSchim__ has joined #u-boot
joaoSchim_ has quit [Ping timeout: 252 seconds]
joaoSchim_ has joined #u-boot
joaoSchim__ has quit [Ping timeout: 265 seconds]
joaoSchim__ has joined #u-boot
joaoSchim_ has quit [Ping timeout: 252 seconds]
<sjg1>
apalos: For OF_HOSTFILE that sounds right for now
<sjg1>
apalos: But I really would like things to be deterministic, so I actually don't like board_fdt_blob_list() very much. Still, it is there for boards that need it, and I think it would clean up the code
<apalos>
yea I already got a patch for riscv
<apalos>
xypron tested it against the unmacthed board and I did some basic testingon qemu
<apalos>
things seems to 'work'
gsz has joined #u-boot
<apalos>
sjg1: and you can't have 'deterministic' things obviously. It's a bunch of different boards
<apalos>
What you can do is limit the options. Let me send that patch in a bit
<sjg1>
apalos: Let me reply to your patch, I think there is a misunderstanding here
Lokis has joined #u-boot
Gravis has joined #u-boot
PatDelaunay has quit [Quit: Client closed]
<sjg1>
apalos: As I see it, our job is to try to do things in a standard way across all boards, so far as possible. Clearly OF_BOARD and OF_PRIOR_STAGE don't help with that, but we can do it, I think
<apalos>
I don't see how
<sjg1>
apalos: sent reply
<apalos>
You get a magical dtb passed in a magical register per *architecture*
<apalos>
even worse per prior stage boot loader
<apalos>
so I think the way OF_BOARD looks now after removing OF_PRIOR_STAGE is fine
<apalos>
ah right I read the mail now
<apalos>
Sure, having a varaible in .data which every board can go and copy the address is better than do housekeeping with gd_t
<apalos>
ideally we could have that variable in .ro_after_init, similar to what the kernel has
<apalos>
which means RW till we fix it up and then switch it to .ro
<apalos>
but we can probably apply an extra level of indirection on that and be 'future proof'
<apalos>
famous last words. We could use a variable in .data and *keep* the callback as well
<apalos>
So if a board is really weird, instead of shooting hardware designers (which is always better), we could fallback to that callback eventually
<apalos>
but the 'norm' would be just an exported fdt address
<sjg1>
apalos: I hope we can have a bloblist pointer in a standard register for each arch, and an option to look in that for a devicetree
GNUtoo has quit [Remote host closed the connection]
GNUtoo has joined #u-boot
matthias_bgg2 has joined #u-boot
LinuxHackerman has joined #u-boot
tlwoerner has quit [Remote host closed the connection]
tlwoerner has joined #u-boot
mmu_man has quit [Ping timeout: 265 seconds]
torez has joined #u-boot
cengiz_io has quit [Ping timeout: 240 seconds]
paulbarker has quit [Ping timeout: 268 seconds]
NishanthMenon_ has quit [Ping timeout: 268 seconds]
cengiz_io has joined #u-boot
paulbarker has joined #u-boot
NishanthMenon_ has joined #u-boot
tnovotny has quit [Quit: Leaving]
monstr has quit [Remote host closed the connection]
<Tartarus>
Guess it needs a re-spin due to the hash stuff, or is that immature enough on review now that I should just revert it? mrnuke wasn't happy with it either
wooosaiii has quit [Remote host closed the connection]
wooosaiii has joined #u-boot
<sjg1>
Tartarus: yes it needs another look first. It is number 2 on my list so should be soon. We need to decide if we want to make CONFIG_IS_ENABLED() support HOST Kconfigs or not. That is the basis of the series
<sjg1>
Tartarus: mrnuke: I think mrnuke is still not convinced
vagrantc has joined #u-boot
fdanis is now known as fdanis_away
<mrnuke>
sjg1: True. I think we should push as much CONFIG_ stuff out of common code as possible
<LinuxHackerman>
Hi folks. I'm trying to get my nanopi m4 booting via SATA. I have a SATA "hat" for it, which has a PCIe SATA controller. This is working great in Linux, but I'm having some trouble getting u-boot to recognise the PCIe PHY (?)
<LinuxHackerman>
I've tried defining DEBUG in a number of relevant-looking files, and so far what I've got is:
<LinuxHackerman>
It's a bit unclear to me why this would be failing, because the device tree does contain the phy, with a status of okay. Can anyone point me in a potentially useful direction to get any further with this? :)
Guest2026 has quit [Ping timeout: 268 seconds]
sszy has quit [Remote host closed the connection]
mmu_man has quit [Ping timeout: 265 seconds]
torez has quit [Remote host closed the connection]
birkoff has joined #u-boot
birkoff is now known as Guest2222
mmu_man has joined #u-boot
gsz has quit [Quit: leaving]
Gravis has quit [Ping timeout: 265 seconds]
guillaume_g has quit [Quit: Konversation terminated!]
sbach has quit [Read error: Connection reset by peer]
sbach has joined #u-boot
<sjg1>
mrnuke: But no more than possible?
<cambrian_invader>
LinuxHackerman: have you defined DEBUG in drivers/phy/phy-uclass.c?
<cambrian_invader>
oh, you have
<cambrian_invader>
perhaps enable it for drivers/core/uclass.c as well?
<cambrian_invader>
you can also enable logging and set the level to 8
<cambrian_invader>
and the default level to 4 or something sane
<cambrian_invader>
and enable the log command as well
<cambrian_invader>
then at runtime you can bump up the level before running your test
<LinuxHackerman>
it's already defined for uclass.c as well
<LinuxHackerman>
but yeah good shout on the log command
mmu_man has quit [Ping timeout: 265 seconds]
<LinuxHackerman>
OK, that's definitely helpful, thanks very much! Now to work out what this means (if anyone has any immediate pointers that can save me some investigation time I'd appreciate them, but otherwise I'll just be looking through the code and hoping I find something)