Clamor has quit [Read error: Connection reset by peer]
Clamor has joined #u-boot
matthias_bgg has joined #u-boot
<umbramalison>
trying to port to a custom board. not getting anything on uart. How do serial consoles work? I mean, the DTS describes one or more uarts. but how does uboot/spl know which to write to?
<umbramalison>
there console parameter in the bootargs, but this is describe by the enumeration such as ttyX, how do I know the enumerated name?
<umbramalison>
currently looking into padconf stuff. the SoC is am5726
<marex>
NishanthMenon: ^
<marex>
umbramalison: have a look at CONFIG_DEBUG_UART, that gives you really simplistic UART access outside of DT entirely
<marex>
umbramalison: good for early debugging
<umbramalison>
marex, i looked at this, but i had no idea what to set as the base address.
<umbramalison>
NishanthMenon: Yes, I had completely overlooked the mux_data file, but within it there is an early_padconf that includes references to UART. there's another example which is relevant to the custom board that I have which is the "phytec am57" but they never put that upstream
<umbramalison>
in the phytec example i think it has the uarts i need. so i'm just about to build that now
<sjg1>
apalos: Re that smbios bug, it seems that there is only a 32-bit address in SMBIOSv2, but for v3 it is 64-bit. So I suppose we need to implement 64-bit SMBIO 3.0? Or just hack it into low memory like x86 does?
<umbramalison>
marex: I think i got it. so in dra7.dtsi, there are all the uarts with their base addresses.
Clamor has quit [Ping timeout: 245 seconds]
Clamor has joined #u-boot
matthias_bgg has quit [Ping timeout: 245 seconds]
prabhakarlad has quit [Quit: Client closed]
dsimic has quit [Quit: Going offline for now]
dsimic has joined #u-boot
matthias_bgg has joined #u-boot
matthias_bgg has quit [Quit: Leaving]
ukky has joined #u-boot
<ukky>
Hi! U-Boot supports up to 3 partitions in MBR format, though it should support 4, without the need for extended partiton.
<apalos>
sjg1: I have no idea tbh, I can have a closer look if you want an idea but I simply noticed the crash
<apalos>
if it's only QEMU that blows up i'd be fine hacking it like x86
<ukky>
Is there a reason why extended partition is required to have 4 MBR partitions? There are 4 partition entries in MBR, so extended partition is not needed to have exactly 4 partitions.
<apalos>
ukky: i have no idea, but try pasting the normal version of the url instead of the shortened one
<apalos>
I am pretty sure the majority of the people would avoid clicking random links in an irc channel
<apalos>
(or at least I hope)
<zear>
doesn't that code count from 0?
<zear>
e.g. it's already 4 partitions
<ukky>
no, 'count' is 1-based total number of MBR partitions
<ukky>
'i' is 0-based current partition index. 'i < count' is still valid inside MBR partition loop.
<ukky>
apalos: That is the only link that I have from wgetpaste
<sjg1>
apalos: I expect that any 64-bit machines with >4GB RAM will have this problem. I'll see if I can add SMBIOS3 easily
<ukky>
zear: thanks! I will try to submit an official PR
<apalos>
sjg1: ok. Is it only > 4gb because iirc that qmeu i managed to implode was running on 4gb
<apalos>
sjg1: on the lwip front, Maxim is fighting with the CI, once that thing passes successfully, he will repost and i will review it
vagrantc has joined #u-boot
f_ has joined #u-boot
Clamor has quit [Ping timeout: 255 seconds]
Clamor has joined #u-boot
matthias_bgg has joined #u-boot
<silurian_invader>
apalos: fwiw it's not shortened
<sjg1>
apalos: OK. I hope he can look at bringing in the source rather than submodule this time?
Clamor has quit [Read error: Connection reset by peer]
Clamor has joined #u-boot
ikarso has joined #u-boot
<apalos>
sjg1: He is using subtrees iirc, but I've left that to Tom to decide
<marex>
apalos: there you go, at least some people will click on that random link :)
<marex>
apalos: but it has no https, so, I cannot access the link, too bad
<marex>
ukky: so, what is the problem anyway ? what are you trying to do ?
<ukky>
marex: I need to create 4 MBR partitions (maximum allowed). Current U-Boot only allows maximum 3 partitions without adding extended partition.
<marex>
ukky: create ... how ?
<ukky>
mbr write mmc 0 'partition_strings'
mckoan is now known as mckoan|away
<marex>
ukky: can you create a test case which can be triggered in sandbox ? that would be easier to reproduce that way for others
<ukky>
I can specify full format string here, but don't know how to use sandbox. It is very easy to reproduce.
<marex>
ukky: make sandbox_defconfig ; make ; ./u-boot -Tc 'commands'
<marex>
ukky: the error is in layout_mbr_partitions(), right ?
<marex>
ukky: I think there's some simplification there, like if you have exactly four partitions, it does 3+1{1} because it is easier to layout and extend
<ukky>
marex: will do that and report; Yes error is in layout_mbr_partitions()
<marex>
ukky: thanks
<marex>
ukky: because I guess if you have exactly four partitions and you want to add one, with 3+1{1} layout it is easy to do, but with 4 base partitions , there is no way to do it except for deleting a partition and adding extended partition instead of it first
<marex>
I guess that's the rationale for current behavior
<marex>
I wonder if it would make sense to extend str_to_partitions to let user specify exactly what layout they desire (base partitions and extended/nested partitions included)
<ukky>
marex: for a production system, when you know exactly 4 partitions are required, there should be no limitation is user wants exactly 4 partitions without extended one.
<ukky>
s/is user/if user/
qqq has quit [Remote host closed the connection]
<marex>
ukky: the current heuristics cannot know that , hence the suggestion to improve the ability to exactly specify what you need
<ukky>
marex: after adding 'mbr' command to default config, here is the output of sandbox'ed U-Boot: http://ix.io/4IeR