<sjg1>
Tartarus: Right, not nefarious...when EFI wants to use memory outside of the U-Boot area it should indeed take account of lmb. But it should not use such memory until we know were are booting with lmb
<sjg1>
It is quite unnecessary, as my simple series shows
<Tartarus>
sjg1: OK? But that implies that EFI is using some unreasonable amount of memory today rather than like how much memory USB uses when I'm going to boot from eMMC anyhow
monstr has quit [Remote host closed the connection]
mmu_man has joined #u-boot
warpme has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
goliath has joined #u-boot
mmu_man has quit [Ping timeout: 265 seconds]
mmu_man has joined #u-boot
frieder has quit [Remote host closed the connection]
<sjg1>
Tartarus: It isn't the amount. It actually only uses about 250 bytes initally. It gets expanded to 12KB because of its allocation mechanism. It is just that I would like that in the U-Boot area, to keep things clean and simple
<Tartarus>
sjg1: It's feeling like we're doing a lot of work that's not "clean and simple" here
<sjg1>
Tartarus: In my series?
<Tartarus>
Yes, your series is not "clean and simple"
jfsimon1981 has quit [Remote host closed the connection]
jfsimon1981 has joined #u-boot
Stat_headcrabbed has quit [Remote host closed the connection]
Stat_headcrabbed has joined #u-boot
prabhakalad has quit [Ping timeout: 252 seconds]
prabhakalad has joined #u-boot
Stat_headcrabbe1 has joined #u-boot
Stat_headcrabbed has quit [Ping timeout: 264 seconds]
ikarso has quit [Quit: Connection closed for inactivity]
___nick___ has joined #u-boot
___nick___ has quit [Client Quit]
___nick___ has joined #u-boot
<Tartarus>
Oh nice, I should bcc someone there I guess moving forward. lwm stopped listing us but I keep doing it anyhow
<sjg1>
Tartarus: What is not clean/simple about it? I'm pretty sure this is how it should have been implemented from the start
<Tartarus>
sng: Oh before I forget again, I think a bunch of the SPL stuff can go away as it because was a side-effect of the event thing, that's gone now?
<Tartarus>
sjg1: Because your line in the sand point of cooperation means we still have to handle the other case, so now there's two cases not one.
<sjg1>
Tartarus: The really nice thing about the patch is that there is only one case. It doesn't even change efi_allocate_pages()! It really couldn't be any simpler IMO
<sjg1>
Tartarus: We start out with a small region for EFI to use, then feed it the rest of memory when firing up the full EFI subsystem
<xypron>
Why would we want this?
<sng>
Tartarus there was a comment from Ilias [1] where he is suggesting that we can do away with the config symbol altogether, by incorporating a check for EFI_LOADER in the lmb_notify() function.
<sng>
Tartarus so the config symbol goes away altogether. the patch can be dropped
<Tartarus>
sng: Lets get v4 going asap then please
<sjg1>
xypron: It avoids the complexity of syncing lmb and EFI. Only one of them is in charge at a time. Also, if EFI is not eventually used for booting, the lmb/efi sync is a waste of time
<sng>
and tweak the lmb_notify() function to have a check for CONFIG_IS_ENABLED(EFI_LOADER)
<Tartarus>
and just depend on or include xypron's patch to document that first flag, and then document your additional ones
<Tartarus>
*one
<sjg1>
Tartarus: I'm going to pull in some x86 patches as Bin seems to have been away for quite a while
<Tartarus>
sjg1: OK
<sng>
Tartarus so can we first get this series in, and then I work on adding documentation on the flags ? are you looking for that documentation as part of this series ?
<Tartarus>
sng: Well, the single flag you add is a 1 line comment to document, so do that today
<Tartarus>
"Tomorrow" check what's not documented well enough in lmb.h and add paragraphs to the rst please
<sng>
Tartarus ah okay. i thought you were looking for a full fledged document about the significance of the flags
<Tartarus>
sng: I am, but that can wait
<sng>
Tartarus okay. let me work on getting a v4 tomorrow
<Tartarus>
Thanks
<sjg1>
Tartarus: One question on Dockerfile...I see it manually downloads all the toolchains - should we use buildman to do that, perhaps? It might help us to keep things in sync
jfsimon1981 has quit [Remote host closed the connection]
<Tartarus>
sjg1: No, then we would have to put buildman in Docker to use it
davlefou has quit [Ping timeout: 265 seconds]
Stat_headcrabbed has quit [Remote host closed the connection]
davlefou has joined #u-boot
<sjg1>
Tartarus: Hmm, isn't that what happens at the end of Dockerfile?
<Tartarus>
No
<Tartarus>
That's just making the config file
vagrantc has joined #u-boot
<sjg1>
Tartarus: Oh I see, it's just adding the requirements into the container. Oh well.
naoki has joined #u-boot
___nick___ has quit [Ping timeout: 252 seconds]
fatal has quit [Remote host closed the connection]
fatal has joined #u-boot
umbramalison has quit [Quit: %So long and thanks for all the fish%]