eduardas has quit [Remote host closed the connection]
eduardas has joined #u-boot
monstr has joined #u-boot
Zapy has quit [Ping timeout: 252 seconds]
jwillikers has joined #u-boot
rtjure__ has joined #u-boot
milkylainen has quit [Quit: Connection closed]
Zapy has joined #u-boot
Zapy has quit [Ping timeout: 252 seconds]
Zapy has joined #u-boot
milkylainen has joined #u-boot
milkylainen has quit [Quit: Connection closed]
milkylainen has joined #u-boot
<Tartarus>
mrnuke: OK, so, merging v2021.10-rc4 in to next. next has https://patchwork.ozlabs.org/project/uboot/list/?series=255852&state=* applied, and I see how to do a functional merge, but as I do that I think there's room for further clean-up for the common case now perhaps? Like, HOSTCC gets one case (as there won't be DM_HASH) but everyone else gets DM_HASH? Or is that likely going in the wrong direction? No big rush
rtjure__ has quit [Quit: Leaving]
<mrnuke>
Tartarus: That changes things. I was planning to move hash algos to linker lists, like I did for cryptos and paddings. WIth DM, there's no opportunity for code sharing with host tools
<Tartarus>
mrnuke: OK, sigh. I kind of noticed after the fact there hadn't been much commentary on the DM_HASH case. Perhaps it needs a bit more thinking to split "here is our accelerator" from "here are the algos" ?
eduardas has quit [Quit: Konversation terminated!]
tre has quit [Remote host closed the connection]
<mrnuke>
Tartarus: I agree it needs more thought. I would have expected a U_BOOT_DRIVER() for each hash algo, rather than a U_BOOT_DRIVER() wich encompassess all possibel algos
<mrnuke>
If I'm trying to use SHA256 in SPL, I might not have the room too add SHA1 and MD5, so I'd have issues using HASH_SOFTWARE, as designed
<mrnuke>
Tartarus: I'm also not very keen on having to enum {} our hash algos, since we already identify them by their strings (e.g. "sha256"). and then associated ops structure.
<Tartarus>
mrnuke: Sigh. Would you mind replying on the ML to the patches? Perhaps it's best if I revert them, or at least get someone started on further clean-up
<mrnuke>
Tartarus: sure. I just don't want to scare the guy off.
blmaier has quit [Quit: ZNC 1.7.5+deb4 - https://znc.in]
blmaier has joined #u-boot
fdanis is now known as fdanis_away
<mrnuke>
Tartarus: reply completed. Final thought is, if we do see a DM_HASH approach in the future, it would be helpful for it to include an implementation for a hardware accelerator.
sszy has quit [Ping timeout: 252 seconds]
vagrantc has joined #u-boot
<Tartarus>
mrnuke: Yeah, I forget why they didn't, or if that was supposed to be/now is a follow-up
<mrnuke>
Tartarus: Okay, we need a way to model "I want hash X, ooh, there's a HW driver for X, use that instead of the software implementation"
<Tartarus>
Yeah
<marex>
Tartarus: where are we with the LMB series ? Maybe you should also pull it into next to see what breaks
<mrnuke>
Tartarus: I have a couple of ideas for that
<Tartarus>
marex: Probably another week, I try and let things sit for 2 weeks before I pull them in, so there's enough time for people to finally comment
<mrnuke>
(1) for_each(HASH_DRIVER, drv) if it supports(hash) return hash; if none, then use_sw_implementation()
<marex>
Tartarus: fine
<Tartarus>
I really do want to get riscv fixed so we can stop disabling fdt relocation
<mrnuke>
(2) One HW driver may implement several HASH_DRIVER for each supported algo
gsz has quit [Quit: leaving]
monstr has quit [Remote host closed the connection]
rtjure__ has joined #u-boot
frieder has quit [Quit: Leaving]
<marex>
mrnuke: hey uh ... does foss.st.com domain work for you ?
<mrnuke>
marex: via HTTP? nope. -ENOENT
<marex>
mrnuke: what about via STMP
<marex>
*SMTP
<mrnuke>
marex: no idea how to test SMTP, but : ping: foss.st.com: Name or service not known
<marex>
mrnuke: well, I got 10 emails after replying to my fav ST engineer that foss.st.com is dead, Jim
<mrnuke>
marex: They failed to recall the california governer, so now he's on to get them
<mrnuke>
marex: I suspect it's either tenmporarily down, or some internal re-org of the servers/domains
<marex>
mrnuke: maybe they are running the thing on some non-free solution ... gasp
<mrnuke>
marex: would you be surprised if that were the case?
<mrnuke>
marex: Had a client once complaining that I shouldn't stay connected to the VPN all the time, because they have a limited number of licenses and other people want to use it
<marex>
mrnuke: fortinet ?
<marex>
mrnuke: it is foss.st.com, so yes, it should run solely on free software :)
<mrnuke>
marex: Might have been sonicwall
<mrnuke>
marex: not really. If I am TyCoon of BigCorpo, and I already have the IT infrastructure, why would I set up a complenely new server farm instead of reusing what I have?
<mrnuke>
Why not be greener and more energy efficient instead of showing the world that FOSS has to be a waste of resources ?
<marex>
mrnuke: just run it all in the cloud then
<mrnuke>
marex: you know what the FSF feels about software-as-a-service, right?
<marex>
mrnuke: I dont really care, as long as GPLv2 does its job
<mrnuke>
marex: The FSF thinks you should care. The FSF believes that the difference is small between running propritary SW and using someone else's computer that you don't control
milkylainen_ has quit [Quit: Connection closed]
sbach has quit [Read error: Connection reset by peer]
sbach has joined #u-boot
milkylainen_ has joined #u-boot
jwillikers has quit [Remote host closed the connection]
jwillikers has joined #u-boot
urja has quit [Read error: Connection reset by peer]
agust has quit [Quit: Leaving.]
urja has joined #u-boot
milkylainen_ has quit [Quit: Ping timeout (120 seconds)]