libv changed the topic of #linux-sunxi to: Allwinner/sunxi development - Did you try looking at our wiki? https://linux-sunxi.org - Don't ask to ask. Just ask and wait for an answer! - This channel is logged at https://oftc.irclog.whitequark.org/linux-sunxi
Ultrasauce has joined #linux-sunxi
nyama has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
apritzel has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
gediz0x539 has joined #linux-sunxi
lurchi_ has joined #linux-sunxi
lurchi__ has quit [Ping timeout: 480 seconds]
Suniel has joined #linux-sunxi
Suniel has quit [Remote host closed the connection]
Suniel has joined #linux-sunxi
apritzel has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
gsz has joined #linux-sunxi
faruk has joined #linux-sunxi
_whitelogger has joined #linux-sunxi
_whitelogger has joined #linux-sunxi
_whitelogger has joined #linux-sunxi
_whitelogger has joined #linux-sunxi
Suniel has quit [Remote host closed the connection]
_whitelogger has joined #linux-sunxi
Suniel has joined #linux-sunxi
swiftgeek has quit [Ping timeout: 480 seconds]
swiftgeek has joined #linux-sunxi
lurchi__ has joined #linux-sunxi
lurchi_ has quit [Read error: Connection reset by peer]
Suniel has quit [Remote host closed the connection]
ssvb has joined #linux-sunxi
Danct12 has quit [Remote host closed the connection]
chewitt has joined #linux-sunxi
Danct12 has joined #linux-sunxi
<ssvb> apritzel: the idea was to be able to keep just a single bootloader binary for both secure and non-secure hardware configurations (unless there's a really good reason not to do this) and the text in https://linux-sunxi.org/TOC0#TOC0_images_generation pretty much explains everything
<apritzel> ssvb: oh, long time no see!
<apritzel> ssvb: well, what would be the reason for a single bootloader binary, if it still needs to be put through a script?
<apritzel> ssvb: so my plan is to have a U-Boot Kconfig switch, so it automatically generates the u-boot-sunxi-with-spl.bin with a TOC0 SPL
<apritzel> for devices which are always secure (Remix Mini) this would be in the defconfig, in other cases people can enable it separately
<ssvb> it doesn't have not be put through a script, because this functionality (conversion on the fly) can be integrated into a flasher software
<apritzel> which flasher software ;-)
<ssvb> I know that right now 'dd' is commonly used as a flasher, but this isn't perfect. Because this is error prone and the users may easily hurt themselves.
<ssvb> Having something slightly more intelligent than 'dd' would be a major improvement for user experience.
<apritzel> ssvb: maybe we can have both approaches? mainline SPL learns to deal with being loaded via a TOC0 (checking for TOC0 magic as well and behaving accordingly)
<apritzel> then we teach the U-Boot build system to directly create TOC0, if requested
<apritzel> or the user creates the default eGON and puts it through your script?
Benjojo has joined #linux-sunxi
<apritzel> which could then possibly lose the TOC0->eGON disguise hack, since the SPL can natively deal with this?
<apritzel> ssvb: and yeah, this: how to best install your firmware seems a somewhat trivial, but still unsolved problem
<ssvb> This "if requested" part is bothering me. In the ancient past, if the users wanted to boot via USB FEL, then they had to build a special U-Boot version via using a separate defconfig. Only a few boards supported this, the users had to be able to build their own bootloader rather than downloading some pre-built binary, etc. And of course, by the time when someone would be interested in building a FEL binary, this functionality was
<ssvb> likely to fail to build or run due to bitrot.
<apritzel> ssvb: I guess we need to work out what the use case for secure boot firmware is
<apritzel> ssvb: AFAICT there are only a very few boards that come with the secure fuse burnt
<apritzel> ssvb: which makes this a somewhat niche use case
<apritzel> ssvb: but I could upstream Remix Mini support and keep testing that, also I have a "converted" Pine64
<ssvb> hmm, so your goal is just to support a few "unfortunate" boards with a pre-burnt e-fuse rather than take advantage of secure boot benefits?
<ssvb> I think jemk mentioned that these boards could be be switched back to egon mode by burning a different efuse bit. I haven't verified this myself.
<apritzel> yeah, heard about this, but indeed didn't dare to try this
<gediz0x539> is Mainline U-Boot SPL capable to boot from FEL on A64?
<apritzel> gediz0x539: yes, since v2021.01
<apritzel> ssvb: and I don't know what the goal is: there are some people interested in explicitly enabling secure boot, but the benefits are questionable, as they seem to be holes/bugs in the BROM
<gediz0x539> apritzel: great. i'm still using sunxi64-fel32
<gediz0x539> it's cool though
<apritzel> gediz0x539: also sunxi-fel can now deal with FIT images, so you can just say: sunxi-fel uboot u-boot-sunxi-with-spl.bin
<apritzel> on every SoC, with v8 or v7 cores
<apritzel> ssvb: bauen1 (still over at Freenode ;-) discovered some BROM bugs (in the H6, at least) that allows booting unsigned payloads
<gediz0x539> it sounds like the easiest-to-use form
<apritzel> gediz0x539: btw: the mainline solution is with an AArch64 SPL, we found some clever way of resetting back to AArch32 before returning to FEL
<apritzel> gediz0x539: yes
<gediz0x539> seamless
faruk has quit []
<ssvb> apritzel: That's a good question about the secure boot benefits. My own primary motivation is to avoid the theoretical risk of https://en.wikipedia.org/wiki/Boot_sector#Boot_sector_viruses (which exists because of the SD card boot priority) or other similar attack vectors.
psydroid has joined #linux-sunxi
<ssvb> apritzel: on the other hand, by improving secure boot support we may be digging our own grave if various vendors start making unbreakable locked down devices
<apritzel> ssvb: indeed! I was always wondering why not more vendors use secure boot for their embedded devices (like this Nintendo console), at least as some hurdle
<apritzel> ssvb: but better not give them funny ideas ;-)
<apritzel> but at the end there might be ways around it anyway, as bauen1 had discovered
_whitelogger has joined #linux-sunxi
<libv> ssvb!
macromorgan has joined #linux-sunxi
<macromorgan> question... I'm trying to run kexec by doing the following: |kexec -s -d zImage --dtb=sun5i-r8-chip.dtb --command-line="root=/dev/sda1 console=ttyS0,115200"| and I'm getting the following error: Could not find a free area of memory of 0x285840 bytes...
<macromorgan> anyone seen this before? Here is my memory: https://paste.debian.net/1198740/
<macromorgan> and here is my free output: https://paste.debian.net/1198741/
<macromorgan> doing this on an R8
<ssvb> libv: What's up with the irc channel migration? Seems like both old and new channels are active at the same time and this looks ugly.
<ssvb> And the message at the top of https://freenode.irclog.whitequark.org/linux-sunxi page suggests to contact them to continue logging the new channel.
<anarsoul> time to set +m on freenode?
<libv> anarsoul: i've been thinking about it, but it seems to hardhanded
<anarsoul> not really
<anarsoul> topic clearly states where everyone went
<anarsoul> and most of regulars are already here
<libv> so +1 from anarsoul
<libv> any other takers?
<MoeIcenowy> I don't think we should set +m on freenode
<MoeIcenowy> it's still the most centralized IRC network
<MoeIcenowy> and gets pre-inputed on the server field on many IRC clients
<MoeIcenowy> currently, because some channels migrate to oftc and others migrate to libera, no other one can replace freenode's place now
<libv> fragmentation... the only real irc network left, ruined for shortsighted profit
<jernej> I don't mind being logged in two servers, but having two active channels for same thing on two servers is confusing...
<jernej> without some kind of force, it will stay splitted, so +1 from me
<libv> +2:-1
warpme_ has joined #linux-sunxi
pentabarf has joined #linux-sunxi
z3ntu has joined #linux-sunxi
MatrixTravelerbot[m] has joined #linux-sunxi
jernej_ has joined #linux-sunxi
veremitz has joined #linux-sunxi
jernej has quit [Ping timeout: 480 seconds]
jernej_ is now known as jernej
cmeerw has joined #linux-sunxi
jernej has quit [Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net]
jernej has joined #linux-sunxi
pentabarf has quit []
<rellla> fragmentation is bad, so +1 from me, too. time flies and people will find it, once they found out what irc is...
gsz has quit []
pentabarf has joined #linux-sunxi
<jelly> also, there are 27 users in the eponymous channel on libera; it would probably be nice to take over that one and point users over here at least in the topic
jernej_ has joined #linux-sunxi
jernej has quit [Ping timeout: 480 seconds]
cmeerw has quit [Ping timeout: 480 seconds]
pentabarf has quit []
<smaeul> indeed, since it is broken, I'm interested in secure boot mostly as a curiosity. in fact, I had to exploit H616 SBROM before I could dump it :)