00:35 <Xgigamen1> https://swedishembedded.com/insights-slip-networking/ i still wanted to forward the message, cause x86 slightly harder way of getting tapping into bios interrupts, is why i did that research partly in cambodia, sweds have seemed to follow and this looks good to me.

00:40 <Xgigamen1> I tried with renode, but hit some bugs or i mean i just was not capable of doing the needed , more documentation is needed, but cortex m3 is available on soft float ABI in qemu too, so it should be possible to get slip going for proprietary phones too.

00:40 <Xgigamen1> they have cortex a9 too

00:41 <Xgigamen1> on qemu

00:43 <Xgigamen1> whole dump reveals that a53 is also available on qemu, i try to get renode working later the same, did not get uart leds to lit up there

00:46 <Xgigamen1> Ideally would want to have uart connection to proprietary hw failing to boot at oob, to hard reset over uart, ipmi power messages would be nice to oob over such solutions.

00:47 <Xgigamen1> i am probably going to openbmc u-boot and linux bootups

00:48 <Xgigamen1> it's one of the last things i request a bit attention and care taken, though it is bit evil towards companies, but for maintaining sake, it is so much simpler this way for users.

00:52 <Xgigamen1> meanwhile the ttl silicon labs two boards as null cable seems to be working, and is reasonably cheap solution, i have two original ones not from china, but local retailers had it, i was just wrapping my head around that leds are controlled in application , so they blink as to how programmer wants

00:52 <Xgigamen1> something is likely hardcoded in driver as default though

00:54 <Xgigamen1> it's usb type-a uart chip two of them connected together, but one has tx-rx swapped, 6pin ttl, seems to work

00:55 <Xgigamen1> zero/null modem, i did my last research on it, would request an oob power management hooks to be sent, which is reasonably doable

00:58 <Xgigamen1> i updated to 5.4 kernel, where the connection seems to be very reliable, but this went to userio, which is irrelevant as of now, character device based stuff , gpiochip is likely deprecated or something

00:59 <Xgigamen1> so on the host solutions like the above and file_xfer projects as well as file sending and all like this works now

01:00 <Xgigamen1> SLIP tcp/ip should indeed be more reliable, i think the author of the blog is well aware and correct on this, but it is slower, but it no longer drops packages once the slip has been fully pulsed, i assume, cause i am in a bit of hurry, i am getting old with no forward progress

01:01 <Xgigamen1> and i need to go to work

01:06 <Xgigamen1> but if i feel like instead of cracking complex pieces of code to go to work, then we are getting somewhere, meaning that things are at all time best, and most things are either done or easily doable

01:09 <Xgigamen1> i think the openipmi code is already available too though, but i did not test, it's in the base of gensio

01:12 <Xgigamen1> but things look like we are finally getting somewhere away from maintenance hell, things look good, and i consider my effort anymore not needed

01:14 <Xgigamen1> I only do some more testing but battery reset and ipmi code over uart, should be all available, just untested yet, but i carry this out, yeah it looks good to me though

01:15 <Xgigamen1> it was some journey, i participated, though others employed played a bigger part at implementations

01:16 <Xgigamen1> I initiated a lot , designed a bit, and proposed a lot, and many people did implementations which seems like success to me

01:29 <Xgigamen1> so if that hack is possible with even simpler cable like usb-otg, actually yes, but this needs an exploit otherwise it gets stuck to the bootrom, usb-otg is hidden in fw, but zephyr has it too, swedishembedded could do a blog entry for this too

01:32 <Xgigamen1> it would as you had an uart on the cpu, i am pretty sure this code is available too

01:33 <Xgigamen1> i have seen also

01:40 <Xgigamen1> They have such a thing as micropython to help, it works single address space bootloader environment only, as said, if you want it to work in protected mode, or os mode, it needs libdimentio to be pulsed and target exploited.

01:40 <Xgigamen1> it needs a port of an exploit

01:41 <Xgigamen1> cause i think otherwise the permissions are causing issues...and some things can not be authored.

01:42 <Xgigamen1> like battery resets would work on bootloader environment only

02:42 <jenatali> :( the new GitLab version doesn't show high-precision MR numbers after 1k

03:22 <robclark> jenatali: apparently using `lowp` instead of `highp`?

03:23 <jenatali> :P

04:34 <mupuf> daniels, anholt_: FYI, we now have builds for x86_64, arm64, ARMv6, and RISCV64: https://gitlab.freedesktop.org/mupuf/boot2container/-/pipelines/873328

04:36 <mupuf> As for "assume local storage", this is also no longer true: b2c.filesystem=myNFS_drive,type=nfs,src=10.0.0.1:/,opts=nodev|ro b2c.cache_device=myNFS_drive

04:38 <mupuf> I fully agree with daniels, one does not simply remove LAVA... but one can use lava+b2c to avoid having to maintain the rootfs that annoys me so damn much :p

04:40 <mupuf> same for the "baremetal" infra :) The first order of biz is to reduce complexity by standardizing on the use of container. When this is done, we'll be able to simplify a lot of things in Mesa CI which should make reproducibility outside of CI waaaay easier

04:59 <mupuf> anholt_: oh, and regarding the "assume local storage", the default is not to use any local storage at all... but that's a little silly to redownload a rootfs all the time

05:01 <mupuf> The valve team's objective for the year wrt the infra is to provide a turn-key solution for fd.o projects. We believe our approach is generic-enough, and if it isn't, the sooner we know the better

07:13 <polygon2> Above this heaven , so far it turned out very good, above this I do not get grant anymore, if course the most maybe fascinating research i have looked at is the neutrino fused nuclear methods, but it elsewhere seemed i paid a little effort and i got my results, if i put myself a grade, its impressive 4- in 5grade system, i am ok with this.

07:15 <polygon2> Most of upsetting stuff others did, and defending against fraud, is not my priority yet, but at the end of the year, super equipment comes through, you did alot of good projects and finest ones are openbmc and DIVINE on llvm.

07:18 <polygon2> to defend against fraud is not my job, but preparing to a possible war, which is low probability yet here, is first priority, fraud prevention have workers and employees, and i am not paid to do that.

07:18 <polygon2> i nearly care not at all, cause i have not performed that

07:18 <polygon2> so i do not get sentences either

07:21 <polygon2> if you can leave me or not bother me at physical locations very nastily attack assaultingly or bloodily, i really in that case am not big teller what fraud you do, cause i know that this department and me are pretty skilled

07:21 <polygon2> it's financial world and their workers are real scientists who are hired for cybersecurity

07:22 <polygon2> if i get framed we will easily riddle out the methods

07:28 <polygon2> The war with russians is low-probability cause of the date of the conflict, as 24 february 2022 independence day of estonias first republic, where russia clearly refers with this move, that ukraine was never a country and column 5, yet they took all our regional sea, so the best solution there since russians were 32 years late to react, is that they get some smaller percentage like 1/4 1/3 of the annexed area.

07:28 <polygon2> cause annexed area really used to belong to them

07:35 <polygon2> i hope to see peace treaty after little bit of more rumbling to state, that it was already part of ukraine for too long 31 years, but russians were cheated on totally, and some area they should get hence smaller because of delays

07:36 <polygon2> it was total below the belt kick to russians imo, look what area they have not at all access towards prime channels to europe

07:43 <polygon2> it's easy to say we do not diserve that, i do not diserve life, cause someone wants that, but to be honest we know that russians got ripped off there dramatically, and putin is not about so soft , and knows his cards

07:43 <polygon2> me i am maybe softy and over too careful and not so warish

07:49 <polygon2> i have lived my whole life in the center border line of center of the city and old town, it's the most prestige place in estonia, it might end up at sale this apartment, norway, sweden , denmark cross the nordic areas around the countries have similar but yet even better real estate and view places

07:51 <polygon2> everything where i have stayed has been bought, it's filled with good energy, cause i as owner or co-stayer or having lived there, am such a friendly and non monstrous non gangster guy

07:54 <polygon2> actually i am not the king despite living all the history at the very best location that estonia had to offer, i am not anything special or decider about things, but the location is healing you at times, the view and stuff the resonance

07:57 <polygon2> the old town according to the labels in during series of wars with the scandinavians was predominantly built by danish

07:58 <polygon2> but germans build some of it too, and designed the language to big degree or shaped and influenced

08:00 <polygon2> so if you want that strategic location, place a bid to my mom, like rich people collecting real estate at strategical locations, in this poorer era, this is the best of estonia

08:01 <polygon2> scandinavian locations and old downs are better though, definitely someone wants and also has the money to operate in estonia so it should go as warm pie without problems to such people

08:04 <polygon2> americans are building their embassy here, stated as most advanced and secured embassy in europe comes to estonia, it is quite near my location is better

08:13 <polygon2> I am yeah packing my things at best position, and tbh could not ask for anything better, i was listened, things turned out so good towards the end, and over my expectations, i performed 4- but result was 5+

08:15 <polygon2> my research was nearly towards five, but my own achievements were not so good, but result was 5+ cause others listened to me, and implemented everything i requested

08:17 <polygon2> so how can a cripple be so happy, several years owing to death cause of injury, cause i performed over my shadow and performed so good , that is why i am so happy

08:19 <polygon2> I had to compensate for the pain and anomalies with wisdom, and boy i made an impact

08:19 <polygon2> so that is why there the hobby was selected, for me to feel better

08:21 <polygon2> if you are an ex athlete just peak in to labs or channels and stay there, you'll see what talents we have in world, and what they really do is purely amazing

08:23 <polygon2> they are at embedded but can not be overflowed with critics, sweds are nicely in charge still, in 2022 spring i talked about it and bit earlier, very good blog entry, makes a lot sense, we are at very good hands

08:23 <polygon2> and mostly i can not be more flattered

08:24 <polygon2> to be listened, even through directly i could not made a real implementation my own, promptly was done by other hackers , i made the word jam

08:26 <polygon2> so their performance in real world computer code implementation is 5 , research was 5 to me, i do not know what grade of creativity and research they had

08:29 <polygon2> those departments are mostly seperate, cause coders are very occupied and need a feed

13:50 <thaller> whot, hi. I have a question about ci-templates+NetworkManager. We have a "pipeline_schedules" that runs a BUILD_CI_IMAGES job (on `main` branch) with setting "AUTHFILE" environment variable. That is related to https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/blob/20cf4ff31a8fcf37c085074d9a9d29a6798dc25e/.gitlab-ci.yml#L421

13:51 <thaller> I don't think it actually works, so I would remove all of this. But is that a problem? Does that mean, our container images pile up? We really don't care about old cached images (and we could even just prune them every X weeks, regardless whether they are used -- they could be re-generated)

13:51 <thaller> we use the cached images to avoid the overhead of installing them over and over. But beyond that, no need to keep them.

17:15 <rodrigovivi> daniels: there are multiple folks with similar complaining: "Tried subscribing to mailing list using https://lists.freedesktop.org/mailman/listinfo/intel-xe. But not receiving any confirmation email."

17:16 <rodrigovivi> but not everyone since we have other folks able to subscribe... any idea of what might be going on?

17:16 <rodrigovivi> any config I could check/test?

17:47 <daniels> rodrigovivi: it’s not config, but likely just Intel mail servers rejecting our email

17:47 <daniels> you can use the admin interface to forcibly subscribe them

17:53 <rodrigovivi> hmmm... yeap, likely... and yes, we are using this interfaces for these cases...