**** BEGIN LOGGING AT Fri May 21 02:59:56 2021 May 21 03:07:27 Segway to false hopes? May 21 03:07:30 ha. May 21 03:07:46 But...that ProtoCape was made by GHI at one point. May 21 03:15:29 It is similar in looks to the rest of their Capes they produced... May 21 03:15:57 I think there is a LED on it too! May 21 03:16:34 Speakin' of that LED on the ProtoCape from GHI, did anyone ever get the LED lit up? May 21 03:17:00 I remember breakin' my brain on it. Oops, never got it lit...argh! May 21 13:52:58 HI GoodMorning: how to check cellular modem ppp link quality or cell signal strength. Any iwconfig for wifi like app? May 21 13:55:44 modem manager (mmcli ) can show signal strength, not sure installing modem manager will have any conflict with ppp? May 21 13:58:22 if you're using usb then it probably doesn't, if you're using an uart then it definitely will May 21 14:00:03 well, hmm May 21 14:00:33 apparently ppp is mostly handled in the kernel, so maybe it can insert commands into the data stream using an escape sequence May 21 14:00:50 I clearly don't know enough about how ppp works on linux (I've never used it) May 21 14:02:02 we are using uart. I think I may sent escape to modem, then use AT+CSQ to get singal strength May 21 14:02:33 yeah except that's not something *you* can do, either pppd or the kernel has to do that May 21 14:03:27 since normal data transmission and reception needs to be halted before the escape is sent and resumed once command mode is exited May 21 14:04:32 Understood May 21 14:04:53 but like I said, I don't know enough about ppp on linux to have any idea whether this is a thing May 21 14:04:59 easiest would be to just Try It And See May 21 14:05:43 if ppp is indeed handled in the kernel and it has a mechanism for interjecting commands while in data mode, then presumably that's what modemmanager will use May 21 14:05:44 Well, it has signal strength logged at the connection time. May 21 14:17:16 tigerxy: what modem are you using? May 21 14:21:23 Quectel BG96 cellular modem May 21 14:22:05 (connected via uart) May 21 14:23:22 yes, we are uing uart connection May 21 14:25:17 seems like it has multiple uarts May 21 14:25:21 and also usb May 21 14:26:56 with uarts ports: one for cell, one for GPS, and one uart debug port. With USB, I think it has four virtual usb-to-uart ports May 21 14:28:48 some modems have two virtual uarts on usb allowing one to be used for data and the other for AT commands at the same time May 21 14:29:14 it's not connected via usb May 21 14:29:26 maybe they should reconsider that May 21 14:29:46 there are also modems that can multiplex data and commands on a single uart May 21 14:29:51 yeah that's what I said, but the hardware designers only considered what made the hardware design simpler, with no regard for the software side ;P May 21 14:29:52 maybe this one has such a mode May 21 14:30:41 unless this thing has already been manufactured and sold, a redesign seems like a good idea May 21 14:31:35 interrupting a ppp data stream to query the signal strength isn't a good idea May 21 14:31:46 mru, that is the uart port for cell. But if pppd is connected, then need to escape sequence +++, then send at commands as zmatt pointed out May 21 14:32:08 yes, and that would disturb the data traffic too much May 21 14:32:25 but there may be a different mode that allows multiplexing without long delays May 21 14:32:40 you'll have to read the manuals May 21 14:32:44 that mode would also need to have kernel driver support May 21 14:32:52 if it exists May 21 14:33:00 (I've see no mention of it) May 21 14:33:01 the kernel is software May 21 14:34:41 If using modemmanger, then mmcli manages escape and interrupts data stream and returns signal strength May 21 14:35:10 does it? you've tested this? then what question do you still have? May 21 14:35:13 yes, but that will pause the data stream for a second or so May 21 14:35:21 if it even does that at all May 21 14:35:52 I don't see a plausible way that could work unless there's kernel support for this May 21 14:36:01 but this is why I said: try it and see. if it works, it works May 21 14:36:04 I didn't test that, it is based on manual. But I am using pppd, and connman, not NetworkManager May 21 14:36:24 why do you care about NetworkManager? (it has nothing to do with modemmanager/mmcli) May 21 14:37:08 I thought modemmanager need networkmanager supports ( I could be wrong) May 21 14:37:33 you are May 21 14:38:13 networkmanager might use modemmanager, definitely not the other way around May 21 14:39:09 too many things in linux. I am a newbie. May 21 14:40:13 nah that's not an excuse... I didn't know what modemmanager is or does until you mentioned its name and I googled it May 21 14:42:01 Well, you are quick and have better knowledge of linux. Linux is very much new to me. May 21 14:43:04 Google is great. Easy to find many answers at your finger tips May 21 14:43:27 this doesn't really have anything to do with "linux" though... like, this is just some piece of software, one I've also never used (or, like I said, even heard of before today) May 21 14:46:04 I am thinking occasionally to send escape and AT+CSQ to query signal strength using a script, if no application is using cell. May 21 14:46:26 getting that to play nicely with pppd won't be easy May 21 14:46:57 that's the single worst idea I've heard so far May 21 14:47:04 if anything is going to work, it's using modemmanager May 21 14:47:24 ok, then I give up May 21 14:47:31 if there's a way to safely interject commands while in data mode, it probably knows how May 21 14:48:05 why? I'm puzzled why you are asking these questions instead of trying to see if modemmanager works May 21 14:48:06 if the modem uses plain at/ppp with no extras, switching mode necessarily adds an annoyingly long delay May 21 14:49:03 like I said several times already... if you want to know if modemmanager can query the signal strength while pppd is connected, TRY IT AND SEE May 21 14:49:18 if it has dual "uarts" over usb, redesign the board to use that May 21 14:51:09 it does May 21 14:51:13 because the script for pppd is ready to use. So I started with pppd. Now everything is done with pppd.  So if there is a solution for checking signal strength, then it is easier. So I asked first. For modemManager, it is like I have to start all over again May 21 14:52:23 mru: with usb you'd probably use qmi for comands and cdc-ecm for data anyway May 21 14:52:54 rather than at commands or ppp May 21 14:53:49 tigerxy: I don't see why anything would have to change May 21 14:53:54 actually, I had modemmanager working, when I started with dev board, which I used usb connection. But hardware is designed not to use USB, so I picked pppd dor uart May 21 14:54:43 why did you not test with uart connection if the final hardware design was going to use uart? prototyping something that's different from the final hardware is not smart May 21 14:55:52 anyway, I'm afk May 21 14:57:02 tested. pppd with uart is working. Getting signal strength is not something must-have. I am evaluating antenna, so i am thinking to get signal strength easier way May 21 14:58:30 if you use modemmanager instead of a connection script (so that pppd is only used in data mode) you can probably even just disconnect pppd briefly (without actually disconnecting the modem's data connection) to query signal strength May 21 14:59:23 ok. guys, thank you all May 21 14:59:47 be sure to thank the hardware guys for making life miserable on the software side :P May 21 15:00:28 hardware designer is outside consultant May 21 15:00:41 lol May 21 15:01:03 well, now I'm really afk May 21 15:01:56 ok. @zmatt. Nice chatting with you May 21 15:23:55 hi there May 21 17:18:10 hi all -- do you guys know if the bbb can boot from the eMMC boot partitions? (mmcblk1boot0 and mmcblk1boot1) May 21 17:18:20 it cannot May 21 17:19:44 (you mean in the sense of bootrom loading u-boot I assume? since that's pretty much all that would fit in those) May 21 17:22:24 zmatt: yes, I mean flashing the bootloader (u-boot or barebox) in mmcblk1boot{0,1} rather than a user partition of mmcblk1 May 21 17:24:35 bootrom looks for a configuration header in sector 0, 256, 512, and 768 of eMMC and if found expects the MLO (raw executable with 8-byte header containing load address and length) in the immediately following sector(s) May 21 17:25:05 so that can still be exploited to have dual bootloaders for atomic bootloader updates May 21 17:29:38 zmatt: yes I currently the MLO written at offset 128 and have 2 (user) MBR boot partitions can I can switch between, I was wondering if I could use the eMMC boot partitions for that purpose, and thus benefit from the ext_csd registers to switch between the two instead of tweaking the MBR active partition flag May 21 17:30:55 you could have a tiny shim loader in between bootrom and u-boot SPL that knows how to do that May 21 17:35:05 I think either the SoC supports toggling this boot register, or it does not May 21 17:35:31 the SoC has no relevance in that May 21 17:35:36 and I think the bbb SoC unfortunately does not support this (hence the mmcblk1bootX partitions are useless) May 21 17:37:13 hi right, that's an eMMC register, not a SoC register. Still I think the eMMC on bbb does not have such support. Does it? May 21 17:37:40 it does, otherwise there would be no boot partitions May 21 17:38:47 zmatt: then why did you say that the bbb cannot boot from the eMMC boot partitions? May 21 17:40:22 like I said, bootrom makes no use of them, but that doesn't mean you can't "toggle this boot register" (which requires no involvement of the SoC, you can do it with a linux commandline utility), and if you wanted to you could absolutely have an intermediate loader (placed in the main eMMC partition or some other place from which bootrom can load it, like SPI flash) that loads the next stage (u-boot ... May 21 17:40:28 ...SPL) from the active eMMC boot partition May 21 17:40:51 it seems doubtful it would be worth the effort, but it's certainly possible May 21 17:42:17 I see what you mean now May 21 17:44:43 if the barebox MLO can boot from these boot partitions, it'd be worth it. If it cannot, it's indeed not worth it to add yet another stage to boot from there May 21 17:45:50 I'm guessing barebox is, just like u-boot, too obese to fit into an MLO itself and therefore uses 2 loading stages (like u-boot and its SPL) ? May 21 17:48:01 in which case the SPL (or whatever barebox calls it) would need to know how to load the main executable from some offset (e.g. 128KB) from the start of the active boot partition... May 21 17:49:25 (that would be necessary regardless, i.e. also if bootrom *did* support loading MLO from the active eMMC boot partition) May 21 17:50:12 zmatt: that's what it does, yes. But wouldn't it be "simple and small enough" to extend the SPL/MLO to boot from the eMMC boot partitions rather than an offset from the start of the active boot partition? May 21 17:51:12 probably but I assumed you'd want to be able to have atomic updates for u-boot/barebox in its entirely, i.e. including its SPL May 21 17:51:35 zmatt: the SPL isn't updated May 21 17:52:00 that's kind of the point of having a separated first/second stage bootloader (beside the size constraint) May 21 17:52:09 if that's acceptable then sure... I thought u-boot didn't like that but I'm wrong May 21 17:52:09 no the point is the size constraint May 21 17:52:15 I'm pretty sure that's the entire reason for SPL May 21 17:52:36 my understanding was that SPL and u-boot.img should always be the same version May 21 17:52:37 otherwise updating the SPL is kinda chicken and egg problem, you mess the SPL and you're screwed May 21 17:54:04 not really, there are multiple ways to ensure that copy 1 of SPL will load copy 1 of u-boot.img and ditto for copy 2 May 21 17:54:26 zmatt: I meant that if the size wasn't a constraint, the reason why you would use a separate boot stage is to split the invariable boot logic (e.g. fixed offset) from the main bootloader code (which can be updated, deals with recovery, whatever features you may think of) May 21 17:54:39 (and it would definitely not be an issue if you'd use eMMC boot partitions, which in fact is why I assumed you cared about it) May 21 17:55:45 and what would be the benefit of that over not using an SPL and having bootrom just directly load u-boot (if the size constraint wouldn't be an issue) May 21 17:56:52 zmatt: you're correct, if the eMMC boot partitions where 300Mb instead of 16Mb, there would be no need for splitting the SPL from its bootloader May 21 17:57:18 uhh what, I was talking about the size limit of what bootrom can load May 21 17:57:25 eMMC boot partition is plenty large for any bootloader May 21 17:58:40 zmatt: erf you're right, I mixed up mb and kb... barebox images are ~200kb which would obviously fit in the 16mb boot partitions ^^ May 21 17:59:00 zmatt: bootrom load is the constraint indeed May 21 17:59:57 which itself is due to limited SRAM size (128KB, some of which is reserved so iirc the max executable size bootrom can load is 107KB or something like that) May 21 18:00:14 109 KB May 21 19:35:42 zmatt: sorry, was afk. So having the MLO look for a bootable image on the eMMC boot partition is totally feasible I assume May 21 19:59:05 I'm not sure what you're asking since your phrasing is ambiguous, although I'm pretty sure it's asking something I already answered :P May 21 20:01:11 for all reasonable interpretations of that question the answer is yes, but if you mean have u-boot SPL on main eMMC load u-boot.img from the active eMMC boot partition then that sounds inadvisable (since again, I'm pretty sure you're supposed to update both SPL and u-boot.img together... maybe it's different with barebox, I'm not familiar with it) May 21 20:04:53 if I wanted to use the eMMC boot partitions for u-boot then I'd put both SPL and u-boot.img there (at offset 0 and 128KB respectively) and make a tiny tiny loader that loads u-boot SPL from the active eMMC boot partition and jumps to it May 21 20:07:10 (bootrom will have done enough initialization at this point that the shim loader probably doesn't need to perform any further initialization and can immediately issue the eMMC command to read the bootloader and jump to it, May 21 20:10:08 the only caveat is that it needs to overwrite the location where bootrom loaded the shim, so it either needs to relocate itself first or ensure it gets cached in L2 so the data in ram can be overwritten by the mmc controller's dma) May 21 20:10:26 but I may be biased because I already know how to do these things May 21 20:10:43 (I've written small bits of baremetal code for the am335x before) May 21 20:22:17 though again I'm not sure any of this is worth the effort, atomic bootloader updates can also be achieved without using the boot partitions (though I'll admit using them would be more elegant) May 21 20:51:36 zmatt: you're totally right. I think using an existing mechanism to update the MBR active flag of a redundant (user) boot partition is way simpler. May 21 20:51:59 oh you're using boot partitions? May 21 20:52:21 I thought bootrom only examined the first partition? I probably just misremember May 21 20:53:49 okay yeah I misremember May 21 20:53:53 zmatt: the bootrom looks for a bootable code at 4 offsets, which in turn differs given the bootloader. Barebox for example look for the first active vfat partition on the same medium May 21 20:54:33 (and try to load a "barebox.bin" file from there) May 21 20:55:01 you're confusing raw mmc boot method and fat partition method May 21 20:55:06 the two are mutually exclusive May 21 20:55:08 while u-boot tries both this approach and loading a bootable (u-boot) image at the next offset May 21 20:55:58 bootrom first looks for a configuration header at 4 fixed offsets... if one is found then neither bootrom nor u-boot SPL will look for a FAT partition May 21 20:56:06 zmatt: barebox doesn't implement the raw mmc boot method on bbb for the 3rd stage bootloader (aka barebox.bin) if I'm not mistaken. May 21 20:57:19 if none is found then bootrom falls back to looking for a bootable FAT12/16/32 partition May 21 20:58:15 (and it requires that there is only one, the TRM says it will fail if there are multiple) May 22 01:54:27 What version of GLIBC does the Debian distros use currently from BBB.io/latest-images? May 22 01:54:46 2.29 or 2.28? May 22 01:56:12 which debian are you talking about? May 22 01:56:24 Debian 4.19 Buster May 22 01:57:50 I can update to 5.4 if we use the 2.29 version of GLIBC. May 22 01:58:15 Hello GenTooMan: Are you feeling any better yet? May 22 02:00:25 set_, it's sort of like falling down a hill sometimes it's quite bumpy sometimes it feels cushy but ... its still falling down a hill :D May 22 02:00:47 Yikes sir, I am sorry. Hills when falling down them cannot be fun. May 22 02:00:55 set_, not better sorry. May 22 02:00:59 oh. May 22 02:01:05 Dang... May 22 02:01:46 Well, like the doctors say, eat grains like a grain eater...Granola, Bran, Wheat, and such. May 22 02:02:20 I know it is not flavored meats but it supposedly helps clear out the system at high rates. May 22 02:03:08 GenTooMan: I am reflashing now. I can check then on my question. May 22 02:03:21 The console image is almost loaded now. May 22 02:03:42 I get like 32/sec. download speeds. This is terrible. May 22 02:03:56 2021 and 32Kb/sec. is garbage. May 22 02:04:20 That is if the 32Kb comes through or not. May 22 02:04:34 set_, back in the day like 20 years ago 32kb/sec was awesome :D May 22 02:04:50 20 years, yes sir. May 22 02:05:00 I remember the first computer I had. It had a game called Defender. May 22 02:05:24 I was like, "Wow!" I can move from left to right and downward and upward on the screen. May 22 02:05:34 Now! It is like almost real individuals you can control. May 22 02:06:11 Yikes. But I remember in '16 the slowness of the upload and download speeds. May 22 02:06:31 The computer would take 15 minutes to boot. May 22 02:06:33 Ha. May 22 02:06:47 Now, we can boot BBBs in 3 sec! May 22 02:07:24 hmm I have had a BBB running for a year next too me... I just didn't use it all the time :D May 22 02:07:50 ha. Nice! What is it doing runnin' all the time? May 22 02:08:51 Bitcoin? Ha. May 22 02:09:15 set_, the reality of bitcoin "mining" is that it's fraud. May 22 02:09:58 I explained that people all "excited" about crypto currency "mining" and they were "shocked" May 22 02:10:15 Ha. I know nothing about for good reason. May 22 02:10:17 "that to people" May 22 02:10:57 well a lot of people are in a fantasy. May 22 02:11:53 GenTooMan: I understand. I had a person work on this house for a bit. Supposedly, he thinks he is rich b/c of bitcoin. May 22 02:13:34 set_, a currency has to have value being able to come up with currency numbers by calculating valid bitcoin values means it has no value. May 22 02:13:54 Ha. May 22 02:14:29 Like an apple. When you are hungry, it is worth a lot. But...when I just ate steak and a ceaser salad, apples are worthless. May 22 02:14:59 just like the US dollar can and will fail if people continue to assume "inflation" and lying to people is OK May 22 02:15:14 that's reality. May 22 02:16:02 raility is brutal and most people don't want to live in it. May 22 02:16:35 GenTooMan: things have value is people decide it has value May 22 02:16:37 *if May 22 02:16:50 Yep and yes sir, reality is not so easy to some people. Take me for instance, I am okay w/ it at times. May 22 02:17:12 But other times, I think it is $%*& %*$&! May 22 02:17:36 GenTooMan: I'm curious what your reasoning is that bitcoin mining is "fraud" May 22 02:17:42 zmatt, exactly ... May 22 02:18:33 zmatt, think about it if you can create monetary units by computing numbers... does it really have any value? May 22 02:22:12 zmatt, what is fraud, deceit trickery. So does bitcoin have value? "yelling" it has value doesn't make it have value. May 22 02:22:17 GenTooMan: that entire sentence makes almost no sense. mining is the act of maintaining the distributed ledger, for which a known reward (in _bitcoin_) is integrated into the system to incentivize mining thus allowing the system to function. mining has no relevance to the monetary value of bitcoin May 22 02:23:11 like, why do things like gold and silver have monetary value? why does anything have monetary value? May 22 02:23:35 bitcoin has monetary value because people collectively decided it has value May 22 02:24:47 "Gold costs too much, I am going w/ bitcoin." May 22 02:24:52 ha. May 22 02:25:08 Hey guys... May 22 02:25:17 the fact that it is a scarce resource is an important factor... just like it is with gems or precious metals, except that with bitcoin the total amount there will be at any point in time is known in advance to everyone May 22 02:25:42 The image is built, the board is booting, and the SDK "will" work! May 22 02:25:55 note that the act of mining doesn't create bitcoin, it is not the case that the more mining happens the more bitcoin there will be. you don't "create monetary units by computing numbers" May 22 02:27:12 rather, the creation of bitcoin is an integral part of the protocol, the amount of bitcoin created is fixed in the design, and the protocol gives it to the miners as compensation for their role in making the system work May 22 02:27:36 If this dev. desktop will ever update. Blah. May 22 02:27:40 zmatt, actually that's the case of what they are doing. Mathematically speaking there is a limit to how much bitcoin can be mined as the actual numbers will be exhausted, as there is a limit to the number of valid bitcoin numbers. May 22 02:27:56 GenTooMan: no, that's not how it works May 22 02:28:03 I should have never brought this up. May 22 02:28:16 * GenTooMan grins at set_ May 22 02:28:19 You all quit it. Bitcoin is boosie. May 22 02:28:24 Ha. May 22 02:29:11 Why would either of you learn how bitcoin works? I do not get it. Are we actually tired of cash money, e.g. we have to look elsewhere? May 22 02:30:01 I am know you two are not disussing the tiredness of cash but the way bitcoin works in reality is the subject. May 22 02:30:04 I understand. May 22 02:30:43 But...why understand it? Are you two miners of bitcoin? I am serious. May 22 02:31:01 GenTooMan: approximately every 10 minutes 6.25 bitcoin is created, and this will remain the case until May 6th 2024 May 22 02:31:07 regardless of how much mining is going on May 22 02:34:33 so roughly 155664 bit coin left? May 22 02:34:54 well * 6.25 May 22 02:35:10 no, they don't run out at that date, the "6.25" changes at that date May 22 02:36:26 the mining reward per block is halved every 210000 blocks (approx every 4 years) May 22 02:37:08 it'll still take more than a century to actually become zero May 22 02:38:20 (the reward per block started at 50 bitcoin when the network was created but has halved 3 times since then) May 22 02:43:41 the total amount of bitcoin that will ever exist is 21 million May 22 02:45:23 (or rather 20999999.9769 if my calculation is correct) **** ENDING LOGGING AT Sat May 22 02:59:57 2021