maemish_ | https://usercontent.irccloud-cdn.com/file/Oyyo13Ys/MVIMG_20230214_031557.jpg | 07:58 |
---|---|---|
maemish_ | https://usercontent.irccloud-cdn.com/file/t9zCHSs9/MVIMG_20230214_031505.jpg | 07:59 |
maemish_ | https://usercontent.irccloud-cdn.com/file/WsUbEGyK/MVIMG_20230214_031557.jpg | 08:00 |
maemish_ | Epiphany browser starts quickly but after it opens and you see the page it crashes. | 08:00 |
maemish_ | This is seen in xterm after that | 08:01 |
freemangordon | get a backtrace | 08:03 |
Wizzup | freemangordon: could easily be out of ram, too | 09:27 |
Wizzup | trying to browse the modern web on a n900 is like using a 2004-2005 celeron (which is actually way more powerful) to try and browse the modern web | 09:28 |
Wizzup | maybe some super simple thing like dillo can work, but it won't support all the new stuff | 09:28 |
sicelo | yeah. realistically, people should mostly just stop using/obtaining N900 if browsing is a main requirement. there's not a lot that we can do about the web getting bloated, and that trend won't stop | 09:51 |
n900 | but still awesome as an irc client! :) | 10:33 |
Wizzup | :p | 10:34 |
uvos__ | Wizzup: ping | 11:02 |
uvos__ | maemish_: sicelo: yeah even the mapphones are pushing it when used to browse modern webpages | 11:03 |
uvos__ | im suprised qtwebrowser manages to dsiplay any page of any complexity on n900 | 11:04 |
uvos__ | ff just ooms immidatly so its better than that :) | 11:04 |
freemangordon | right | 11:07 |
maemish_ | But Maemo 5 with Opera Mobile, Fennec or Micro-b with nginx loads just fine without problems. So it is not just the question of bloated web. | 11:11 |
freemangordon | maemish_: right, but none of the 'modern' browsers is optimized for 256 MB of ram | 11:13 |
freemangordon | also, which exactly page microb or opera have no issue loading? besides tmo that is | 11:13 |
freemangordon | also, do you have swap enabled? also, make sure to disable ramzswap | 11:14 |
maemish_ | I have been using Opera just fine for browsing. Or Fennec. And Micro-b with nginx works suprisingly well. | 11:14 |
freemangordon | for some reason it is enabled on n900 making things even worse | 11:14 |
freemangordon | ok, but lets have one site that loads ok in fremantle and use it as a benchmark | 11:15 |
freemangordon | "fine for browsing" is too general | 11:15 |
maemish_ | I have not done any tweaks for swap. Thought there is swap by default. | 11:15 |
freemangordon | nope, iirc | 11:16 |
freemangordon | besides ramzswap and that shall be disabled | 11:16 |
freemangordon | so make sure to disable ramzswap and enable emmc swap | 11:16 |
freemangordon | and try again | 11:16 |
freemangordon | but don;t expect miracles, FF is FUBAR after v78 | 11:17 |
maemish_ | Is there any info how to? Like to an idiot level? | 11:17 |
freemangordon | give me the output of "swapon -s" | 11:17 |
maemish_ | Not having device at hand. At work. Will do in the evening. Thanks for bothering to help. | 11:18 |
freemangordon | ok, I will be a bit busy in the evening, but in general you need swapon/swapoff commands | 11:19 |
freemangordon | and you have to identify emmc swap partition | 11:19 |
freemangordon | (s)fdisk -l | 11:19 |
maemish_ | Have been active on the forum helping others without using N900 for a year. Lot to remember. | 11:20 |
maemish_ | Your info made this idiot to say "Huh?" | 11:21 |
freemangordon | hmm? | 11:21 |
freemangordon | why is that | 11:21 |
maemish_ | But thanks. Gonna see if I can with that info and with help from my friend go forward. It is no problem. | 11:22 |
freemangordon | I mean - maybe ask google how to list enabled swap etc | 11:22 |
freemangordon | ok | 11:22 |
maemish_ | Thanks. | 11:22 |
sixwheeledbeast | `blkid -t TYPE=swap` ? | 11:24 |
* freemangordon just learned something new :) | 11:25 | |
sixwheeledbeast | so performance is better if swap is on the device? | 11:26 |
freemangordon | for leste, yes | 11:26 |
sixwheeledbeast | ah | 11:26 |
freemangordon | like for fremantle swap has to be on uSD ;) | 11:26 |
sixwheeledbeast | that's why i wondered. | 11:27 |
freemangordon | sixwheeledbeast: on fremantle /opt is on emmc | 11:28 |
uvos__ | idk if thats true | 11:28 |
freemangordon | uvos__: what is "thats"? | 11:28 |
uvos__ | for d4 a modern sdcard is mutch faster with random writes | 11:28 |
uvos__ | than its emmc | 11:28 |
freemangordon | not for n900 | 11:28 |
uvos__ | and seqential is mutch faster too | 11:29 |
uvos__ | if you enable 100mhz hack | 11:29 |
freemangordon | we talk n900 here | 11:29 |
uvos__ | n900 can do this hack too | 11:29 |
freemangordon | but, swap on emc means 2 io channels | 11:29 |
uvos__ | freemangordon: i would not be so sure, i have a sdcard that can saturate the bus at 50Mhz with random writes | 11:29 |
freemangordon | one for root and another for swap | 11:29 |
uvos__ | ie 25MB/s | 11:29 |
freemangordon | on n900 emmc does more than 15MB/s | 11:30 |
freemangordon | iirc | 11:30 |
uvos__ | well thats al lot slower than a good application class sdcard | 11:30 |
Wizzup | uvos__: pong | 11:30 |
freemangordon | combined with separate channel gives better performance in general | 11:30 |
freemangordon | swap io should not wait for rottfs io | 11:30 |
uvos__ | freemangordon: sure maybe, but its not that clear cut as you make it out to be | 11:30 |
freemangordon | *rootfs | 11:30 |
freemangordon | also, you talk about some hacks and special cards | 11:31 |
uvos__ | freemangordon: no hacks | 11:31 |
uvos__ | freemangordon: for faster random | 11:31 |
uvos__ | but yes you need a "special" ie modern application class sdcard | 11:31 |
freemangordon | uSD on n900 is 2 lines, IIRC | 11:31 |
freemangordon | emmc is 4 | 11:31 |
uvos__ | right | 11:31 |
uvos__ | but this dosent matter if emmc cant saturate its bus | 11:32 |
freemangordon | sure | 11:32 |
freemangordon | but emmc on n900 is *very* fast | 11:32 |
uvos__ | Wizzup: just wonderning what you wanted | 11:32 |
uvos__ | freemangordon: 15 years ago, sure | 11:32 |
freemangordon | can;t remember the benchmarks from back then | 11:32 |
maemish_ | You guys can't even believe how lucky I feel to follow this conversation. | 11:33 |
Wizzup | uvos__: just a sec | 11:34 |
maemish_ | Peak behind the scenes to esoteric info. | 11:34 |
freemangordon | uvos__: unfortunately google does not give any results for n900 emmc benchmark | 11:34 |
uvos__ | freemangordon: ok well we should disable zswap and add a script that asks the user where to place swap maybe | 11:35 |
freemangordon | but I can assure you that copy over usb mass storage is done with 16MB/s | 11:35 |
freemangordon | right | 11:35 |
freemangordon | zswap on n900 is very bad idea | 11:35 |
uvos__ | thats ~seqential then and really slow, | 11:36 |
bencoh | 16MB/s is very decent for a 2009 device | 11:36 |
uvos__ | compeared to a modern sdcard | 11:36 |
freemangordon | not really sequential | 11:36 |
sixwheeledbeast | someone on tmo claimed 13000Kb/s using flasher3.5 output ? | 11:36 |
uvos__ | shoudl be fairly seqentuail, depending on allocations on the card ofc | 11:36 |
bencoh | I'd say it counts as sequential as well yeah | 11:37 |
freemangordon | sixwheeledbeast: this is not good benchmark | 11:37 |
freemangordon | because we have USB playing there | 11:37 |
sixwheeledbeast | fair enough | 11:37 |
maemish_ | https://youtu.be/FSsGyuqOZRU | 11:37 |
freemangordon | sixwheeledbeast: I remember running benchmarks locally on the device back then | 11:38 |
freemangordon | but neither can remember nor I can find the results :) | 11:38 |
freemangordon | uvos__: still, my gut feeling tells me that a separate io channel for swap will be better than even the fastest uSD card, as it is not only the raw speed thet matters | 11:39 |
sixwheeledbeast | i'd like to stick to my cycling swap defragging method if required, so just interested in performance for leste having all on the card. | 11:39 |
freemangordon | maybe provide so benchmarks if possible | 11:40 |
freemangordon | *some | 11:40 |
sixwheeledbeast | it does make sense bandwidth wise i agree | 11:40 |
uvos__ | freemangordon: that might be true, is deffinatly true when the sdcard is otherwise loaded, but idk how this presents itself in regular usage | 11:41 |
freemangordon | IIRC omap3 runs its mmc bus on max 54MHz, right? | 11:41 |
uvos__ | something to play around with | 11:41 |
uvos__ | freemangordon: yes, scard spec requires 1.8v trancievers if you want to do more | 11:41 |
uvos__ | which we dont have | 11:41 |
uvos__ | *sdcard | 11:41 |
freemangordon | so, theoretical bus bandwith for uSD is 54/4 = 13.5 MB/s, no? | 11:42 |
uvos__ | bus with should be 4 and 8 not 2 and 4 | 11:42 |
uvos__ | iirc, def true for d4 | 11:43 |
freemangordon | umm, have to check it, but I think on n900 it is 2 and 4 | 11:43 |
uvos__ | i thin 2 isent valid for sdhc | 11:43 |
freemangordon | does anyone have n900 boot dmesg log around? | 11:44 |
freemangordon | kernel tells us the bus width | 11:44 |
freemangordon | "mmc1: switch to bus width 1 failed" | 11:46 |
bencoh | maybe, lemme check | 11:47 |
bencoh | hmm, nope, sorry. are you looking for fremantle or leste btw? | 11:48 |
freemangordon | well, we are looking for upstream kernel :) | 11:48 |
bencoh | ah, nevermind then | 11:49 |
freemangordon | joerg: do you remember uSD vs eMMC bus widhts on n900? | 11:49 |
freemangordon | uvos__: right, according to schematic we have MMC_DATA0 to MMC_DATA3 for MMC1 and MMC_DATA0 to MMC_DATA7 for MMC2 | 11:52 |
freemangordon | so uSD is 4 bits | 11:52 |
bencoh | mmc1 has 4, mmc2 has 8 | 11:52 |
freemangordon | :nod: | 11:53 |
bencoh | looks like we're looking at the same pdf :] | 11:53 |
freemangordon | joerg: nevermind | 11:53 |
freemangordon | bencoh: yeah | 11:53 |
freemangordon | uvos__: so, max theoretical transfer is 27MB/s | 11:53 |
uvos__ | freemangordon: right so any modern uhs sdcard can massivley outperform emmc | 11:54 |
uvos__ | seqentally | 11:54 |
uvos__ | and A class ones with do mutch faster random | 11:54 |
freemangordon | I am still not convinced, because those 16MB/s are over USB | 11:55 |
freemangordon | if we bench eMMC on the device, I will not be surprised if we see like 25MB/s | 11:55 |
freemangordon | maybe I can do that soon | 11:55 |
freemangordon | remember back then in HDD times? all the recommendations were to have swap on a separate HDD (and on a separate IDE channel), even if that HDD is slower than rootfs one? | 11:56 |
uvos__ | sure | 11:56 |
freemangordon | the same here | 11:56 |
uvos__ | but on d4 emmc can do about 40MB/s | 11:57 |
uvos__ | sequentally | 11:57 |
freemangordon | d4 is another beer | 11:57 |
uvos__ | and like 10 random | 11:57 |
freemangordon | HW wise those are incomparable | 11:57 |
* sixwheeledbeast still uses spinning disks with two swap partitions on another disk... | 11:57 | |
uvos__ | so if n900 is same ballpark the increased random speed helps more than io contention hurts | 11:57 |
freemangordon | uvos__: afaik eMMC on n900 has very high random speed compared to uSD cards | 11:58 |
freemangordon | everything but 256 MB of rootfs is on eMMC | 11:58 |
uvos__ | i have a sdcard that can saturate random writes on d4 at 100MHz bus speed | 11:58 |
freemangordon | wow | 11:58 |
uvos__ | ie about 45MB/s | 11:59 |
uvos__ | its not that special even | 11:59 |
uvos__ | just a A2 card | 11:59 |
freemangordon | may I have a model? I am struggling to find a good uSD card here, no matter the price | 11:59 |
freemangordon | *the model | 11:59 |
uvos__ | sure | 11:59 |
uvos__ | but i dont have it on me rn | 11:59 |
freemangordon | no hurry | 11:59 |
uvos__ | ill post it later | 11:59 |
freemangordon | ok | 12:00 |
joerg | >>mmc1 has 4, mmc2 has 8<< just watch out for mmcs getting renamed, at least in fremantle | 14:05 |
joerg | they swap places/names/numbers | 14:06 |
joerg | iirc max speed on uSD in N900 always maxed out at 12.5MB/s, in line with 25MHz@4bit ? | 14:46 |
joerg | there are faster uSD interfaces but... prolly not for OMP34xx | 14:47 |
joerg | OMAP* | 14:47 |
uvos__ | joerg: should be 50MHz (max speed of standart sdhc @ 5V) @4bit | 14:49 |
uvos__ | practicly about 22MB/s or so with overhead | 14:49 |
joerg | yeah, but... maybe a signal integrity issue? I don't know | 14:52 |
joerg | anyway the 12some MB/s I recall quite reliably | 14:53 |
joerg | it's what mc etc shown me when I copied huge amounts of data to/from uSD | 14:54 |
joerg | possibly for write though, which must be even slower than read | 14:55 |
uvos__ | sure your usd card must be able to saturate the bus ofc | 14:56 |
* joerg idly wonders what uSD if any is in Iron900 | 14:57 | |
joerg | hmmm, mount doesn't suggest there's ANY uSD in my IroN900 | 14:59 |
joerg | :-/ https://termbin.com/fhlz3 | 15:04 |
maemish_ | freemangordon: https://usercontent.irccloud-cdn.com/file/0fGdAzS2/MVIMG_20230214_163318.jpg | 15:33 |
maemish_ | freemangordon: How do I identify emmc swap partition? | 15:49 |
maemish_ | And change away from using zram? | 15:50 |
buZz | maemish_: emmc swap? | 16:20 |
buZz | i dont think by -default- any emmc is used for swap? | 16:20 |
buZz | uvos__: any luck in finding that sphone selectbox issue? | 16:21 |
buZz | could just hardcode it to ofono in https://github.com/maemo-leste/sphone/blob/master/src/modules/gui/gtk2/ui-messages-gtk.c#L209 | 16:23 |
buZz | :P | 16:23 |
buZz | i guess this? https://github.com/maemo-leste/sphone/blob/master/src/modules/gui/gtk2/ui-messages-gtk.c#L149 | 16:24 |
buZz | seems to default to '0' | 16:24 |
buZz | ah no, thats 'column' , wonder what that means in this context | 16:26 |
uvos__ | buZz: havent looked | 16:26 |
uvos__ | im sure its farily trivial | 16:26 |
buZz | yeah it must be something with that hildon_touch_selector i guess | 16:26 |
uvos__ | https://github.com/maemo-leste/sphone/blob/e851cd95c689b6fbf3c72c4aca82d62af6cf632d/src/modules/gui/gtk2/ui-messages-gtk.c#L209maybe check with gdb what backend_name is | 16:28 |
uvos__ | maybe | 16:29 |
uvos__ | not sure how the selector can show ofono as the current selection but then not return ofono to hildon_touch_selector_get_current_text | 16:30 |
uvos__ | maybe we should check if the non-hildon path works | 16:30 |
uvos__ | i gues the one in the dialer works | 16:36 |
uvos__ | the only difference is that https://github.com/maemo-leste/sphone/blob/e851cd95c689b6fbf3c72c4aca82d62af6cf632d/src/modules/gui/gtk2/ui-messages-gtk.c#L149 is not called | 16:37 |
uvos__ | if thats it that seams like a bug in libhildon\ | 16:37 |
uvos__ | buZz: if you have the time maybe comment out that line and see what happens | 16:37 |
uvos__ | ah i get it | 16:40 |
uvos__ | yeah | 16:40 |
uvos__ | so when sending a new message msg->backend is NULL | 16:40 |
uvos__ | er uninialized | 16:41 |
uvos__ | so we pass some random colllumb as selection | 16:41 |
uvos__ | ofc this dosent exist | 16:41 |
uvos__ | thats the bug | 16:41 |
uvos__ | yeah non-hildon path works fine on my laptop | 16:44 |
uvos__ | but the bug is in sphone in the hildon path | 16:44 |
buZz | hmhm | 16:45 |
buZz | does sphone have src packages on our repos? | 16:45 |
uvos__ | idk | 16:45 |
uvos__ | ;D | 16:45 |
uvos__ | i have no idea where src packages come frome | 16:45 |
uvos__ | i gues it should | 16:45 |
uvos__ | ci should build one | 16:45 |
uvos__ | idk if it gets copied to the repo or not | 16:45 |
uvos__ | just clone it from git (imo) | 16:46 |
buZz | apt source sphone, works | 16:46 |
buZz | there's another hildon_touch_selector_set_active() in the code | 16:48 |
buZz | line 75 | 16:48 |
buZz | i'll comment out 149 and try it | 16:49 |
uvos__ | buZz: that ones fine | 16:49 |
uvos__ | it just selects whatever is the first backend loaded | 16:49 |
uvos__ | if you comment out 149 the bug should dissapear | 16:50 |
uvos__ | but the backend wont be selected correctly if you click reply | 16:50 |
uvos__ | (it will be the first backend instead of the one of the message) | 16:50 |
buZz | but i disabled commtest in sphone.ini ;) | 16:50 |
uvos__ | sure yeah | 16:50 |
uvos__ | it will allways be ofono | 16:51 |
uvos__ | and atm thats the only backend anyhow | 16:51 |
buZz | good | 16:51 |
buZz | ok, got my build installed, lets see | 16:56 |
buZz | 'sudo reboot' really takes ages sometimes :D | 16:59 |
uvos__ | would have been no need for a reboot | 16:59 |
uvos__ | killall sphone should have been enough | 16:59 |
buZz | thanks for the help on this btw, uvos__ | 17:00 |
buZz | its been such a annoyance :P | 17:00 |
buZz | just weird that 'sudo reboot' takes so long vs 'sudo poweroff' being near instant | 17:01 |
uvos__ | poweroff also takes long sometimes | 17:01 |
buZz | ahhhhh sweet, it works now \o | 17:02 |
buZz | with 149 commented out | 17:02 |
buZz | want me to PR that? | 17:02 |
buZz | maybe i should try it with commtest enabled again | 17:03 |
buZz | oh, commtest -is- enabled again :D hahaha | 17:03 |
uvos__ | buZz: no commenting it out is not a solution | 17:04 |
uvos__ | i need to write a real fix | 17:04 |
buZz | alright, either way, with that line removed its functional | 17:04 |
uvos__ | buZz: yeah you dident follow my instructions and not edit /usr/share/sphone/sphone.inii | 17:05 |
uvos__ | :) | 17:05 |
buZz | yes, i am a rebel | 17:05 |
buZz | :D | 17:05 |
uvos__ | you supposed to edit ~/.sphone/sphone.ini | 17:05 |
uvos__ | :P | 17:05 |
buZz | i'm just happy it works :) | 17:05 |
buZz | so i can finally initiated conversations | 17:06 |
joerg | poweroff might only fake being fast | 17:14 |
buZz | it seems to be legit, it can powerup aswell when it finishes | 17:15 |
buZz | i think i can boot whole device before a 'sudo reboot' even powers off the device :P | 17:16 |
buZz | s/powers off/powercycles/ | 17:16 |
joerg | it's easy to shut off RF, LEDs and LCD | 17:16 |
buZz | right but you can see it being legit if you press powerbutton and motorola logo comes :) | 17:16 |
joerg | yep | 17:16 |
joerg | make sure you got no power source (USB) connected | 17:18 |
joerg | or "power off" may be complete fake | 17:19 |
buZz | yeah i rarely do | 17:19 |
uvos__ | poweroff dosent fake anything | 17:23 |
uvos__ | mce (ie shutdown via power menu) fakes it a bit by instantly turning off the display | 17:24 |
buZz | sometimes i'm considering trying to get boot animations | 17:24 |
uvos__ | but it keeps the status led on untill real power off | 17:24 |
buZz | or maybe just a bootsplash screen during kernel msgs | 17:24 |
uvos__ | buZz: plymouth should just work ootb | 17:24 |
uvos__ | on d4 | 17:24 |
uvos__ | so just install it | 17:25 |
buZz | yeah? just apt install plymouth? | 17:25 |
buZz | i'll try | 17:25 |
uvos__ | and activate the service | 17:25 |
uvos__ | should work, but for early boot you would need plymouth in initramfs | 17:25 |
uvos__ | i dont like hideing the kernel anyhow | 17:25 |
buZz | 'package plymouth is not available' | 17:25 |
uvos__ | kernel messages that is | 17:25 |
uvos__ | buZz: hmm | 17:26 |
buZz | i'm still on buster | 17:26 |
uvos__ | dunno if its even in devuan | 17:26 |
bencoh | it is | 17:27 |
buZz | https://git.devuan.dev/devuan/plymouth | 17:27 |
uvos__ | cant find it on my d4 either | 17:27 |
uvos__ | strange | 17:27 |
buZz | uvos__: are you on chimeara? | 17:27 |
uvos__ | no | 17:27 |
buZz | oh ok | 17:27 |
buZz | maybe the package couldnt get built for arm? | 17:28 |
bencoh | uvos__: is there a way to disable that poweroff black screen and get system/kernel logs instead? | 17:28 |
uvos__ | well not with this device anyhow | 17:28 |
uvos__ | bencoh: besides using poweroff _ | 17:28 |
bencoh | I'd rather have my device report about poweroff going awry than just a black screen | 17:28 |
uvos__ | ? | 17:28 |
bencoh | uvos__: yeah | 17:28 |
buZz | it exists on my x86 devuan ; plymouth/stable 0.9.5-2+devuan2 amd64 | 17:28 |
bencoh | I mean, if poweroff is the only answer it's fine | 17:28 |
uvos__ | bencoh: no | 17:28 |
uvos__ | not atm | 17:29 |
bencoh | okay | 17:29 |
bencoh | is poweroff really fine though, or does it skip on some stuff that leste ecosystem needs to take care of before shutdown? | 17:29 |
uvos__ | should be fine, the shutdown dbus signal is not activated | 17:30 |
uvos__ | but if something relies on that so hard that it causes problems its a bug | 17:30 |
uvos__ | i dont think there is any user of this signal at all | 17:31 |
uvos__ | besides mce and dsme | 17:31 |
uvos__ | buZz: hmm | 17:32 |
Wizzup | bencoh: yes it skip stuff, but it's probably not super harmful | 17:34 |
buZz | Wizzup: any idea if plymouth is available on chimaera? | 17:43 |
buZz | (on d4) | 17:43 |
Wizzup | rafael2k_: do you need me to do anything for the pp atm? | 18:40 |
Wizzup | other than fixing the arm64 image builder for chimaera:) | 18:40 |
rafael2k_ | nope, all good! | 18:52 |
rafael2k_ | A suspend button in the power menu would be great (may be just triggering suspend directly) | 18:53 |
rafael2k_ | btw, my pp keyboard did not arrive yet... but it coming | 18:53 |
Wizzup | check | 18:56 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!