Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2024-10-25

00:27:15 Quit bleb (Ping timeout: 276 seconds)
00:30:13 Join bleb [0] (~cm@user/bleb)
00:43:03 Quit _bilgus (Read error: Connection reset by peer)
00:43:45 Join _bilgus [0] (~bilgus@syn-162-154-213-134.res.spectrum.com)
00:50:08 Quit pixelma (Quit: .)
00:50:08 Quit amiconn (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
00:51:13 Join pixelma [0] (marianne@p4fe76284.dip0.t-ipconnect.de)
00:51:13 Join amiconn [0] (jens@p4fe76284.dip0.t-ipconnect.de)
01:00
01:52:11***Saving seen data "./dancer.seen"
02:00
02:54:07 Quit D_the_n (Ping timeout: 264 seconds)
03:00
03:04:17 Quit othello7 (Ping timeout: 252 seconds)
03:29:36 Quit kugel (Ping timeout: 252 seconds)
03:31:25 Join kugel [0] (~kugel@ip4d146a3a.dynamic.kabel-deutschland.de)
03:47:48 Join OlsroFR [0] (~OlsroFR@user/OlsroFR)
03:49:19OlsroFR+speachy Battery life is still very terrible. 4H47 only before shut down using the battery bench and playing mp3. It's not better than before your patch.
03:49:40OlsroFRit was not fully charged at the start of the test; I started the test with 93% of battery
03:50:26OlsroFRwith 100% it would probably have lasted 5H or 5H30, which is far below what I should get with power management working with a Mini. It should last easily 10 hours at least
03:52:12***Saving seen data "./dancer.seen"
03:52:29 Quit OlsroFR (Client Quit)
05:00
05:04:42 Join jacobk_ [0] (~quassel@47-186-105-237.dlls.tx.frontiernet.net)
05:04:48 Quit jacobk (Ping timeout: 276 seconds)
05:52:15***Saving seen data "./dancer.seen"
07:00
07:52:18***No seen item changed, no save performed.
09:00
09:07:08 Join dconrad [0] (~dconrad@152.117.104.217)
09:36:39 Nick jacobk_ is now known as jacobk (~quassel@47-186-105-237.dlls.tx.frontiernet.net)
09:52:19***No seen item changed, no save performed.
10:00
10:13:42 Join othello7 [0] (~Thunderbi@pool-100-36-176-164.washdc.fios.verizon.net)
10:16:59 Join chris_s [0] (~chris_s@2a04:4e41:65:15::333b:7a16)
10:19:37 Join IPG [0] (~InvoxiPla@94.4.117.11)
10:21:31chris_sspeachy: I haven't put much more than an hour or two on it, but g#5986 has caused no issues so far on either an iFlash-Solo-equipped iPod video or iPod 4g (power mgmt unsupported, flush cache: extended)
10:21:35rb-bluebotGerrit review #5986 at https://gerrit.rockbox.org/r/c/rockbox/+/5986 : ata: Rework how flushing, sleeping, and power off interacts by Solomon Peachy
10:26:38speachychris_s: thanks!
10:27:12speachyOlsroFR: I find the "no difference in battery life" surprising, because it _should_ be powering things off.
10:27:36CH23i wonder how worn their battery is
10:28:28speachyonce my mini2g charges enough for it to come to life, I have a logf-enabled build to make sure it _is_ power-cycling the storage device.
10:31:13speachykicking myself for not doing that to begin with.
10:36:53chris_sfwiw, as a test I did insert splashes before L1438 (ide_power_enable(false)) and it was called as expected - subjectively battery life seems a bit better, but I haven't had time to check systematically
10:51:18speachygood to know, thanks
10:58:06speachymade some additional tweaks −− re-query the identify info after we've configured everything. ipod6g also now explicitly sets PM and other such parameters for CE-ATA devices too.
10:58:37speachyspeaking of, anyone out there with one of those ipod6gs that has a CE-ATA drive?
11:00
11:01:39 Quit jacobk (Ping timeout: 276 seconds)
11:03:56speachyOlsroFR: I updated that zip file, please grab it and scroll into the debug menu, "show log file"
11:04:54speachyon mine, SLEEPNOW -> FLUSH CACHE -> OFF
11:05:16speachy(devices that support ATA PM you'll see SLEEPNOW->FLUSH CACHE->SLEEP->OFF)
11:06:05speachyI can generate builds for other ipods if anyone is willing to try 'em out. Also need to make sure there are no regressions on old drives.
11:08:50speachymy 5g broke the ZIF conector on the mainboard so I can't test that out myself until I replace that
11:10:17 Quit chris_s (Quit: Client closed)
11:10:58 Join chris_s [0] (~chris_s@2a09:bac2:29da:2478::3a2:7)
11:29:26 Join D_the_n [0] (~D_the_n@user/D-the-n:33076)
11:35:30 Quit D_the_n (Ping timeout: 272 seconds)
11:52:20***Saving seen data "./dancer.seen"
12:00
12:00:41 Quit braewoods__ (Read error: Connection reset by peer)
12:00:55 Join braewoods__ [0] (~braewoods@user/braewoods)
12:02:26 Quit IPG (Ping timeout: 248 seconds)
12:06:18hactar|antyou can do that speachy? impressive
12:20:25 Quit npmania (Read error: Connection reset by peer)
12:20:35 Join npmania [0] (~npmania@210.123.73.190)
12:33:35 Join OlsroFR [0] (~OlsroFR@user/OlsroFR)
12:37:37OlsroFRhere is what I see: ata SLEEPNOW 801 ata FLUSH CACHE 801 ata OFF 1051
12:37:38OlsroFR+speachy
12:39:01OlsroFRI started a new battery benchmark with that build. This time, the ipod is fully charged
12:48:17 Quit rogeliodh9101 (Quit: The Lounge - https://thelounge.chat)
12:48:39 Join rogeliodh9101 [0] (~rogeliodh@rogeliodh.dev)
12:51:23 Quit chris_s (Quit: Client closed)
13:00
13:08:01 Join IPG [0] (~InvoxiPla@94.4.117.11)
13:12:39 Join jacobk [0] (~quassel@utdpat242031.utdallas.edu)
13:17:06 Quit jacobk (Ping timeout: 248 seconds)
13:22:51speachyTrzyzet: We already do mention that, just not directly.
13:22:58 Join jacobk [0] (~quassel@utdpat242064.utdallas.edu)
13:23:16speachyhactar|ant: what's impressive? breaking the ZIF connector?
13:27:16speachyOlsroFR: that's what I'd have expected; we flush the cache (don't issue a sleep) and two seconds later, kill power to the storage device.
13:27:50OlsroFR+speachy I am wondering if the fact that I got flush cache: normal is a good thing here
13:27:58OlsroFRI had flush cache extended with the microSD adapter
13:28:01OlsroFRis there any different ?
13:28:07OlsroFRdifference*
13:28:48speachythat's because the card you're using now is <128GiB.
13:29:03OlsroFRit is a 128GB
13:29:43speachyGiB is 1024*1024*1024, "GB" is 1000*1000*1000
13:30:12OlsroFROK
13:30:24OlsroFRso everything logical here
13:30:56OlsroFRright now the battery is at 84%
13:31:10OlsroFRbacklight off, playing music in mp3
13:31:24speachyturns out these SD<->CF adapters do set various capability flags depending on the size of the storage volume
13:32:24speachywhen you had the 1TB in there, you needed to use LBA48 commands, so the adapter flipped on stuf that only matters if LBA48 needs to be used.
13:32:35speachysmarter than I would have expected
13:32:49hactar|antbeing capable of fixing that
13:32:59speachyFLUSH_CACHE_EXT is basically the same as FLUSH_CACHE but for LBA48.
13:33:35speachyhactar|ant: oh, screw that. "Fix" in this context means "swap the mainboard" −− it has no audio output anyway
13:34:00speachyI don't have the eyesight or a steady enough hand to do that sort of delicate surface mount work.
13:34:10hactar|antohhh lol ok
13:34:38speachy(I think the tools I have are enough, if wielded by someone with a steadier hand
13:34:50OlsroFRhttps://gerrit.rockbox.org/r/c/rockbox/+/5986/2/firmware/drivers/ata.c#870 look your code +speachy
13:35:00OlsroFR    /* If device doesn't support PM features, don't try to sleep. */
13:35:00OlsroFR    if (!ata_disk_can_sleep())
13:35:01OlsroFR        return 0; // XXX or return a failure?
13:35:06OlsroFRyou should also check canflush here
13:35:35OlsroFRright now you allow the timer to start, but once it call the function, it return 0 rather than powering down the adapter...
13:35:37speachyno. the command that function issues will lock the adapter up.
13:36:23speachyata_perform_sleep() returns 0 if it succeeds or the device doesn't support sleeping
13:37:32speachybasically "sleeping" is independent of "power off" as the latter doesn't require any cooperation from the storage device. we just kill the power.
13:38:58 Quit jacobk (Ping timeout: 248 seconds)
13:41:50OlsroFRI downloaded your patch to get more insight from all of this
13:43:45OlsroFRof your code is allright, and show log proved it. You do the "ide_power_enable(false);" just under the logging
13:43:55OlsroFROk*, your
13:45:03speachyI want to make more structural changes to the ATA code but one must proceed quite cautiously.
13:46:29OlsroFRyour approach is allright. I tried to do something like this when I investigating this, but couldn't know how to look properly the flush event and did also my work under an bugged hardware environment with that 1TB stuff
13:47:10speachyyeah. I suspect a large portion of the user problem reports are due to flaky hardware
13:47:38speachybut strictly speaking we're issuing commands that the hardware claims to not support
13:48:29OlsroFROh, even ATA flush commands are not supported with sd adapters ?
13:48:35speachyactually I take that back, it does claim to support FLUSH
13:49:16speachysome of the identify dumps I was given early on were corrupted due to PIO timings not working for non-hard disks.
13:50:11speachyflush is optional in ATA-4, mandatory in ATA-5, and a capability flag (that should always be asserted) was added to ATA-6.
13:50:49OlsroFRinteresting. Should be fine so to hook on to check that data are written correctly on disk. I hope I will get more battery now. 8 hours would be ok. I will also do a test on Stock OS with the same music content, I am curious now, to compare. Technically speaking, shutting down the disk should help a lot with battery life on those old iPods
13:50:49speachy...some of these things don't set _any_ of the "ATA version" bits.
13:51:15speachydepends on the SD card. some are a lot hungrier than others.
13:51:50speachywould be informative to probe the power feed into one of these things to get some hard data
13:52:23speachycome to think of it I might still have a SD paddleboard for that very purpose.
13:52:25***Saving seen data "./dancer.seen"
13:52:42OlsroFRtarkan did some compare in the past on Stock OS
13:53:01OlsroFRhttps://www.iflash.xyz/iflash-ipod-sd-vs-cf-power-consumption-analysis/
13:54:03 Join jacobk [0] (~quassel@utdpat242060.utdallas.edu)
13:54:50speachytoo bad they didn't include a stock HDD.
13:55:26OlsroFRhe did also, but here: https://www.iflash.xyz/runtime-shootout-2016-quad-dual-solo-msata-vs-original-hard-drive/
13:56:09speachybut I'm particularly interested in seeing the realtime power draw of a single off->on->reads->write->sleep->off cycle
13:56:29OlsroFRalso https://pastebin.com/0CK0bczT here is the benchmark that I made on Rockbox with the 1TB setup with an older build without any kind of shutdown so the adapter was ON the whole time. It was stable for playing music all along.
13:57:12OlsroFR6h32 before shutting down while playing FLACS music
13:59:29speachymodern mSATA ought to be more power efficient too
14:00
14:00:35speachy(and I suspect the old ipod firmware doesn't try to take advantage of the ATA advanced PM feature set)
14:06:54 Quit jacobk (Ping timeout: 276 seconds)
14:17:16OlsroFRbattery is now at 65%; it's draining fast...
14:23:54 Join lebellium [0] (~lebellium@2a01cb0405d07f00902200cf1da8619d.ipv6.abo.wanadoo.fr)
14:29:41speachyflacs are not great due to the much larger files
14:29:50speachy(for power management I mean)
14:34:27 Quit OlsroFR (Ping timeout: 256 seconds)
14:58:20 Join OlsroFR [0] (~OlsroFR@user/OlsroFR)
14:58:22OlsroFRright now, the same ipod is just playing mp3 files. So theorically, battery life should be far better than 6h30.
14:58:26OlsroFRthe percentage is now 44%. It is lowering so fast... :(
14:58:35 Quit OlsroFR (Client Quit)
14:58:55speachyscreen on?
15:00
15:05:07 Join OlsroFR [0] (~OlsroFR@user/OlsroFR)
15:05:09OlsroFRbacklight off in both cases
15:05:18OlsroFRbut on the mini, screen has to remain ON while music playback
15:18:07OlsroFRbattery: 30% now
15:36:12 Join othello8 [0] (~Thunderbi@pool-100-36-176-164.washdc.fios.verizon.net)
15:37:48 Quit othello7 (Ping timeout: 252 seconds)
15:37:48 Nick othello8 is now known as othello7 (~Thunderbi@pool-100-36-176-164.washdc.fios.verizon.net)
15:52:26***Saving seen data "./dancer.seen"
15:58:16 Quit OlsroFR (Quit: Client closed)
16:00
16:11:22 Join jacobk [0] (~quassel@utdpat241106.utdallas.edu)
16:13:58 Join Everything [0] (~Everythin@178-133-77-41.mobile.vf-ua.net)
16:14:55speachyscreen on does eat into battery life considerably, even without the backlight.
16:22:41edhelas_speachy it's not possible to turn it off on the Mini like on the 6g ?
16:23:13 Part Everything
16:24:08speachyhmm, that seems like an oversight.
16:26:10 Quit lebellium (Quit: Leaving)
16:31:12 Quit jacobk (Ping timeout: 276 seconds)
16:32:24speachyhuh, never really paid attention to that before.
16:43:26 Join OlsroFR [0] (~OlsroFR@user/OlsroFR)
16:43:38OlsroFRbattery is now 6%
16:44:25OlsroFRthe screen itself should not be so much an issue. My blue ipod mini 2G who was also a new battery but a real CF card can last 15 hours easily on Rockbox...
17:00
17:03:51 Quit braewoods__ (Read error: Connection reset by peer)
17:03:53 Join braewoods [0] (~braewoods@user/braewoods)
17:10:23OlsroFRbattery is now 3%, end is near
17:16:58 Join Moriar [0] (~moriar@107-200-193-159.lightspeed.stlsmo.sbcglobal.net)
17:18:12 Quit braewoods (Remote host closed the connection)
17:20:04 Join Everything [0] (~Everythin@178-133-29-214.mobile.vf-ua.net)
17:20:36 Join braewoods [0] (~braewoods@user/braewoods)
17:23:53OlsroFR1% now
17:32:19OlsroFR0% now (but still alive)
17:41:30 Join jacobk [0] (~quassel@47-186-105-237.dlls.tx.frontiernet.net)
17:43:18 Quit braewoods (Read error: Connection reset by peer)
17:43:27 Join braewoods_ [0] (~braewoods@user/braewoods)
17:45:41 Quit braewoods_ (Read error: Connection reset by peer)
17:45:56 Join braewoods_ [0] (~braewoods@user/braewoods)
17:47:56 Quit braewoods_ (Read error: Connection reset by peer)
17:48:02 Join braewoods__ [0] (~braewoods@user/braewoods)
17:52:28***Saving seen data "./dancer.seen"
17:57:45 Quit dconrad (Remote host closed the connection)
18:00
18:00:19 Join dconrad [0] (~dconrad@152.117.104.217)
18:02:19OlsroFRipod is now dead :(
18:04:20OlsroFRresult is 5h15 before the shut down
18:04:56 Quit dconrad (Ping timeout: 255 seconds)
18:07:20edhelas_OlsroFR not sure if its a good idea to completely empty the battery
18:07:37OlsroFRI had to do it to benchmark the patch correctly
18:16:01 Quit OlsroFR (Ping timeout: 256 seconds)
18:18:38 Quit cstine (Quit: The Lounge - https://thelounge.chat)
18:19:08 Join cstine [0] (~cstine@150.136.136.191)
18:33:01 Quit braewoods__ (Read error: Connection reset by peer)
18:33:09 Join braewoods_ [0] (~braewoods@user/braewoods)
18:48:16 Quit Everything (Quit: leaving)
18:52:31 Join massiveH [0] (~massiveH@2600:4040:a982:dc00:ad10:8482:60ec:6ffb)
19:00
19:52:29***Saving seen data "./dancer.seen"
20:00
20:13:49 Join dconrad [0] (~dconrad@152.117.104.217)
20:18:38 Quit dconrad (Ping timeout: 272 seconds)
21:00
21:45:33 Quit Moriar (Quit: Leaving.)
21:52:31***Saving seen data "./dancer.seen"
22:00
22:38:47 Join jeans1826 [0] (~jeans1826@2601:447:c200:5cc0:a59f:8bc2:c79d:c9ff)
22:39:00jeans1826hi - could i ask a question about rockbox here?
22:40:07jeans1826my old sansa mp3 player finally died, and they stopped producing the model - so i need some kind of replacement, and i was thinking of getting an ipod and installing rockbox on it
22:40:42jeans1826my big thing i cant figure out is if its possible to put podcasts/audiobooks onto it, and for it to keep my place on the individual files if i re open the file later?
22:56:09 Join dconrad [0] (~dconrad@152.117.104.217)
23:00
23:00:22 Quit dconrad (Ping timeout: 252 seconds)
23:43:18jeans1826is anybody here?
23:52:34***Saving seen data "./dancer.seen"

Previous day | Next day