IRC log of #maemo for Monday, 2017-11-27

*** pkill9 has quit IRC00:15
*** Pali has quit IRC01:00
*** atk has quit IRC02:00
*** atk has joined #maemo02:00
*** xorly has quit IRC02:28
*** Natch has quit IRC02:44
*** florian_kc has quit IRC02:46
*** Natch has joined #maemo02:47
*** pkill9 has joined #maemo03:04
*** pkill9 has quit IRC03:54
*** dafox has quit IRC04:10
*** Kabouik_ has quit IRC04:32
*** Maxdaman1us is now known as Maxdamantus07:17
*** FalconSpy has quit IRC07:43
*** FalconSpy has joined #maemo08:10
*** FalconSpy has joined #maemo08:10
*** spiiroin has quit IRC08:36
*** Oksanaa has quit IRC09:06
*** spiiroin has joined #maemo09:09
*** geaaru has quit IRC09:09
*** Marshall_Banana has joined #maemo09:13
*** herekun has quit IRC09:17
*** herekun has joined #maemo09:20
drathirDocScrutinizer05: if i may ask n900 what type of simlock have?09:31
*** xes__ is now known as xes09:40
*** Kabouik has joined #maemo09:44
bencohit usually comes sim-unlocked09:55
KotCzarnyand since it has bb5 modem, all rules apply09:56
*** Marshall_Banana has quit IRC10:02
*** florian has joined #maemo10:04
*** heroux has quit IRC10:22
*** heroux has joined #maemo10:38
*** Smily has joined #maemo10:41
*** Kabouik has quit IRC10:47
*** ketas has quit IRC11:00
*** ketas has joined #maemo11:04
*** jskarvad has joined #maemo11:08
*** jskarvad has quit IRC11:08
*** jskarvad has joined #maemo11:08
*** qwazix has quit IRC11:20
*** qwazix has joined #maemo11:21
*** Kilroo has quit IRC11:25
*** eijk_ has joined #maemo11:39
*** Kilroo has joined #maemo11:40
*** cyteen has quit IRC11:44
*** cyteen has joined #maemo11:48
*** cyteen has quit IRC12:01
*** cyteen has joined #maemo12:14
*** cyteen has quit IRC12:19
*** eMHa has joined #maemo12:28
*** pigeon has quit IRC12:34
*** pigeon has joined #maemo12:36
*** cyteen has joined #maemo12:44
*** opt1mus has quit IRC12:45
*** xorly has joined #maemo12:53
*** eMHa has quit IRC13:00
*** Wikiwide has joined #maemo13:21
WikiwideWhy does my N900 have camera's videos mixed with all other films in Media Player, and not placed into category of "Recorded by device camera"?13:22
*** xorly has quit IRC13:37
*** eMHa has joined #maemo13:39
*** freemangordon has quit IRC13:57
*** Wikiwide has left #maemo14:00
*** pigeon has quit IRC14:49
*** pigeon has joined #maemo14:52
*** eijk has joined #maemo14:52
*** eijk_ has quit IRC14:53
*** pkill9 has joined #maemo15:03
*** dafox has joined #maemo15:19
DocScrutinizer05I think because it doesn't filter subgroups out form "all" group15:31
DocScrutinizer05beauty of fracker err tracker15:31
DocScrutinizer05the tracker concept is broken by design, in my book15:32
*** KotCzarny has quit IRC15:38
*** geaaru has joined #maemo15:46
*** mkz has joined #maemo15:53
mkzis it ok if I mirror ?15:54
mkzthat is, recursive wget or rsync15:54
mkzdon't want to be that guy which hogs bandwidth without permission15:55
bencohif anything, you should probably use apt-mirror15:56
bencoh(or something close to it)15:57
bencohfirst sync will take ages, but then on it should be fine15:58
DocScrutinizer05generally it's ok to mirror repos and whatever you like16:01
DocScrutinizer05but listen to advice regarding best way to do so16:01
*** spiiroin has quit IRC16:02
*** ketas- has joined #maemo16:13
*** spiiroin has joined #maemo16:16
Wizzupnew cssu in cssu testing it seems, cool. can't upgrade to it due some conflict with curl though (on cssu thumb)16:18
enycWizzup: hrrm somebody else whad some crazy dependency mess too16:54
enyc15:23  * DocScrutinizer05 feels massive headache - prolly getting too old for such shit16:54
enycok maye that wasn't the rpoblem16:55
*** pkill9 has quit IRC17:01
Wizzupenyc: I think there needs to be a new curl build for thumb17:04
Wizzupso it might just take a bit more time17:04
*** pkill9 has joined #maemo17:14
*** xorly has joined #maemo17:22
*** vectis3 has quit IRC17:34
*** vectis3 has joined #maemo17:34
*** auenf has quit IRC17:42
*** auenf has joined #maemo17:45
*** Kabouik has joined #maemo17:48
*** KotCzarny has joined #maemo17:53
*** Vajb has quit IRC18:33
*** Vajb has joined #maemo18:35
*** Pali has joined #maemo18:36
*** herekun has quit IRC18:59
*** herekun has joined #maemo19:04
*** ketar has joined #maemo19:08
*** ketar has quit IRC19:10
*** ketar has joined #maemo19:10
*** ketar has quit IRC19:11
*** herekun has quit IRC19:12
*** pagurus has quit IRC19:14
*** pagurus has joined #maemo19:15
*** herekun has joined #maemo19:32
*** pkill9 has quit IRC19:37
*** pkill9 has joined #maemo19:54
*** eMHa has quit IRC20:25
*** florian has quit IRC20:42
*** Natch has quit IRC21:01
*** pigeon has quit IRC21:11
*** pigeon has joined #maemo21:13
*** pkill9 has quit IRC21:20
*** freemangordon has joined #maemo21:22
*** eMHa has joined #maemo21:28
*** Natch has joined #maemo21:43
*** jonwil has joined #maemo21:48
*** jskarvad has quit IRC21:53
*** florian has joined #maemo22:05
*** Kabouik_ has joined #maemo22:17
*** Kabouik has quit IRC22:20
*** jonwil has quit IRC22:28
*** eijk has quit IRC22:43
*** Oksana has joined #maemo22:49
CatButtshey DocScrutinizer0523:02
CatButtsyou mentioned softfloat is not 'that' much slower than hardfloat23:03
CatButtshow would it hold against something like DSP23:03
CatButtssay, a multichannel softsynth23:04
DocScrutinizer05CatButts: softFP uses whatever available, DSP or arithmetic coprocessor or even software implementation of float math23:04
DocScrutinizer05softfp is hardfp wrapped into a library call ABI23:05
DocScrutinizer05if the platform has a hw accel for FP23:05
DocScrutinizer05so your question doesn't make much sense to me23:06
CatButtsso it still ends up in hardware, but there is a capability check at runtime?23:06
DocScrutinizer05not at runtime, at compile time23:06
DocScrutinizer05at least for maemo, that's not done at runtime afaik23:07
CatButtsso on the n900, I still benefit from hardware floats?23:08
DocScrutinizer05softfp is a library, the implementation of that library may differ with target platform, the API stays the same23:08
CatButtsah, I see meow23:08
DocScrutinizer05hardfp directly calls the float hw from your code, without any library in between23:09
DocScrutinizer05(simplified picture)23:09
CatButtsI initially thought that softfp meant software emulated float calculations23:09
DocScrutinizer05the overhead of softfp over hardfp is basically the library wrapper code, for maemo23:10
CatButtsnot a lib call23:10
CatButtsnow it's clear in my head23:10
DocScrutinizer05or let me put it that way: your maemo osftfp library is built using hardfp code23:11
freemangordonguys, it is not like that23:12
freemangordonit is about the calling convention23:13
freemangordonsoftfp means float args are passed through int registers23:13
freemangordonhardfp means float args are passed through FP registers23:13
DocScrutinizer05yes, you call a copoc directly with parameters in coproc registers, for hardfp. You call a library with parameters on stack or whatever, for softfp23:13
freemangordonno, you call a library (or a local function) both ways23:14
freemangordonjust that function parameters are put in CPU for softfp or FPU for hardfp registers23:15
freemangordonwith softfp you can emulate FP in software23:15
freemangordonbut you can use HW FPU if it is available as well23:16
DocScrutinizer05"soft" vs "softfp" vs "hard"23:16
freemangordonDocScrutinizer05: remember how was 8087 emulated?23:16
freemangordonback then23:16
DocScrutinizer05wasn't that softirq23:17
DocScrutinizer05software interrupt23:17
freemangordonno, it was generatin "illegal instruction" interrupt, whatever it was called :)23:18
freemangordonthat irq was handled by FPU emulation lib23:18
DocScrutinizer05hmm, pretty much a sowtware interrupt23:18
DocScrutinizer05an instruction causing an interupt23:19
freemangordonhowever, soft(fp) vs hardfp in terms of cortex a8 is just about where FP parameters are passed23:20
DocScrutinizer05so not as huge a difference as the uneduicated user might expect23:20
freemangordonwell, it was a prefix at a start of every FPU instruction, 0x0F or somesuch23:20
freemangordonthere is a huge diff for FPU intense function calls23:21
freemangordonsome 20% iirc23:21
freemangordonbut no difference for most of the system23:21
DocScrutinizer05yeah, probably you need 30 instead 8 instructions for a complete FP op23:21
DocScrutinizer05no 1000 instructions though23:22
DocScrutinizer05like you'd need for software emulation23:22
freemangordonit is not the instructions themselves, but the memory access23:22
freemangordonI am not sure you can copy from integer reg to FPU reg directly, without going through memory, however, even if it is possible, it takes time23:24
freemangordonand leads to pipeline stalls23:24
DocScrutinizer05I'd give ballpark estimations of  "soft":1000;  "softfp":1000 or 30 (depending if hw accel available);"hard" (only when available): 823:25
freemangordonCatButts: the last ~20 lines is a kind of an answer to your question23:25
DocScrutinizer05whatever time unit23:25
DocScrutinizer05according to you it's rather  "soft":1000;  "softfp":1000 or 12 (depending if hw accel available);"hard" (only when available): 1023:26
freemangordonDocScrutinizer05: well, keep in mind that for bug FP functions penalty is relatively small comapred to function execution time23:26
CatButtsso it's a tad bit slowe, but no tragedy23:26
CatButtseven with intensive float calculations?23:27
freemangordonfloat calculations themselves are not affected23:27
DocScrutinizer05that's *only* looking at the float instructions23:27
freemangordonjust the function calls23:27
DocScrutinizer05the rest of the system doesn't care what your float does23:28
DocScrutinizer05a program that doesn't use floats will not run any fatser on a hardfp system,23:28
DocScrutinizer05CatButts: so yes, *only* "with intensive float calculations" (100% intensity)23:29
* DocScrutinizer05 idly wonders if you could wrap the softfp functions with an IRQ-lock and save all FP registers and restore them before leavinbg the softfp function, so hardfp and softfp code could coexist23:32
DocScrutinizer05sucks big time, but would make coexistence possible23:33
* CatButts plays with cubic interpolation and manages to turn it into ear-raping noise23:33
DocScrutinizer05wait, wouldn't an IRQ either not touch FP registers at all, or (in case it's the schedukler) save the FP registers as well and restore them for the newly scheduled process?23:34
DocScrutinizer05so except for shared libs, is there *anything* that comflicts between softfp-processes and hard(fp)-processes?23:35
DocScrutinizer05and even for shared libs, would the ABI change for soft/hard?23:37
*** thuttu77 has quit IRC23:37
DocScrutinizer05freemangordon: ^^^23:37
*** eijk has joined #maemo23:38
freemangordonthere is nothing that conflicts23:38
DocScrutinizer05HAHAH  >>FPU instructions could be emulated by the kernel so that FPU-less systems could run with this ABI but as far as we know this does not exist<<23:38
freemangordonnot exactly true iirc, there are "some" instructions emulated in the kernel23:39
DocScrutinizer05ooh, this direction of "not true" :-D23:39
DocScrutinizer05oooh, "pipeline stalls"23:40
freemangordoniirc, some vfpv3 instructions are emulated if hw is vfpv223:40
freemangordonbut don;t quote me on that one, it was a while I looked at kernel FP support code23:41
*** pkill9 has joined #maemo23:41
DocScrutinizer05CatButts: (at runtime) I stand corrected: >>The GCC -mfloat-abi=softfp flag allows use of VFP while remaining compatible with soft-float code. This allows selection of appropriate routines at runtime based on the availability of VFP hardware.<<23:50
DocScrutinizer05>>The runtime linker,, supports a mechanism for selecting runtime libraries based on features reported by the kernel. For instance, it's possible to provide two versions of libm, one in /lib and another one in /lib/vfp, and will select the /lib/vfp one on systems with VFP.<<23:51
DocScrutinizer05BUT... >>...the ARM EABI and ELF specification has no way to tell which of soft, softfp, hard is used to build it<<23:53
DocScrutinizer05HMMPH   >>This needs to be looked into.. is it only a part of the version 5 EABI and not implemented? Could there possibly be scope for a GNU EABI extension for this, or a fix for the linker?<<23:56

Generated by 2.15.1 by Marius Gedminas - find it at!