| *** lbt_ has joined #sailfishos-porters | 01:55 | |
| *** lbt_ has joined #sailfishos-porters | 01:55 | |
| *** lbt has quit IRC | 01:55 | |
| *** phdeswer has quit IRC | 06:39 | |
| *** phdeswer has joined #sailfishos-porters | 07:54 | |
| situ | vgrade_: Looks like we are seeing lipstick crashes because of uniniitalizes variables | 12:54 |
|---|---|---|
| situ | https://github.com/siteshwar/android_frameworks_native/blob/cm-10.2/services/surfaceflinger/DisplayHardware/HWComposer.cpp#L732 | 12:54 |
| situ | compare above to https://github.com/siteshwar/qt5-qpa-hwcomposer-plugin/blob/hammerhead/hwcomposer/hwcomposer_backend_v11.cpp#L83 | 12:55 |
| situ | thp: Is that correct ? | 12:56 |
| *** plundstr has left #sailfishos-porters | 13:10 | |
| morphis | situ: what exactly do you mean in the code? | 13:25 |
| situ | morphis: From surfaceflinger code look at line 733 to 761 and 794 to 816 | 13:28 |
| situ | We are not doing that in qpa plugin. | 13:28 |
| morphis | ah | 13:29 |
| morphis | we're doing it but differently | 13:29 |
| morphis | surfaceflinger builds its display list from time to time | 13:29 |
| morphis | we build it once and set variables to the right values when needed | 13:29 |
| morphis | situ: http://bpaste.net/show/mKEVEs4QLJ1m5zJ96g3j/ is what I currently need to get it working on a HP touchpad with hwcomposer 1.3 | 13:30 |
| morphis | with two additional changes on top of hardware/qcom/display-caf: https://github.com/webOS-ports/android_hardware_qcom_display-caf/commits/wop-11.0 | 13:31 |
| situ | hmm ok | 13:32 |
| *** phdeswer has quit IRC | 13:33 | |
| *** phdeswer has joined #sailfishos-porters | 14:11 | |
| *** vgrade has joined #sailfishos-porters | 14:45 | |
| vgrade | morphis: this one https://github.com/webOS-ports/android_hardware_qcom_display-caf/commit/c02650576fdeed8af1f74d4dc49f07b8a8c4f577 | 14:45 |
| vgrade | is interesting | 14:45 |
| vgrade | this is directly related to the issue we saw and tried to work around with https://github.com/siteshwar/qt5-qpa-hwcomposer-plugin/commit/414d865a622a26de9d1d1c3378bd8dea03cabc9f | 14:48 |
| vgrade | ie reducing the number of times round the loop | 14:48 |
| vgrade | situ: ^ | 14:49 |
| situ | vgrade: But I saw another very stragen thing, let me show you | 14:51 |
| situ | So in qpa plugin num_displays variable is used only at 2 places | 14:53 |
| situ | https://github.com/siteshwar/qt5-qpa-hwcomposer-plugin/blob/master/hwcomposer/hwcomposer_backend_v11.cpp | 14:53 |
| situ | line 83 and 86 | 14:54 |
| vgrade | looks like more than 2 places | 14:56 |
| situ | https://github.com/siteshwar/android_hardware_libhardware/blob/cm-10.2/modules/hwcomposer/hwcomposer.cpp#L70 | 14:56 |
| situ | https://github.com/siteshwar/android_hardware_libhardware/blob/cm-10.2/modules/hwcomposer/hwcomposer.cpp#L81 | 14:56 |
| situ | none of the 2 methods actually use that variable | 14:57 |
| vgrade | situ: those look like cm10 | 14:59 |
| situ | Sorry, Same case with hybris-11.0 https://github.com/siteshwar/android_hardware_libhardware/blob/hybris-11.0/modules/hwcomposer/hwcomposer.cpp#L70 | 14:59 |
| situ | But those are static methods hmm | 15:00 |
| vgrade | are those in a layer above the specific hardware implementation | 15:01 |
| * vgrade needs to go back to day job. | 15:02 | |
| *** vgrade has quit IRC | 15:02 | |
| situ | morphis: on your device, what happens if you do 'cat /dev/fb2' ? | 15:18 |
| morphis | situ: the one you're looking at is not the qcom hwcomposer implementation | 15:29 |
| morphis | look at https://github.com/CyanogenMod/android_hardware_qcom_display/tree/cm-11.0 for cm-11.0 && nexus5 | 15:30 |
| w00tc0d3 | hey | 16:50 |
| *** w00tc0d3 has quit IRC | 16:50 | |
| *** w00tc0d3 has joined #sailfishos-porters | 16:50 | |
| *** w00tc0d3 is now known as netchip | 16:50 | |
| netchip | today, school has ended for real :D | 16:51 |
| netchip | And, MSameer, I discovered new parts in the C programming language book xD | 16:51 |
| situ | morphis: but in what case this will happen : | 17:20 |
| situ | open("/dev/graphics/fb2", O_RDWR|O_LARGEFILE) = -1 ENOTSUPP (Unknown error 524) | 17:20 |
| situ | In my kernel config | 17:24 |
| situ | # CONFIG_FB_VIRTUAL is not set | 17:24 |
| situ | Shoud I try setting it ? | 17:24 |
| MSameer | netchip: what have you discovered :) | 17:33 |
| MSameer | netchip: and congrats for getting rid of school | 17:33 |
| MSameer | situ: it helps to check the code and see what returns the ENOTSUPP | 17:34 |
| situ | MSameer: I don't see anything returning ENOTSUPP in code. | 17:35 |
| MSameer | odd | 17:35 |
| MSameer | link to a git repos I can clone? | 17:35 |
| situ | What if kernel support is not there for a feature ? | 17:35 |
| MSameer | who makes that device node then? | 17:36 |
| situ | MSameer: How would I know which kernel driver is being used ? | 17:36 |
| MSameer | situ: ls -lh /dev/graphics/fb2 | 17:36 |
| situ | Yes, but which kernel driver is it using ? | 17:37 |
| MSameer | should be possible via udev/sysfs | 17:37 |
| MSameer | /sys/class/graphics/fb2 | 17:38 |
| MSameer | on jolla phone: | 17:38 |
| MSameer | cat /sys/class/graphics/fb2/name | 17:38 |
| MSameer | msmfb43_a0001 | 17:38 |
| MSameer | so it's msmfb "something" | 17:38 |
| situ | mdssfb_a0000 | 17:40 |
| situ | How would I track it kernel source ? | 17:40 |
| MSameer | do you have a link to the git repos? | 17:40 |
| MSameer | I don't know but I can try to help if I have the code | 17:40 |
| situ | MSameer: This is my manifest https://github.com/siteshwar/android/blob/hybris-11.0/default.xml | 17:41 |
| MSameer | cat kernel/lge/hammerhead/.git/config | 17:43 |
| situ | MSameer: It's actually ./out/target/product/hammerhead/obj/KERNEL_OBJ/.config | 17:44 |
| MSameer | that's the kernel .config | 17:44 |
| MSameer | I am looking for the git config to know the clone url | 17:44 |
| MSameer | I don't have enough space to download the whole manifest :/ | 17:45 |
| situ | MSameer: ohh sorry, misread your command | 17:45 |
| situ | MSameer: http://pastie.org/9355400 | 17:46 |
| MSameer | I am sure there is a smarter way but I don't know it :) | 17:46 |
| situ | I am building from hybris-11.0 branch | 17:46 |
| MSameer | git@github.com:siteshwar/android_kernel_lge_hammerhead.git | 17:46 |
| MSameer | cloning | 17:46 |
| MSameer | thanks | 17:46 |
| MSameer | here I did mknod fb2 c 29 2 | 17:50 |
| MSameer | tried strace cat fb2 and got: | 17:50 |
| netchip | MSameer: I discovered the reference part. :) It's like a summary, but it's more goes more in details | 17:50 |
| MSameer | open("fb2", O_RDONLY|O_LARGEFILE) = -1 ENODEV (No such device) | 17:50 |
| situ | MSameer: which device ? | 17:50 |
| MSameer | netchip: that's good :) | 17:50 |
| MSameer | situ: just tried to see what happens if I create a device node which does not correspond to a real device driver | 17:51 |
| situ | ok | 17:51 |
| situ | I am pretty sure it could be the kernel optioni which I showed earlier. | 17:51 |
| MSameer | so it's a different error AFAICT | 17:51 |
| MSameer | situ: why don't you try then? | 17:51 |
| situ | CONFIG_FB_VIRTUAL | 17:51 |
| situ | I was waitning for someone to confirm :P | 17:52 |
| situ | kernel building takes lot of time. | 17:52 |
| MSameer | drivers/video/msm$ rgrep ENOTSUPP | wc -l | 17:53 |
| MSameer | 22 | 17:53 |
| MSameer | heh | 17:53 |
| MSameer | situ: any dmesg output when the open() fails? | 17:56 |
| situ | MSameer: Nothing related to video driver | 17:58 |
| MSameer | do you happen to have the full dmesg output BTW? | 17:58 |
| MSameer | I am almost sure it's some file under drivers/video/msm | 17:58 |
| MSameer | but I don't yet know which one | 17:58 |
| situ | let me check | 17:59 |
| situ | MSameer: Now I am getting | 18:01 |
| situ | sorry wait | 18:02 |
| situ | sh-3.2# cat /dev/fb2 | 18:02 |
| situ | cat: /dev/fb2: No such device | 18:02 |
| situ | I had set display count to 1 with this commit https://github.com/siteshwar/qt5-qpa-hwcomposer-plugin/commit/acb78f7bd113a928a9b3dba20138c3fa2d2974c0 | 18:03 |
| situ | Reverting it now | 18:03 |
| MSameer | situ: ls -lh /dev/graphics/fb2 | 18:03 |
| situ | MSameer: http://pastie.org/9355440 | 18:04 |
| MSameer | you are missing the device node | 18:04 |
| MSameer | try: cd /dev && mknod fb2 29 2 | 18:05 |
| situ | It's there | 18:05 |
| situ | sh-3.2# ls -l /dev/fb2 | 18:06 |
| situ | crw-rw---- 1 system graphics 29, 2 Jul 12 20:55 /dev/fb2 | 18:06 |
| MSameer | then something is missing from the kernel | 18:06 |
| MSameer | I guess I can't help here :( | 18:07 |
| situ | MSameer: ok, will try with CONFIG_FB_VIRTUAL option | 18:07 |
| MSameer | sorry I could not help :/ | 18:07 |
| situ | No problems, will catch you guys later. | 18:07 |
| netchip | totally offtopic and irrelevant for me, but I'll still ask the question: is it worth to get a PhD in Eruope? | 18:59 |
| * MSameer does not have a PhD | 19:00 | |
| * MSameer does not even have a masters degree :( | 19:00 | |
| * MSameer did not even study computer science or engineering :'( | 19:01 | |
| netchip | I did some research on this, and I found out it's veru useful for women, but not so much for men. | 19:01 |
| netchip | well, you could always do that, MSameer | 19:01 |
| MSameer | netchip: any tips? but then: does it help much ? :D | 19:02 |
| netchip | something like maths or computer science? | 19:02 |
| netchip | any tips? | 19:02 |
| netchip | what do you mean? | 19:02 |
| MSameer | you said that I could always do it | 19:02 |
| netchip | yes | 19:03 |
| MSameer | if it's about books and practical experience then I think I might have that :) | 19:03 |
| netchip | I don't know if it's helpful, but my intuition says it is | 19:03 |
| MSameer | Academia is a waste of lifetime IMHO | 19:03 |
| netchip | maybe, but the degree you get is not | 19:04 |
| netchip | and well, in several branches you need to do a study. Learning calculus/linear algebra/combinatorics without college might not be easy | 19:04 |
| MSameer | teachers are persons from planet X. using vim is easier than dealing with them :P | 19:05 |
| MSameer | netchip: my POV is: If I discover that I miss something then I try to catch up | 19:05 |
| netchip | that's a good POV :) | 19:05 |
| MSameer | but that is me and whatever has lead me to this has happened. If you are in the beginning then think carefully and decide | 19:06 |
| MSameer | I don't know about Europe to be honest so I cannot tell | 19:06 |
| netchip | getting a master degree is something I really want to do, after that I have no plans :P | 19:06 |
| MSameer | then go for it by all means :) | 19:07 |
| vgrade_ | situ: https://github.com/siteshwar/android_kernel_lge_hammerhead/blob/hybris-11.0/drivers/video/msm/msm_fb.c#L3265 | 19:08 |
| MSameer | netchip: but also make sure you concentrate on the practical aspects | 19:08 |
| vgrade_ | seems to depend on https://github.com/siteshwar/android_kernel_lge_hammerhead/blob/hybris-11.0/drivers/video/msm/msm_fb.c#L3157 | 19:08 |
| vgrade_ | which we don't have in defconfig | 19:08 |
| vgrade_ | mako has it however | 19:09 |
| MSameer | vgrade_: but that's an ioctl (seems so) | 19:09 |
| netchip | MSameer: I'm always unsure which ECs are useful, and which aren't | 19:09 |
| MSameer | open() does not involve any ioctl calls unless the driver does something | 19:09 |
| MSameer | netchip: in what sense? | 19:10 |
| vgrade_ | MSameer: was it on open, was just checking for ENOTSUPP | 19:10 |
| MSameer | vgrade_: [20:20] <situ> open("/dev/graphics/fb2", O_RDWR|O_LARGEFILE) = -1 ENOTSUPP (Unknown error 524) | 19:10 |
| netchip | MSameer: Once in a month, someone from school (usually a teacher) offers me to do a project on a certain subject | 19:11 |
| netchip | and I always wonder if it's useful | 19:11 |
| MSameer | netchip: I'd say if it adds to your knowledge or lets you exercise some skills then it is useful | 19:12 |
| netchip | then I didn't waste 3 months on a project, yaay :P | 19:12 |
| netchip | (I gave a presentation about the inner workings of a solar cell) | 19:12 |
| MSameer | netchip: but sometimes ECs don't seem to have any practical benefit but that's the mistake of the teacher not you | 19:12 |
| netchip | (I have had a lot of positive feedback :)) | 19:12 |
| netchip | that's true, MSameer | 19:13 |
| MSameer | netchip: and you are not yet in college so no one knows what would interest you the most. I'd say that diversity in topics is always needed at this stage | 19:13 |
| MSameer | I guess it helps you to discover what your passion is (and even if you already know what it is, it shows you other stuff) | 19:14 |
| netchip | Unfortunately, my passion is not one subject | 19:15 |
| netchip | it's not that I like one thing | 19:15 |
| MSameer | then maybe it is something common among a handful of subjects | 19:15 |
| netchip | programming, physics, maths - it's all interesting | 19:15 |
| netchip | it's abstract thinking | 19:15 |
| netchip | but that's very broad | 19:15 |
| netchip | or wide (don't know the right word) | 19:16 |
| MSameer | humans don't have passion towards one thing only ;) | 19:16 |
| MSameer | netchip: too abstract :D | 19:16 |
| netchip | Hm? | 19:16 |
| MSameer | very broad or wide or abstract :D | 19:17 |
| MSameer | bbl | 19:17 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!