EOS M100 porting - page 13 - DryOS Development - CHDK Forum  

EOS M100 porting

  • 135 Replies
  • 20646 Views
*

Offline c_joerg

  • *****
  • 1129
Re: EOS M100 porting
« Reply #120 on: 30 / August / 2021, 13:26:16 »
Advertisements
That build crashes as well.


It's really weird. I couldn't get it right at first either, no matter which version I take. And I have a number of RAWOPINT log files.
It's the interval. At 20 (2s) the script runs with „m100-100a-1.5.0-5575 Oct 12 2020 17:13:37“

Also your latest source (m100-100a-1.5.0-5616 Aug  8 2021 18:15:50) works with 2s
« Last Edit: 30 / August / 2021, 13:33:59 by c_joerg »
M100 100a, M3 101a, G9x II (1.00c), 2*G1x (101a,100e), S110 (103a), SX50 (100c), SX230 (101a), S45,
Flickr https://www.flickr.com/photos/136329431@N06/albums
YouTube https://www.youtube.com/channel/UCrTH0tHy9OYTVDzWIvXEMlw/videos?shelf_id=0&view=0&sort=dd

*

Offline Caefix

  • *****
  • 568
  • Sorry, busy deleting test shots...
Re: EOS M100 porting
« Reply #121 on: 30 / August / 2021, 14:57:33 »
That build crashes as well.



Same with 5616 compiled with 5670...  ???
Btw: Other lenses crashing as well: EF-M 22 STM + 11-22 IS STM.

(Something else needs adjustment too, md fields only fit to 3:2. )
Edit: With some extra lines copied from g7x2´s lib.c to 127-214 the md issue seems done.  :xmas

Edit2: @srsa_4c  Thanks for update to #6014 ::) , lib.c for md_ adapted to it.
« Last Edit: 31 / August / 2021, 15:36:03 by Caefix »
All lifetime is a loan from eternity.

*

Offline Caefix

  • *****
  • 568
  • Sorry, busy deleting test shots...
Re: EOS M100 porting
« Reply #122 on: 31 / August / 2021, 12:29:47 »
It's really weird.

Instead of the DefMarkMan Assert :blink:  I got sometimes (even at 2sec) :
Edit: (Focus-related /low light?)
Edit2: I don´t need cont mode for failure if interval >2.5 sec...
« Last Edit: 31 / August / 2021, 15:50:35 by Caefix »
All lifetime is a loan from eternity.

*

Offline srsa_4c

  • ******
  • 4426
Re: EOS M100 porting
« Reply #123 on: 31 / August / 2021, 14:54:30 »
If you have a reliable test case, it would be interesting to know if g7x II has the same problem.
The conditions I know: auto focus, continuous mode, file format: native raw only. Running hooktest.lua (only cont test enabled) reliably crashes the camera then.
If no focus is found, the script times out at the first 'hook_shoot wait'.
In the crash case, the last script log entry is again 'hook_shoot wait', with a timestamp approx. 4 seconds less than the romlog's last entry.


*

Offline philmoz

  • *****
  • 3332
    • Photos
Re: EOS M100 porting
« Reply #124 on: 31 / August / 2021, 18:43:50 »
The G7XII crashes in hooktest.lua when running the continuous mode tests.

Not sure if this is similar to the M100 crash or not. Hooktest and ROMLOG log files attached.

Currently moving house so may not get to look further at this for a while.


Edit: A quick check of the ROMLOG shows the crash is in TakeSemaphore called from our suba_alloc function  ???
« Last Edit: 01 / September / 2021, 05:19:03 by philmoz »
CHDK ports:
  sx30is (1.00c, 1.00h, 1.00l, 1.00n & 1.00p)
  g12 (1.00c, 1.00e, 1.00f & 1.00g)
  sx130is (1.01d & 1.01f)
  ixus310hs (1.00a & 1.01a)
  sx40hs (1.00d, 1.00g & 1.00i)
  g1x (1.00e, 1.00f & 1.00g)
  g5x (1.00c, 1.01a, 1.01b)
  g7x2 (1.01a, 1.01b, 1.10b)

*

Offline Caefix

  • *****
  • 568
  • Sorry, busy deleting test shots...
Re: EOS M100 porting
« Reply #125 on: 01 / September / 2021, 12:36:34 »
I think, :haha/platform/M100/shooting.c should have the same aperture_sizes_ & shutter_speeds_table as M3.
All lifetime is a loan from eternity.

*

Offline srsa_4c

  • ******
  • 4426
Re: EOS M100 porting
« Reply #126 on: 01 / September / 2021, 17:39:28 »
The G7XII crashes in hooktest.lua when running the continuous mode tests.

Not sure if this is similar to the M100 crash or not. Hooktest and ROMLOG log files attached.
Similarity is that both ports crash while doing the continuous test, and that isn't really supposed to happen.
In your case the reason of crash is obvious RAM corruption that affects a CHDK variable. Before seeing that, I thought that the m100 port likely has a badly placed hook. Now I'm not so sure, we might have a common problem that causes RAM corruption.

edit:
If I remember correctly, the PhySw task is running on core1 (unlike most of the other tasks), and has a higher-than-usual priority. In case PhySw ever calls CreateTask (does it?), the created task will also run on the same core. The shooting hooks rely on simple variables rather than, say, semaphores. I wonder if the current logic has problems in a dual core environment.

edit2:
Moving physw to core0 has absolutely no effect on the issue.

edit3:
Moving wait_until_remote_button_is_released almost below capt_seq_hook_set_nr (after the sub_E004C5EE call) eliminates the crash (the original place is inside a semaphore-protected part of the fw code). Even though the script doesn't crash the camera, it still gets stuck at the third 'hook_shoot wait'...
« Last Edit: 03 / September / 2021, 17:28:20 by srsa_4c »

*

Offline Caefix

  • *****
  • 568
  • Sorry, busy deleting test shots...
Re: EOS M100 porting
« Reply #127 on: 08 / September / 2021, 16:14:57 »
 :blink: The critical rewopint setting crashes, hooktest always fine in cont...

Late edit: Latest flciker appended ...  :o
« Last Edit: 13 / September / 2021, 15:43:17 by Caefix »
All lifetime is a loan from eternity.


*

Offline Caefix

  • *****
  • 568
  • Sorry, busy deleting test shots...
Re: EOS M100 porting
« Reply #128 on: 11 / September / 2021, 12:50:50 »
 :blink: There is no flicker in GUI&OSD until the 1.st photo is taken, scripted or not.
Edit: ... with manual lens (Rockstar Fisheye 10 F8 )  :haha
Edit2: With other lenses it´s enough to reach a [red out of focus square]! to trigger flicker...
Edit3: Further hooktest can pass all several times with smooth display.
Edit4: Found no way (AiAf+servo+cont) to crash hooktest.
« Last Edit: 20 / September / 2021, 14:50:43 by Caefix »
All lifetime is a loan from eternity.

*

Offline Caefix

  • *****
  • 568
  • Sorry, busy deleting test shots...
Re: EOS M100 porting
« Reply #129 on: 14 / September / 2021, 12:18:56 »
Unexpected fun detected...  :D
md_tune.bas tells min 10, max 150, avg ~70.
md_detect works fine in <play> && with the touchscreen.
So enlargening and moving a picture gives nice trails in a md grid. (Nice is white dot in nigthly sky)
I can´t really compare it with the M3-X build, but it seems much faster here on M100 classic..

Since my cam´s CRC has "changed", I have turned the check off in debug menu & turned it on again for the updated build.
Now sometimes it happens, that primary.bin is written, but mostly it´s skipped after (always) shortly displayed calc- & warning window, or it seems! stuck because it waits for the button while the ww is already deleted.
Booting to <rec> or <play> doesn´t matter.
All lifetime is a loan from eternity.

 

Related Topics