Canon 1100D / T3

[dfort]

That doesn't sound good. Are you sure one message says "t3_106.sym" and the other one says "t3_1006.sym"

I know it is a PITA but also check the 1.0.5 test build because I put in a pull request to get some modules working on that version.

I'm not seeing these problems in QEMU or maybe I just don't know what to look for.

[DavidSamish]

So sorry, only t3_106 is listed - my bad typo

[dfort]

Even that is weird because if you look at the file structure on the download:



So how is this happening?

I am getting an error message: 'Error loading "ML/MODULES/t3_106.sym" File does not exist."

Are you deleting everything off the card before transferring all the files from the download or just copying over the autoexec.bin file?

[DavidSamish]

That's weird. I started my T3 today, and there is no error message. Maybe it "needed" a couple of restarts. ML seems to be working.

[dfort]

Ok--so how is the camera handling? Is it working as expected?

Turn on the selftest module and run the Debug -> Self tests -> Stubs API test -- problems? Go back to 1.0.5 and try it again. What we're looking for is if the issues we find on 1.0.6 are new or inherited from 1.0.5.

Once we get through this part I'll make a lua_fix build for 1.0.6. In the meantime, you could run the lua_fix build for 1.0.5 that is on the experiments download page.

[EDIT] There have been some download activity on the test builds but no reports yet. I posted a 1.0.6 lua_fix build hoping that someone takes the initiative to do some testing and report back.

For every single camera model available on the Experiments page (lua_fix build), please run:

- api_test.lua (upload the log)
- selftest.mo -> stubs tests (upload the log)
- bench.mo -> memory benchmarks (upload the screenshot)
- overall sanity check (for example, if you decide to take this build out and use it for a couple of hours, please report back)

[dfort]

This is what the memory benchmark tests looks like in QEMU:



[EDIT] If anyone wants to try out 1100D.106 in QEMU, here's the new script:

1100D/debugmsg.gdb

Code Select Expand

# ./run_canon_fw.sh 1100D -d debugmsg
# ./run_canon_fw.sh 1100D -d debugmsg -s -S & arm-none-eabi-gdb -x 1100D/debugmsg.gdb

source -v debug-logging.gdb

# To get debugging symbols from Magic Lantern, uncomment one of these:
#symbol-file ../magic-lantern/platform/1100D.106/magiclantern
#symbol-file ../magic-lantern/platform/1100D.106/autoexec
#symbol-file ../magic-lantern/platform/1100D.106/stubs.o

macro define CURRENT_TASK 0x1a2c
macro define CURRENT_ISR  (MEM(0x670) ? MEM(0x674) >> 2 : 0)

# GDB hook is very slow; -d debugmsg is much faster
# ./run_canon_fw.sh will use this address, don't delete it
# b *0xff06c91c
# DebugMsg_log

b *0xff06fafc
task_create_log

b *0xff1e8638
register_interrupt_log

b *0xff1d34d4
SetEDmac_log

continue


[EDIT 2] Found out that many of the lua tests work in QEMU so I tried running them and it did fine up to the middle of the exposure tests then showed some sort of a srm_malloc_suite error.



The same error is showing up on 1.0.5 and 1.0.6. However, running just the exposure test completed without any errors.

More information on Reply #305 of the How to run Magic Lantern into QEMU?!... topic.

[Danne]

Some testing with the 1100D. Provided autoexe.bin and other info logs from three test builds:
https://bitbucket.org/Dannephoto/magic-lantern/downloads/tests_1100D_new%20firmware.zip

Started with this build since it was already on my cam:
testbuild_105_01_stubtest(zipped folder)
Changeset: 23524a6ed2f9+ (allocate-raw-lv-buffer) tip

Seems more stable than with lua test builds(check autoexe.bin for build info)
Both firmware versions seems to behave the same in my tests. Cannot do any allround tests because Q button stopped working in both lua versions. Check zipped folder:
testbuild_105_02_stubtest_luatest
testbuild_106_03_stubtest_luatest

[dfort]

Thanks for testing. I suspected that 1.0.6 would be about the same as 1.0.5 and that 1.0.5 has some issues.

The selftest module test turned up something I haven't seen before:

Code Select Expand

[FAIL] bmp = bmp_load("ML/CROPMKS/CINESCO2.BMP", 1) => 0x0


That is the only failed test so we aren't in too bad a shape there but the lua tests stopped on the GUI function test on both 1.0.5 and 1.0.6. Let's see what other tests it can pass or fail. Check out this post on the 60D topic for instruction on how to run individual lua tests.

As expected, the benchmark looks very different on the camera than in lua--see Reply #1505 for the same test in QEMU.

[Farangi]

Hey

I'm new here so apologies if I'm posting on the wrong thread but i have a question specific to canon 1100d.

On page 52 of this forum someone suggested that it might be possible to rewire your camera to be able to take a mic...

I Just bought rodé videomic go thinking it could connect to my canon 1100d (it does have  something that looks a lot like a 3.5 input) but when cable wouldn't go in input i looked it up and found that that model  doesnt take external mic... Bit of a bugger as i'm travelling and it was a mission  to get the microphone in the first place...

Has anyone managed to hack their camera's hardware to allow it to take a mic?

Cheers

D

[FastKatt]

Hello, new to photography and camera firmware hacking. But, I've been rooting Android since 2012, and I tuned my 2004 Grand Prix's PCM.

I've found the nightly build for my Rebel T3 (1100D)
I understand the nightly builds should be flashed over a stable build, but I can't find any other builds for my rig?

Could someone point me in the right direction?
Thanks

Sent from my Moto Z (2) using Tapatalk

[Walter Schulz]

I understand the nightly builds should be flashed over a stable build, but I can't find any other builds for my rig?

Not correct. Consider all instructions not found/linked on this site outdated and invalid.
Follow instruction found in -> Top of page -> Downloads -> Download Nightly Builds -> 1100D / Rebel T3 1.0.5 -> Installation

BTW: If you are new to photography consider learning how to use your cam without ML. Learning about exposure time, aperture and ISO is challenging enough (been there, done that). YMMV esp. if you are addicted to video recording.

[FastKatt]

Those are the instructions I followed, before I signed up on the forums. The only thing I did not do is copy my ROM files back to my computer.
I was hoping to have recording bars at the top of the screen, and like the idea of having a battery icon work a percentage (xx%)
I probably wouldn't get into very many of the features of this hack, but still wanted to have it there in case I found I was at a point where I would like to use something.
I am finding there is a lot to be learned, thanks for the advice!

Sent from my Moto Z (2) using Tapatalk

[Walter Schulz]

Those are the instructions I followed, before I signed up on the forums.

Please link/quote the part of instructions talking about "stable build" having to be overwritten to work. We like to correct our errors.

The only thing I did not do is copy my ROM files back to my computer.

There is a chance getting a bricked cam to work using these files.

I was hoping to have recording bars at the top of the screen, and like the idea of having a battery icon work a percentage (xx%)

In movie mode toggle screens by pressing DISP button until ML screen appears.

[a1ex]

Looking for a quick test:
- qemu branch
- compile from minimal/qemu-frsp with "make MODEL=1100D"
- you'll get a small autoexec.bin with no dependencies (no other files required), that's going to capture a full-res silent picture and preview it
- if it works, please upload a picture of the camera screen (just want to check where the preview appears on the screen)
- the picture will not be saved on the card (it should only display a preview).

[dfort]

Tried the qemu-frsp quick test on a 500D and it works. I had to run it a few times because the timing of the button presses seem to make a difference. Anyway, the request is for testing on the 1100D. One of the few people I know who has access to this camera and can compile ML is on vacation so I uploaded a test file to my downloads page. Look for 1100D_qemu-frsp_test.zip

This is for the 1100D.105 and when you unzip it, the autoexec.bin file placed on a bootable card and run on a camera that has the bootflag set.

Note that this will capture a full-res silent picture. If you haven't used this feature before you'll probably be surprised by an overexposed image. That's because FRSP doesn't use the camera shutter and the capture speed is usually much slower than what you have set.

Here is what the test looks like on the 500D. Note that there is a black bar at the top of the image so it doesn't cover the entire screen.

[dfort]

Interesting. So the Canon overlays were captured in the image? Have you tried this a few times and does it always appears the same?

[a1ex]

Maybe you just have to press the buttons quicker. The test program only gives you 5 seconds to do so; after that, it attempts to capture the image regardless of whether you went to LiveView/PLAY or not.

Also make sure you are in M mode, and ISO must not be Auto. Otherwise you may get an error message.

[RhinoNelson]

OK - I deleted the earlier post. Now I believe I captured what you requested.



That is a lamp shade in the upper left of the image. The rest is overexposed but I believe it gives you what you were hopinh for.

[Floont]

After numerous Err 70s that seem to be related to a quirky SD card, I got the following (using 1100D_qemu-frsp_test):



(please excuse the crud-laden screen protector)

[greg_kennedy]

Hello there!  T3 owner and ML user here.  I am willing to test firmware that you throw at me.

My camera runs 1.0.5 right now.  Is 1.0.6 ML just as stable as 1.0.5?  If so, I may upgrade the "official" firmware.

[aplinhares]

Latestbuild isn't 1.05 still?
magiclantern-Nightly.2018Jul03.1100D105.zip

[dfort]

I worked on a port for the 1.0.6 Canon firmware. Don't have the camera so we're depending on user input. You can get a test build from my downloads page.

[dfort]

If you download the current build from the official downloads page you'll see lots of modules aren't building so a while back I submitted a fix for the 1100D.105 unified branch. Don't have the camera so we're depending on user input. You can get a test build from my downloads page. Please post your test results here or on the pull request.

[greg_kennedy]

I worked on a port for the 1.0.6 Canon firmware. Don't have the camera so we're depending on user input. You can get a test build from my downloads page.

Uh oh... I uninstalled ML from 1.0.5 camera by doing the firmware update and waiting one minute.  Then I upgraded the camera to 1.0.6, downloaded your latest test Nightly, placed it on a card, and attempted to re-enable ML booting through the Firmware Update.

Of course, as I now see reading back one page, ML-SETUP.FIR is not updated to 1.0.6 support yet.  So on my camera the firmware update screen hangs with "wrong firmware version, expected 105, pull the battery".

I tried using EOSCard to set BOOTDISK and EOS_DEVELOP but, of course, even if the *card* is marked bootable, the *camera* is not hacked to read it.

How are people getting 1.0.6 ML to boot without a 1.0.6 ML-SETUP.FIR?  Can I get one made somehow?

[Walter Schulz]

Best using 2 cards. One for ML and an obsolete one for Canon firmware.
Format both cards.
Use "obsolete" to downgrade to Canon firmware 1.0.5.
Copy ML for 1.0.5 to "ML card" and install ML on cam. Remove card.
Take "obsolete" and upgrade cam to Canon firmware 1.0.6.
Take "ML card" and copy files for ML (for 1.0.6) to card. Overwrite all files.

Done.