Fix USB handling and timeouts for newer devices #329
Conversation
|
I'm going to take back the last one being completely fixed on USB3.0, but it is significantly better at least when downloading the PIT. USB2.0 works without any problems. I still get the occasional "WARNING: Empty bulk transfer after sending packet failed. Continuing anyway..." as well, but far fewer and they don't really cause problems. |
|
Could you please separate these into 3 separate commits?
I will very happily merge both of these - thanks for the awesome work.
Definitely useful information, unfortunately I can't merge this because older devices do require the use of Also, what model is your S5 and what is the firmware version the secondary bootloader was included with? I ask this because I also own a S5 (G900I) and it's running flawlessly with whatever secondary bootloader I have on there (it's not with me at the moment so I can't provide the version unfortunately). |
Sorry, to clarify, I mean running flawlessly with the current master build i.e. It doesn't require the changes you've mentioned. |
|
Sorry for the delay - this week has been... hectic. I'll certainly split those commits for you later on this week. I actually have Cyanogenmod installed right now, but at the time I was seeing these issues, I had a stock 5.1.1 ROM on a T-Mobile SM-G900T. It was factory reconditioned when I bought it, so I imagine it would have been the latest SBL. I'm not sure how to get the actual SBL version, though - I'm a Linux administrator with some programming experience, not an Android expert. How would I get that information for you? |
|
Any update on this? |
|
I'm having this same issue, Samsung Galaxy S4, international, non-LTE version (GT-i9500) C:.......\Heimdall Suite>heimdall detect C:.......\Heimdall Suite>heimdall flash --RECOVERY ....\cm-12.1-20160417-NIGHTLY-i9500-recovery.img --no-reboot --verbose Copyright (c) 2010-2013, Benjamin Dobell, Glass Echidna This software is provided free of charge. Copying and redistribution is If you appreciate this software and you would like to support future Initialising connection... iMan:iProd:iSer: 1:2:0 interface[0].altsetting[0]: num endpoints = 1 interface[1].altsetting[0]: num endpoints = 2 Initialising protocol... C:.......\Heimdall Suite> |
|
@richieacc You're running 1.4.0, please try the latest from master before commenting on issues that may have already been fixed. |
|
@Benjamin-Dobell Thanks! I read about the update after posting the comment. I have built Heimdall from source, and it worked. |
|
Hey, I just cloned and built Heimdall from source and get this error for a GT-S7580:
Edit: Ok, like always.. After posting this and restarting the device again it successfully goes further now: It's writing this message several times until the phone restarts... |
|
@sosnet did you apply this patch? |
|
@EvanCarroll |
|
@sosnet still shows not merged.. Just hit the merge button. |
And break support for thousands of people's devices? No, thanks 😉 As mentioned, if the pull request is split up into 2-3 smaller pieces it's easier for me to test and merge each of the functional changes individually. |
|
I tried this patch and it didn't work for me: |
|
It doesn't work for me either: Log here. |
Neither does it for me |
|
This patch worked for me to fix USB issues with Galaxy Tab S2 (Samsung SM-T813) on Ubuntu 16.04 LTS. |
|
@procrastinat0r could you share the pcap dump? |
|
SGS5 g900t user here. This patch changes nothing of the behavior for me on the S5. still get ERROR: Failed to receive handshake response. Result: -7
Heimdall v1.4.2
Copyright (c) 2010-2017 Benjamin Dobell, Glass Echidna This software is provided free of charge. Copying and redistribution is If you appreciate this software and you would like to support future Initialising connection...
|
|
@sanimalp |
|
I found that this patch plus one other change seemed to work for me. Specifically, within |
|
@amo13 The device is a Samsung Galaxy Tab 4 (Model SM-T230NU, Android version 4.4.2). The error I was originally getting was this: (I added the part to translate the -7 into something a little more human readable.) In addition to changing the timeout for the ODIN message as mentioned above, I also found that I needed to modify the code to try twice to send the |
That would be awesome! Thanks for the hints. Why the need to flash a CACHE image though? |
|
I have done some more thorough testing and believe that the problem is not timeouts at all. Specifically, I have determined with two Linux machines that I have, if I reset the Samsung tablet into ODIN mode and execute |
|
Oh, I have come across this issue quite a few times with a few different devices. I believe it is also somewhat known and you'll read about other users reporting this behavior in some other issues on here... Still, in my experience of flashing samsung devices with heimdall, even though I have seen this behavior a few times, most devices will also accept further commands. |
This commit should fix #228 #230 #209 #276 #190 #307 #141 #225 #286 for most new devices (they were all related to the same problems). The issues found were: