On 10/18/2016 00:18, Larry Finger wrote:
On 10/17/2016 07:23 PM, Moby wrote:
On 10/11/2016 23:45, Moby wrote:
The general prescription is as follows:
git bisect reset git bisect start git bisect bad git bisect good v4.5
Build and test the resulting kernel, then
git bisect good or git bisect bad, whichever is appropriate. Repeat the previous step.
When yo do the above, what happens? If it fails, please post the console output.
Larry
The bisect is done - I think. In addition to the bug with the Makefile, I also kept running into a kernel bug that kept causing segfaults in wpa_supplicant. I started with kernel 4.8.1, which had the packet loss, and I did a bisect bad on it. Then I checked out v4.5, which did not have packet loss in it, and I did bisect good on it. From then on, almost every bisect resulted in a kernel where wpa_supplicant kept segfaulting. Ikept doing a bisect good on these, since I wanted the bisection to "proceed towards 4.8.1" - I am not sure if this was the right thing to do. However, at the end of all this, I am sitting at version 4.6.0-rc7-1-default+, and git bisect good replies with:
2dcd0af568b0cf583645c8a317dd12e344b1c72a is the first bad commit commit 2dcd0af568b0cf583645c8a317dd12e344b1c72a Author: Linus Torvalds
Date: Sun May 15 15:43:13 2016 -0700 Linux 4.6
:100644 100644 acf6155421cc244913b71c6d3b95cd17690e64ff 0f9cb36d45c2c59a589670679e7fa1d25ff9ee59 M Makefile
I am not sure if all this effort is valid, since the wpa_supplicant segfaults in most revisions preventing me from testing them.
No, that merge is not correct. The place where you ended up only added the v4.6 tag - it had no code associated with it. By arbitrarily calling everything "good", you had no chance to find anything that was "bad". I have no idea what was causing the segfaults for wpa_supplicant. I have not seen anything like that.
Larry
Well, I did have a bisect bad - this was with the version I started the bisect with, 4.8.1. That version showed the packet loss issue. After that, I went to 4.5.1, and it was good - but almost all versions after 4.5.1 had the wpa_supplicant segfault in them and only way I could think of progressing closer to the "culprit commit" was to keep labeling them good. -- --Moby -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org