On Mon, Oct 29, 2018 at 10:16 PM Waldek Kozaczuk <[email protected]> wrote:
> No I have realized that another method, I have added, make_text_writable > uses a loop with r-value. Honestly I am not sure if that has some > side-effect I am not understanding. > This is not an rvalue but rather a "universal reference" - the fact both look like "&&" in C++ is really terrible. I think it should be fine... > > > On Monday, October 29, 2018 at 1:32:18 PM UTC-4, Waldek Kozaczuk wrote: >> >> I ran all the tests twice again on both Ubuntu and Fedora and all passed. >> >> I wonder if the fact I extracted the get_segment_mmap_permissions() >> method somehow inadvertently changes phdr.p_flags and then this assert >> fails. Maybe some gcc optimization and we have different gcc on the build >> machine. Maybe different newer Fedora - I think Fedora 29 is about to be >> released so we have have some beta there? >> > Avi, did you change something in the Jenkin's software setup a few days ago? I'll install Fedora 29 on my own machine and check if I can reproduce this... > >> Waldek >> >> On Monday, October 29, 2018 at 11:57:41 AM UTC-4, Waldek Kozaczuk wrote: >>> >>> This definitely seems possibly related to the changes I made. But how >>> come I cannot reproduce it on both Ubuntu and Fedora on both my machine and >>> virtual machines? >>> >>> Clean? >>> >>> On Monday, October 29, 2018 at 11:55:16 AM UTC-4, Nadav Har'El wrote: >>>> >>>> >>>> >>>> On Mon, Oct 29, 2018 at 5:33 PM <[email protected]> wrote: >>>> >>>>> See < >>>>> http://jenkins.cloudius-systems.com:8080/job/osv-build/1429/display/redirect?page=changes >>>>> > >>>>> >>>>> Changes: >>>>> >>>>> [jwkozaczuk] Support non-fPIC shared objects >>>>> >>>>> [Nadav Har'El] tests makefile: clean up non-fpic compilation >>>>> ... >>>>> LD build/release.x64/boot.bin >>>>> DD loader.img boot.bin >>>>> DD loader.img lzloader.elf >>>>> IMGEDIT build/release.x64/loader.img >>>>> IMGEDIT build/release.x64/loader.img >>>>> AS arch/x64/boot32.S >>>>> LD build/release.x64/loader.bin >>>>> make[1]: Leaving directory '< >>>>> http://jenkins.cloudius-systems.com:8080/job/osv-build/ws/'> >>>>> OSv v0.52.0-5-g65f0e384 >>>>> eth0: 192.168.122.15 >>>>> Assertion failed: (phdr.p_flags & (PF_R | PF_W | PF_X)) == PF_R >>>>> (core/elf.cc: fix_permissions: 495) >>>>> >>>>> >>>> So at this point OSv was trying to run /tools/mkfs.so; tools/cpiod.so >>>> (see scripts/upload_manifest.py) and I think >>>> the first one crashed on this error. I don't know why, how did the >>>> latest changes (which did not change phdr.p_flags) >>>> make this assertion fail? Waldek, do you have any ideas? >>>> >>>> I can't reproduce this on my machine. "make check" succeeds for me. >>>> >>> -- > You received this message because you are subscribed to the Google Groups > "OSv Development" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "OSv Development" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
