>This is a follow up to an email chain that in November: > > http://www.seabios.org/pipermail/seabios/2015-November/009887.html > >It was possible for the SeaBIOS code to get confused if an external >reboot request occurs while seabios is already in the process of >handling a boot or reboot. > >This two patch series attempts to make the SeaBIOS reboot code more >robust. With these patches the code maintains an invariant - if >HaveRunPost is false then the BIOS code will be in a pristine state >and it is okay to run the "post" boot logic; if HaveRunPost is true >then the reboot logic must be run before starting the "post" boot >logic. The reboot handler should always return the BIOS code to a >pristine state prior to clearing HaveRunPost.
Kevin, thanks for your patch. I've tested yesterday for several days. And it works very well solving all the reset racing problems i've met. > >-Kevin > > >Kevin O'Connor (2): > resume: Make KVM soft reboot loop detection more flexible > post: Always set HaveRunPost prior to setting any other global > variable > > src/fw/csm.c | 1 + > src/fw/shadow.c | 17 +++++++++++++++-- > src/fw/xen.c | 1 + > src/post.c | 20 ++++++++++++++++---- > src/resume.c | 9 --------- > src/util.h | 1 + > 6 files changed, 34 insertions(+), 15 deletions(-) > >-- >2.5.0 > _______________________________________________ SeaBIOS mailing list SeaBIOS@seabios.org http://www.seabios.org/mailman/listinfo/seabios