I wrote: "PS: it's possible that that commit doesn't actually fix the
underlying kernel crash, it just means that rr isn't triggering it any more,
and that if you modified EFLAGS via the ptrace interface rather than r11 you'd
get the crash back again."
but looking at the kernel I think that is i
PS: it's possible that that commit doesn't actually fix the underlying
kernel crash, it just means that rr isn't triggering it any more, and
that if you modified EFLAGS via the ptrace interface rather than r11
you'd get the crash back again.
--
You received this bug notification because you are a
I've now completed the kernel git bisect. git bisect says the commit
which fixed this issue is 29722cd4ef666705b2eda1c3ba44435488e509eb
("x86/asm/entry/64: Save R11 into pt_regs->flags on SYSCALL64
fastpath").
This fits in with the discovery on the rr side that the rr commit which
started causing
I retried with a workaround for the rr bug I described in "caveat (2)"
and the kernel still does not lockup, so I am now confident that this
bug is not present in the upstream kernel.
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux i
I tested with linux-headers-4.4.0-040400 / linux-
headers-4.4.0-040400-generic / linux-image-4.4.0-040400-generic
4.4.0-040400.201601101930. The kernel lockups did *not* reproduce.
Two caveats:
(1) I did get this kernel warning in the log:
Jan 21 18:00:08 e104462 kernel: [ 171.577000] --
No, I don't think this was associated with a kernel upgrade, though it's
hard to say for certain as I hadn't run the test suite in some months. I
suspect it's more that rr's test suite got more complex and included
some stress tests that reveal pre-existing kernel bugs.
I will test the upstream ke
Public bug reported:
Running the 'rr' make check on Ubuntu Trusty causes the machine to
become unusable because the kernel crashes. (rr build-and-make-check
instructions: https://github.com/mozilla/rr/wiki/Building-And-
Installing)
This is 3.13.0-74-generic #118-Ubuntu for x86_64.
This is rr bug
7 matches
Mail list logo