aboutsummaryrefslogtreecommitdiff
path: root/qemu-seccomp.c
diff options
context:
space:
mode:
authorPeter Maydell <peter.maydell@linaro.org>2018-05-25 13:18:01 +0100
committerPeter Maydell <peter.maydell@linaro.org>2018-05-25 14:29:07 +0100
commitbdab39e4d26346d708437a59d0cd54e9b828a87f (patch)
tree283fb72713ce14a424c218b9f0bbc97781cf902c /qemu-seccomp.c
parent35ef7b60d6381333f6b3f2184e0b778b35b45e92 (diff)
Correct CPACR reset value for v7 corestz-migration
In commit f0aff255700 we made cpacr_write() enforce that some CPACR bits are RAZ/WI and some are RAO/WI for ARMv7 cores. Unfortunately we forgot to also update the register's reset value. The effect was that (a) a guest that read CPACR on reset would not see ones in the RAO bits, and (b) if you did a migration before the guest did a write to the CPACR then the migration would fail because the destination would enforce the RAO bits and then complain that they didn't match the zero value from the source. Implement reset for the CPACR using a custom reset function that just calls cpacr_write(), to avoid having to duplicate the logic for which bits are RAO. This bug would affect migration for TCG CPUs which are ARMv7 with VFP but without one of Neon or VFPv3. Reported-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- This is sufficient that a save-and-reload while the romulus-bmc machine is in the bootloader will work. On the other hand if I do a save-and-reload after the kernel has started booting then we get the classic "guest hang after reload", so some state is still not being transferred somewhere (probably in a device in the machine model?)
Diffstat (limited to 'qemu-seccomp.c')
0 files changed, 0 insertions, 0 deletions