Using the DAWR on POWER9 can cause xstops, hence we need to disable it. The current CPU_FTR for DAWR is a bit messy. Despite having CPU_FTR_DAWR, currently we assume DAWR exists in the KVM code based on CPU_FTR_ARCH_207. In other places we assume DAWR exists if CPU_FTR_DAWR is set. This attempts to clear up the situation by always using CPU_FTR_DAWR before setting the DAWR (to a non-zero value). DAWR has 5 different ways of being set from userspace. ptrace, h_set_mode, h_set_mode(DAWR), h_set_dabr(), kvmppc_set_one_reg() and xmon. For ptrace, we now advertise zero breakpoints on POWER9 via the PPC_PTRACE_GETHWDBGINFO call. This results in GDB falling back to software emulation of the watchpoint (which is slow). h_set_mode() and h_set_dabr() will now return an error to the guest when on a POWER9 host. Current Linux guests ignore this error, so they will silently not get the DAWR (sigh). The same error codes are being used by POWERVM in this case. kvmppc_set_one_reg() will store the value in the vcpu but won't actually set it on POWER9 hardware. This is done so we don't break migration from P8 to P9, at the cost of silently losing the DAWR on the migration. This is not ideal but hopefully the best overall solution. This approach has been acked by paulus. For xmon, the 'bd' command will return an error on P9. Thanks to Pedro Franco de Carvalho for the initial version of this. v3: Split patches Fix usage of H_UNSUPPORTED v2: Move to CPU feature quirk from Nick's dbginfo.num_data_bps = 0 if !breakpoint_available() from Pedro -- To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
