|
From: | Paolo Bonzini |
Subject: | Re: [Qemu-devel] [RFC PATCH v5 07/31] icount: implement icount requesting |
Date: | Wed, 03 Dec 2014 11:17:49 +0100 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.2.0 |
On 26/11/2014 11:39, Pavel Dovgalyuk wrote: > +int64_t cpu_get_instructions_counter(void) > +{ > + /* This function calls are synchnonized to timer changes, > + calling cpu_get_instructions_counter_locked without lock is safe */ > + int64_t icount = timers_state.qemu_icount; > + CPUState *cpu = current_cpu; > + > + if (cpu) { > + icount -= (cpu->icount_decr.u16.low + cpu->icount_extra); > + } > + return icount; Why do you need to do this if !cpu_can_do_io(cpu)? Perhaps a better name for the functions is - cpu_get_instructions_counter_locked -> cpu_get_icount_raw - cpu_get_instructions_counter -> cpu_get_icount_raw_nocheck This makes it clear that cpu_get_instructions_counter should raise questions to a reviewer. Paolo
[Prev in Thread] | Current Thread | [Next in Thread] |