Re: [Qemu-devel] [PATCH untested for-4.2] memory: fix race between TCG and accesses to dirty bitmap

[Paolo Bonzini]

On 06/08/19 16:23, Peter Maydell wrote:
> On Mon, 29 Jul 2019 at 22:47, Paolo Bonzini <address@hidden> wrote:
>>
>> The race is as follows:
>>
>>       vCPU thread                  reader thread
>>       -----------------------      -----------------------
>>       TLB check -> slow path
>>         notdirty_mem_write
>>           write to RAM
>>           set dirty flag
>>                                    clear dirty flag
>>       TLB check -> fast path
>>                                    read memory
>>         write to RAM
>>
>> and the second write is missed by the reader.
>>
>> Fortunately, in order to fix it, no change is required to the
>> vCPU thread.  However, the reader thread must delay the read after
>> the vCPU thread has finished the write.  This can be approximated
>> conservatively by run_on_cpu, which waits for the end of the current
>> translation block.
>>
>> A similar technique is used by KVM, which has to do a synchronous TLB
>> flush after doing a test-and-clear of the dirty-page flags.
>>
>> Reported-by: Dr. David Alan Gilbert <address@hidden>
>> Signed-off-by: Paolo Bonzini <address@hidden>
>> ---
>>         I tested this some time ago, and enough has changed that I don't
>>         really trust those old results.  Nevertheless, I am throwing out
>>         the patch so that it is not forgotten.
> 
> This patch looks almost the same (maybe identical except for the
> commit message title?) as the patch "memory: introduce
> memory_global_after_dirty_log_sync" which you sent out at almost
> the same time as this one. Which patch should we be reviewing?

Yes, it's the same except for the commit message title.  I forgot a "-1"
after editing the .patch file.

Paolo