[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH] Fix xbzrle vs last_sent_block update
From: |
Juan Quintela |
Subject: |
Re: [Qemu-devel] [PATCH] Fix xbzrle vs last_sent_block update |
Date: |
Fri, 11 Dec 2015 13:25:50 +0100 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/24.5 (gnu/linux) |
Peter Maydell <address@hidden> wrote:
D> On 10 December 2015 at 16:31, Dr. David Alan Gilbert (git)
> <address@hidden> wrote:
>> From: "Dr. David Alan Gilbert" <address@hidden>
>>
>> My fix (84e7b80a) replaced the last_sent_block update that I'd
>> removed earlier; however it was too aggressive in the xbzrle case.
>>
>> save_xbzrle_page might return '0' to mean that the page didn't
>> need sending since it was the same as the last sent version;
>> in this case we can't update 'last_sent_block' since we didn't
>> actually send it.
>>
>> Symptom: 'Illegal RAM offset 1018000' as we try and send a page
>> to the wrong RAMBlock; potentially that could be a data
>> corruption if you were really unlucky.
>>
>> Fixes: 84e7b80a05c0c44b90533c6cd2f1db5c932ccf77
>>
>> Signed-off-by: Dr. David Alan Gilbert <address@hidden>
>> ---
>> migration/ram.c | 11 ++++++++++-
>> 1 file changed, 10 insertions(+), 1 deletion(-)
>>
>> diff --git a/migration/ram.c b/migration/ram.c
>> index 1eb155a..0490f00 100644
>> --- a/migration/ram.c
>> +++ b/migration/ram.c
>> @@ -716,6 +716,9 @@ static int save_zero_page(QEMUFile *f, RAMBlock *block,
>> ram_addr_t offset,
>> * ram_save_page: Send the given page to the stream
>> *
>> * Returns: Number of pages written.
>> + * < 0 - error
>> + * >=0 - Number of pages written - this might legally be 0
>> + * if xbzrle noticed the page was the same.
>> *
>> * @f: QEMUFile where to send the data
>> * @block: block that contains the page we want to send
>> @@ -1249,7 +1252,13 @@ static int ram_save_target_page(MigrationState *ms,
>> QEMUFile *f,
>> if (unsentmap) {
>> clear_bit(dirty_ram_abs >> TARGET_PAGE_BITS, unsentmap);
>> }
>> - last_sent_block = block;
>> + /* Only update last_sent_block if a block was actually sent; xbzrle
>> + * might have decided the page was identical so didn't bother
>> writing
>> + * to the stream.
>> + */
>> + if (res > 0) {
>> + last_sent_block = block;
>> + }
>> }
>>
>> return res;
>
> This sounds like we should probably put this into 2.5; I'm happy
> to do so if it gets review by tomorrow afternoon and Juan/Amit
> agree.
Yeap, did the review by. Do you want a pull request, or just pick it
directly?