Re: [Qemu-devel] [PATCH v5 3/4] block: add block timer and block throttling algorithm

[Stefan Hajnoczi]

On Fri, Aug 12, 2011 at 6:35 AM, Zhi Yong Wu <address@hidden> wrote:
> On Tue, Aug 9, 2011 at 4:57 PM, Ram Pai <address@hidden> wrote:
>> On Tue, Aug 09, 2011 at 12:17:51PM +0800, Zhi Yong Wu wrote:
>>> Note:
>>>       1.) When bps/iops limits are specified to a small value such as 511 
>>> bytes/s, this VM will hang up. We are considering how to handle this 
>>> senario.
>>>       2.) When "dd" command is issued in guest, if its option bs is set to 
>>> a large value such as "bs=1024K", the result speed will slightly bigger 
>>> than the limits.
>>>
>>> For these problems, if you have nice thought, pls let us know.:)
>>>
>>> Signed-off-by: Zhi Yong Wu <address@hidden>
>>> ---
>>>  block.c     |  347 
>>> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++--
>>>  block.h     |    6 +-
>>>  block_int.h |   30 +++++
>>>  3 files changed, 372 insertions(+), 11 deletions(-)
>>>
>>> diff --git a/block.c b/block.c
>>> index 24a25d5..8fd6643 100644
>>> --- a/block.c
>>> +++ b/block.c
>>> @@ -29,6 +29,9 @@
>>>  #include "module.h"
>>>  #include "qemu-objects.h"
>>>
>>> +#include "qemu-timer.h"
>>> +#include "block/blk-queue.h"
>>> +
>>>  #ifdef CONFIG_BSD
>>>  #include <sys/types.h>
>>>  #include <sys/stat.h>
>>> @@ -58,6 +61,13 @@ static int bdrv_read_em(BlockDriverState *bs, int64_t 
>>> sector_num,
>>>  static int bdrv_write_em(BlockDriverState *bs, int64_t sector_num,
>>>                           const uint8_t *buf, int nb_sectors);
>>>
>>> +static bool bdrv_exceed_bps_limits(BlockDriverState *bs, int nb_sectors,
>>> +        bool is_write, double elapsed_time, uint64_t *wait);
>>> +static bool bdrv_exceed_iops_limits(BlockDriverState *bs, bool is_write,
>>> +        double elapsed_time, uint64_t *wait);
>>> +static bool bdrv_exceed_io_limits(BlockDriverState *bs, int nb_sectors,
>>> +        bool is_write, uint64_t *wait);
>>> +
>>>  static QTAILQ_HEAD(, BlockDriverState) bdrv_states =
>>>      QTAILQ_HEAD_INITIALIZER(bdrv_states);
>>>
>>> @@ -90,6 +100,68 @@ int is_windows_drive(const char *filename)
>>>  }
>>>  #endif
>>>
>>> +/* throttling disk I/O limits */
>>> +void bdrv_io_limits_disable(BlockDriverState *bs)
>>> +{
>>> +    bs->io_limits_enabled = false;
>>> +    bs->req_from_queue    = false;
>>> +
>>> +    if (bs->block_queue) {
>>> +        qemu_block_queue_flush(bs->block_queue);
>>> +        qemu_del_block_queue(bs->block_queue);
>>> +    }
>>> +
>>> +    if (bs->block_timer) {
>>> +        qemu_del_timer(bs->block_timer);
>>> +        qemu_free_timer(bs->block_timer);
>>> +    }
>>> +
>>> +    bs->slice_start[0]   = 0;
>>> +    bs->slice_start[1]   = 0;
>>> +
>>> +    bs->slice_end[0]     = 0;
>>> +    bs->slice_end[1]     = 0;
>>> +}
>>> +
>>> +static void bdrv_block_timer(void *opaque)
>>> +{
>>> +    BlockDriverState *bs = opaque;
>>> +    BlockQueue *queue = bs->block_queue;
>>> +
>>> +    qemu_block_queue_flush(queue);
>>> +}
>>> +
>>> +void bdrv_io_limits_enable(BlockDriverState *bs)
>>> +{
>>> +    bs->req_from_queue = false;
>>> +
>>> +    bs->block_queue    = qemu_new_block_queue();
>>> +    bs->block_timer    = qemu_new_timer_ns(vm_clock, bdrv_block_timer, bs);
>>> +
>>> +    bs->slice_start[BLOCK_IO_LIMIT_READ]  = qemu_get_clock_ns(vm_clock);
>>> +    bs->slice_start[BLOCK_IO_LIMIT_WRITE] = qemu_get_clock_ns(vm_clock);
>>
>> a minor comment. better to keep the slice_start of the both the READ and 
>> WRITE
>> side the same.
>>
>>    bs->slice_start[BLOCK_IO_LIMIT_WRITE] = 
>> bs->slice_start[BLOCK_IO_LIMIT_READ];
>>
>> saves  a call to qemu_get_clock_ns().
>>
>>> +
>>> +    bs->slice_end[BLOCK_IO_LIMIT_READ]    =
>>> +                      qemu_get_clock_ns(vm_clock) + BLOCK_IO_SLICE_TIME;
>>
>>    bs->slice_end[BLOCK_IO_LIMIT_READ] = bs->slice_start[BLOCK_IO_LIMIT_READ] 
>> +
>>                        BLOCK_IO_SLICE_TIME;
>>
>> saves one more call to qemu_get_clock_ns()
>>
>>> +    bs->slice_end[BLOCK_IO_LIMIT_WRITE]   =
>>> +                      qemu_get_clock_ns(vm_clock) + BLOCK_IO_SLICE_TIME;
>>
>>
>>    bs->slice_end[BLOCK_IO_LIMIT_WRITE] = 
>> bs->slice_start[BLOCK_IO_LIMIT_WRITE] +
>>                        BLOCK_IO_SLICE_TIME;
>>
>> yet another call saving.
>>
>>
>>> +}
>>> +
>>> +bool bdrv_io_limits_enabled(BlockDriverState *bs)
>>> +{
>>> +    BlockIOLimit *io_limits = &bs->io_limits;
>>> +    if ((io_limits->bps[BLOCK_IO_LIMIT_READ] == 0)
>>> +         && (io_limits->bps[BLOCK_IO_LIMIT_WRITE] == 0)
>>> +         && (io_limits->bps[BLOCK_IO_LIMIT_TOTAL] == 0)
>>> +         && (io_limits->iops[BLOCK_IO_LIMIT_READ] == 0)
>>> +         && (io_limits->iops[BLOCK_IO_LIMIT_WRITE] == 0)
>>> +         && (io_limits->iops[BLOCK_IO_LIMIT_TOTAL] == 0)) {
>>> +        return false;
>>> +    }
>>> +
>>> +    return true;
>>> +}
>>
>> can be optimized to:
>>
>>        return (io_limits->bps[BLOCK_IO_LIMIT_READ]
>>                || io_limits->bps[BLOCK_IO_LIMIT_WRITE]
>>                || io_limits->bps[BLOCK_IO_LIMIT_TOTAL]
>>                || io_limits->iops[BLOCK_IO_LIMIT_READ]
>>                || io_limits->iops[BLOCK_IO_LIMIT_WRITE]
>>                || io_limits->iops[BLOCK_IO_LIMIT_TOTAL]);
> I want to apply this, but it violate qemu coding styles.

Perhaps checkpatch.pl complains because of the (...) around the return
value.  Try removing them.

Stefan