qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Yoshiaki Tamura <tamura.yoshiaki@lab.ntt.co.jp>
To: Avi Kivity <avi@redhat.com>
Cc: aliguori@us.ibm.com, kvm@vger.kernel.org,
	ohmura.kei@lab.ntt.co.jp, mtosatti@redhat.com,
	qemu-devel@nongnu.org
Subject: [Qemu-devel] Re: [PATCH v2 2/6] Introduce bit-based phys_ram_dirty for VGA, CODE, MIGRATION and MASTER.
Date: Mon, 12 Apr 2010 18:39:48 +0900	[thread overview]
Message-ID: <4BC2EA64.6060700@lab.ntt.co.jp> (raw)
In-Reply-To: <4BC2D340.2030402@redhat.com>

Avi Kivity wrote:
> On 04/06/2010 03:51 AM, Yoshiaki Tamura wrote:
>> Replaces byte-based phys_ram_dirty bitmap with three bit-based
>> phys_ram_dirty
>> bitmap. On allocation, it sets all bits in the bitmap.
>>
>>
>
>> index c74b0a4..9733892 100644
>> --- a/exec.c
>> +++ b/exec.c
>> @@ -110,7 +110,7 @@ uint8_t *code_gen_ptr;
>>
>> #if !defined(CONFIG_USER_ONLY)
>> int phys_ram_fd;
>> -uint8_t *phys_ram_dirty;
>> +unsigned long *phys_ram_dirty[NUM_DIRTY_FLAGS];
>> static int in_migration;
>>
>> typedef struct RAMBlock {
>> @@ -2825,10 +2825,32 @@ ram_addr_t qemu_ram_alloc(ram_addr_t size)
>> new_block->next = ram_blocks;
>> ram_blocks = new_block;
>>
>> - phys_ram_dirty = qemu_realloc(phys_ram_dirty,
>> - (last_ram_offset + size)>> TARGET_PAGE_BITS);
>> - memset(phys_ram_dirty + (last_ram_offset>> TARGET_PAGE_BITS),
>> - 0xff, size>> TARGET_PAGE_BITS);
>> +/* temporarily copy from qemu-kvm.git/qemu-kvm.h */
>> +#define ALIGN(x, y) (((x)+(y)-1)& ~((y)-1))
>> +#define BITMAP_SIZE(m) (ALIGN(((m)>>TARGET_PAGE_BITS),
>> HOST_LONG_BITS) / 8)
>
> Please put in some header file, maybe qemu-common.h.

OK.  BTW, is qemu-kvm.h planned to go upstream?

>> +
>> + if (BITMAP_SIZE(last_ram_offset + size) !=
>> BITMAP_SIZE(last_ram_offset)) {
>> + phys_ram_dirty[MASTER_DIRTY_FLAG] =
>> + qemu_realloc(phys_ram_dirty[MASTER_DIRTY_FLAG],
>> + BITMAP_SIZE(last_ram_offset + size));
>> + phys_ram_dirty[VGA_DIRTY_FLAG]
>> + = qemu_realloc(phys_ram_dirty[VGA_DIRTY_FLAG],
>> + BITMAP_SIZE(last_ram_offset + size));
>> + phys_ram_dirty[CODE_DIRTY_FLAG] =
>> + qemu_realloc(phys_ram_dirty[CODE_DIRTY_FLAG],
>> + BITMAP_SIZE(last_ram_offset + size));
>> + phys_ram_dirty[MIGRATION_DIRTY_FLAG] =
>> + qemu_realloc(phys_ram_dirty[MIGRATION_DIRTY_FLAG],
>> + BITMAP_SIZE(last_ram_offset + size));
>> + memset((uint8_t *)phys_ram_dirty[MASTER_DIRTY_FLAG] +
>> + BITMAP_SIZE(last_ram_offset), 0xff, BITMAP_SIZE(size));
>> + memset((uint8_t *)phys_ram_dirty[VGA_DIRTY_FLAG] +
>> + BITMAP_SIZE(last_ram_offset), 0xff, BITMAP_SIZE(size));
>> + memset((uint8_t *)phys_ram_dirty[CODE_DIRTY_FLAG] +
>> + BITMAP_SIZE(last_ram_offset), 0xff, BITMAP_SIZE(size));
>> + memset((uint8_t *)phys_ram_dirty[MIGRATION_DIRTY_FLAG] +
>> + BITMAP_SIZE(last_ram_offset), 0xff, BITMAP_SIZE(size));
>> + }
>>
>
> Should be nicer as a loop calling a helper to allocate each bitmap. This
> patch won't work by itself, will it? I think you need to merge it with
> the succeeding patches.

Yeah.  I originally wrote as you suggested, but I needed to skip 0x03 because of 
the reason written below.

> +/* Use DIRTY_FLAG as indexes of bit-based phys_ram_dirty.
> +   0x03 is empty to make it compatible with byte-based bitmap. */
> +#define MASTER_DIRTY_FLAG    0x00
>  #define VGA_DIRTY_FLAG       0x01
>  #define CODE_DIRTY_FLAG      0x02
> -#define MIGRATION_DIRTY_FLAG 0x08
> +#define MIGRATION_DIRTY_FLAG 0x04

If you think if (i != 0x03) is better, I would modify it to a loop.

And sorry, you need to merge the succeeding patches to make it work.

  reply	other threads:[~2010-04-12  9:40 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-04-06  0:51 [Qemu-devel] [PATCH v2 0/6] Introduce bit-based phys_ram_dirty, and bit-based dirty page checker Yoshiaki Tamura
2010-04-06  0:51 ` [Qemu-devel] [PATCH v2 1/6] Modify DIRTY_FLAG value to use as indexes of bit-based phys_ram_dirty Yoshiaki Tamura
2010-04-06  0:51 ` [Qemu-devel] [PATCH v2 2/6] Introduce bit-based phys_ram_dirty for VGA, CODE, MIGRATION and MASTER Yoshiaki Tamura
2010-04-12  8:01   ` [Qemu-devel] " Avi Kivity
2010-04-12  9:39     ` Yoshiaki Tamura [this message]
2010-04-12 10:17       ` Avi Kivity
2010-04-06  0:51 ` [Qemu-devel] [PATCH v2 3/6] Modifies wrapper functions for byte-based phys_ram_dirty bitmap to bit-based phys_ram_dirty bitmap Yoshiaki Tamura
2010-04-12  8:10   ` [Qemu-devel] " Avi Kivity
2010-04-12 10:58     ` Yoshiaki Tamura
2010-04-12 11:09       ` Avi Kivity
2010-04-13  8:01         ` Yoshiaki Tamura
2010-04-13  9:20           ` Avi Kivity
2010-04-13 10:49             ` Yoshiaki Tamura
2010-04-06  0:51 ` [Qemu-devel] [PATCH v2 4/6] Introduce cpu_physical_memory_get_dirty_range() Yoshiaki Tamura
2010-04-06  0:51 ` [Qemu-devel] [PATCH v2 5/6] Use cpu_physical_memory_set_dirty_range() to update phys_ram_dirty Yoshiaki Tamura
2010-04-06  0:51 ` [Qemu-devel] [PATCH v2 6/6] Use cpu_physical_memory_get_dirty_range() to check multiple dirty pages Yoshiaki Tamura

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4BC2EA64.6060700@lab.ntt.co.jp \
    --to=tamura.yoshiaki@lab.ntt.co.jp \
    --cc=aliguori@us.ibm.com \
    --cc=avi@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=mtosatti@redhat.com \
    --cc=ohmura.kei@lab.ntt.co.jp \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).