From: Fam Zheng <famz@redhat.com>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: qemu-devel@nongnu.org, stefanha@redhat.com, mst@redhat.com
Subject: Re: [Qemu-devel] [PATCH 23/29] bitmap: add atomic set functions
Date: Tue, 26 May 2015 19:54:16 +0800 [thread overview]
Message-ID: <20150526115416.GV13749@ad.nay.redhat.com> (raw)
In-Reply-To: <1430152117-100558-24-git-send-email-pbonzini@redhat.com>
On Mon, 04/27 18:28, Paolo Bonzini wrote:
> From: Stefan Hajnoczi <stefanha@redhat.com>
>
> Use atomic_or() for atomic bitmaps where several threads may set bits at
> the same time. This avoids the race condition between threads loading
> an element, bitwise ORing, and then storing the element.
>
> When setting all bits in a word we can avoid atomic ops and instead just
> use an smp_mb() at the end.
>
> Most bitmap users don't need atomicity so introduce new functions.
>
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> Message-Id: <1417519399-3166-2-git-send-email-stefanha@redhat.com>
> [Avoid barrier in the single word case, use full barrier instead of write.
> - Paolo]
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
> include/qemu/bitmap.h | 2 ++
> include/qemu/bitops.h | 14 ++++++++++++++
> util/bitmap.c | 37 +++++++++++++++++++++++++++++++++++++
> 3 files changed, 53 insertions(+)
>
> diff --git a/include/qemu/bitmap.h b/include/qemu/bitmap.h
> index f0273c9..3e0a4f3 100644
> --- a/include/qemu/bitmap.h
> +++ b/include/qemu/bitmap.h
> @@ -39,6 +39,7 @@
> * bitmap_empty(src, nbits) Are all bits zero in *src?
> * bitmap_full(src, nbits) Are all bits set in *src?
> * bitmap_set(dst, pos, nbits) Set specified bit area
> + * bitmap_set_atomic(dst, pos, nbits) Set specified bit area with atomic ops
> * bitmap_clear(dst, pos, nbits) Clear specified bit area
> * bitmap_find_next_zero_area(buf, len, pos, n, mask) Find bit free area
> */
> @@ -226,6 +227,7 @@ static inline int bitmap_intersects(const unsigned long *src1,
> }
>
> void bitmap_set(unsigned long *map, long i, long len);
> +void bitmap_set_atomic(unsigned long *map, long i, long len);
> void bitmap_clear(unsigned long *map, long start, long nr);
> unsigned long bitmap_find_next_zero_area(unsigned long *map,
> unsigned long size,
> diff --git a/include/qemu/bitops.h b/include/qemu/bitops.h
> index 90ca8df..95f98fb 100644
> --- a/include/qemu/bitops.h
> +++ b/include/qemu/bitops.h
> @@ -16,6 +16,7 @@
> #include <assert.h>
>
> #include "host-utils.h"
> +#include "atomic.h"
>
> #define BITS_PER_BYTE CHAR_BIT
> #define BITS_PER_LONG (sizeof (unsigned long) * BITS_PER_BYTE)
> @@ -39,6 +40,19 @@ static inline void set_bit(long nr, unsigned long *addr)
> }
>
> /**
> + * set_bit_atomic - Set a bit in memory atomically
> + * @nr: the bit to set
> + * @addr: the address to start counting from
> + */
> +static inline void set_bit_atomic(long nr, unsigned long *addr)
> +{
> + unsigned long mask = BIT_MASK(nr);
> + unsigned long *p = addr + BIT_WORD(nr);
> +
> + atomic_or(p, mask);
> +}
> +
> +/**
> * clear_bit - Clears a bit in memory
> * @nr: Bit to clear
> * @addr: Address to start counting from
> diff --git a/util/bitmap.c b/util/bitmap.c
> index 9c6bb52..6838d49 100644
> --- a/util/bitmap.c
> +++ b/util/bitmap.c
> @@ -11,6 +11,7 @@
>
> #include "qemu/bitops.h"
> #include "qemu/bitmap.h"
> +#include "qemu/atomic.h"
>
> /*
> * bitmaps provide an array of bits, implemented using an an
> @@ -177,6 +178,42 @@ void bitmap_set(unsigned long *map, long start, long nr)
> }
> }
>
> +void bitmap_set_atomic(unsigned long *map, long start, long nr)
> +{
> + unsigned long *p = map + BIT_WORD(start);
> + const long size = start + nr;
s/size/end/ ?
Otherwise,
Reviewed-by: Fam Zheng <famz@redhat.com>
> + int bits_to_set = BITS_PER_LONG - (start % BITS_PER_LONG);
> + unsigned long mask_to_set = BITMAP_FIRST_WORD_MASK(start);
> +
> + /* First word */
> + if (nr - bits_to_set > 0) {
> + atomic_or(p, mask_to_set);
> + nr -= bits_to_set;
> + bits_to_set = BITS_PER_LONG;
> + mask_to_set = ~0UL;
> + p++;
> + }
> +
> + /* Full words */
> + while (nr - bits_to_set >= 0) {
> + *p = ~0UL;
> + nr -= bits_to_set;
> + mask_to_set = ~0UL;
> + p++;
> + }
For full words, will memset be faster by any chance?
> +
> + /* Last word */
> + if (nr) {
> + mask_to_set &= BITMAP_LAST_WORD_MASK(size);
> + atomic_or(p, mask_to_set);
> + } else {
> + /* If we avoided the full barrier in atomic_or(), issue a
> + * barrier to account for the assignments in the while loop.
> + */
> + smp_mb();
> + }
> +}
> +
> void bitmap_clear(unsigned long *map, long start, long nr)
> {
> unsigned long *p = map + BIT_WORD(start);
> --
> 1.8.3.1
>
>
next prev parent reply other threads:[~2015-05-26 11:54 UTC|newest]
Thread overview: 88+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-04-27 16:28 [Qemu-devel] [PATCH v2 00/29] Dirty bitmap atomic access and optimizations Paolo Bonzini
2015-04-27 16:28 ` [Qemu-devel] [PATCH 01/29] memory: the only dirty memory flag for users is DIRTY_MEMORY_VGA Paolo Bonzini
2015-05-26 7:00 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 02/29] g364fb: remove pointless call to memory_region_set_coalescing Paolo Bonzini
2015-05-26 12:53 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 03/29] display: enable DIRTY_MEMORY_VGA tracking explicitly Paolo Bonzini
2015-05-26 7:13 ` Fam Zheng
2015-05-26 8:14 ` Paolo Bonzini
2015-05-26 11:24 ` Paolo Bonzini
2015-05-26 11:52 ` Peter Maydell
2015-05-26 11:56 ` Paolo Bonzini
2015-05-26 12:08 ` Peter Maydell
2015-05-26 12:10 ` Paolo Bonzini
2015-05-26 12:16 ` Peter Maydell
2015-05-31 20:32 ` Mark Cave-Ayland
2015-04-27 16:28 ` [Qemu-devel] [PATCH 04/29] display: add memory_region_sync_dirty_bitmap calls Paolo Bonzini
2015-05-26 7:36 ` Fam Zheng
2015-05-26 8:14 ` Paolo Bonzini
2015-04-27 16:28 ` [Qemu-devel] [PATCH 05/29] memory: differentiate memory_region_is_logging and memory_region_get_dirty_log_mask Paolo Bonzini
2015-05-26 7:46 ` Fam Zheng
2015-05-26 11:35 ` Paolo Bonzini
2015-04-27 16:28 ` [Qemu-devel] [PATCH 06/29] memory: prepare for multiple bits in the dirty log mask Paolo Bonzini
2015-05-26 7:55 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 07/29] framebuffer: check memory_region_is_logging Paolo Bonzini
2015-05-26 8:02 ` Fam Zheng
2015-05-26 11:16 ` Paolo Bonzini
2015-04-27 16:28 ` [Qemu-devel] [PATCH 08/29] ui/console: remove dpy_gfx_update_dirty Paolo Bonzini
2015-05-26 8:03 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 09/29] memory: track DIRTY_MEMORY_CODE in mr->dirty_log_mask Paolo Bonzini
2015-05-26 8:06 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 10/29] kvm: accept non-mapped memory in kvm_dirty_pages_log_change Paolo Bonzini
2015-05-26 8:10 ` Fam Zheng
2015-05-26 8:16 ` Paolo Bonzini
2015-05-26 8:22 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 11/29] memory: include DIRTY_MEMORY_MIGRATION in the dirty log mask Paolo Bonzini
2015-05-26 8:40 ` Fam Zheng
2015-05-26 9:07 ` Paolo Bonzini
2015-05-26 9:22 ` Fam Zheng
2015-05-26 9:26 ` Paolo Bonzini
2015-05-26 9:48 ` Fam Zheng
2015-05-26 9:07 ` Paolo Bonzini
2015-04-27 16:28 ` [Qemu-devel] [PATCH 12/29] kvm: remove special handling of " Paolo Bonzini
2015-05-26 8:56 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 13/29] ram_addr: tweaks to xen_modified_memory Paolo Bonzini
2015-05-26 12:52 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 14/29] exec: use memory_region_get_dirty_log_mask to optimize dirty tracking Paolo Bonzini
2015-05-26 10:42 ` Fam Zheng
2015-05-26 10:58 ` Paolo Bonzini
2015-05-26 11:12 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 15/29] exec: move functions to translate-all.h Paolo Bonzini
2015-05-26 10:50 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 16/29] translate-all: remove unnecessary argument to tb_invalidate_phys_range Paolo Bonzini
2015-05-26 10:51 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 17/29] cputlb: remove useless arguments to tlb_unprotect_code_phys, rename Paolo Bonzini
2015-05-26 10:53 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 18/29] translate-all: make less of tb_invalidate_phys_page_range depend on is_cpu_write_access Paolo Bonzini
2015-04-27 16:28 ` [Qemu-devel] [PATCH 19/29] exec: pass client mask to cpu_physical_memory_set_dirty_range Paolo Bonzini
2015-05-26 11:08 ` Fam Zheng
2015-05-26 11:28 ` Paolo Bonzini
2015-04-27 16:28 ` [Qemu-devel] [PATCH 20/29] exec: invert return value of cpu_physical_memory_get_clean, rename Paolo Bonzini
2015-05-26 11:26 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 21/29] exec: only check relevant bitmaps for cleanliness Paolo Bonzini
2015-05-26 11:31 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 22/29] memory: do not touch code dirty bitmap unless TCG is enabled Paolo Bonzini
2015-05-26 11:33 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 23/29] bitmap: add atomic set functions Paolo Bonzini
2015-05-26 11:54 ` Fam Zheng [this message]
2015-05-26 11:58 ` Paolo Bonzini
2015-05-26 12:36 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 24/29] bitmap: add atomic test and clear Paolo Bonzini
2015-05-26 12:27 ` Fam Zheng
2015-05-26 12:32 ` Fam Zheng
2015-05-26 12:33 ` Paolo Bonzini
2015-04-27 16:28 ` [Qemu-devel] [PATCH 25/29] memory: use atomic ops for setting dirty memory bits Paolo Bonzini
2015-05-26 12:29 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 26/29] migration: move dirty bitmap sync to ram_addr.h Paolo Bonzini
2015-05-26 12:34 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 27/29] memory: replace cpu_physical_memory_reset_dirty() with test-and-clear Paolo Bonzini
2015-05-26 12:42 ` Fam Zheng
2015-05-26 12:45 ` Paolo Bonzini
2015-04-27 16:28 ` [Qemu-devel] [PATCH 28/29] memory: make cpu_physical_memory_sync_dirty_bitmap() fully atomic Paolo Bonzini
2015-05-26 12:45 ` Fam Zheng
2015-04-27 16:28 ` [Qemu-devel] [PATCH 29/29] memory: strengthen assertions on mr->terminates Paolo Bonzini
2015-05-26 12:49 ` Fam Zheng
2015-05-26 12:52 ` Paolo Bonzini
2015-05-26 12:56 ` Fam Zheng
2015-05-25 21:50 ` [Qemu-devel] [PATCH v2 00/29] Dirty bitmap atomic access and optimizations Paolo Bonzini
2015-05-26 12:58 ` Fam Zheng
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=20150526115416.GV13749@ad.nay.redhat.com \
--to=famz@redhat.com \
--cc=mst@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@redhat.com \
/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).