qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Philippe Mathieu-Daudé" <philmd@redhat.com>
To: qemu-devel@nongnu.org
Cc: "Peter Maydell" <peter.maydell@linaro.org>,
	"Michael S. Tsirkin" <mst@redhat.com>,
	"Philippe Mathieu-Daudé" <philmd@redhat.com>,
	"Huacai Chen" <chenhuacai@kernel.org>,
	"Richard Henderson" <richard.henderson@linaro.org>,
	"Peter Xu" <peterx@redhat.com>,
	"Stefan Hajnoczi" <stefanha@redhat.com>,
	"Paolo Bonzini" <pbonzini@redhat.com>,
	"Bibo Mao" <maobibo@loongson.cn>
Subject: [RFC PATCH 25/25] hw/virtio: Optimize accesses on vring aligned pointers
Date: Tue, 18 May 2021 20:36:55 +0200	[thread overview]
Message-ID: <20210518183655.1711377-26-philmd@redhat.com> (raw)
In-Reply-To: <20210518183655.1711377-1-philmd@redhat.com>

By the virtio spec [*] the vring is aligned, so the 'idx' and 'flag'
fields are also 16-bit aligned. Therefore we can use the load/store
*aligned* API to set the value.

[*] https://docs.oasis-open.org/virtio/virtio/v1.1/csprd01/virtio-v1.1-csprd01.html#x1-430008

Reported-by: Bibo Mao <maobibo@loongson.cn>
Inspired-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 hw/virtio/virtio.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index e02544b2df7..ebee9c4e643 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -296,7 +296,7 @@ static inline uint16_t vring_avail_flags(VirtQueue *vq)
         return 0;
     }
 
-    return virtio_lduw_phys_cached(vq->vdev, &caches->avail, pa);
+    return virtio_lduw_phys_cached_aligned(vq->vdev, &caches->avail, pa);
 }
 
 /* Called within rcu_read_lock().  */
@@ -309,7 +309,8 @@ static inline uint16_t vring_avail_idx(VirtQueue *vq)
         return 0;
     }
 
-    vq->shadow_avail_idx = virtio_lduw_phys_cached(vq->vdev, &caches->avail, pa);
+    vq->shadow_avail_idx = virtio_lduw_phys_cached_aligned(vq->vdev,
+                                                           &caches->avail, pa);
     return vq->shadow_avail_idx;
 }
 
@@ -359,7 +360,7 @@ static uint16_t vring_used_idx(VirtQueue *vq)
         return 0;
     }
 
-    return virtio_lduw_phys_cached(vq->vdev, &caches->used, pa);
+    return virtio_lduw_phys_cached_aligned(vq->vdev, &caches->used, pa);
 }
 
 /* Called within rcu_read_lock().  */
@@ -369,7 +370,7 @@ static inline void vring_used_idx_set(VirtQueue *vq, uint16_t val)
     hwaddr pa = offsetof(VRingUsed, idx);
 
     if (caches) {
-        virtio_stw_phys_cached(vq->vdev, &caches->used, pa, val);
+        virtio_stw_phys_cached_aligned(vq->vdev, &caches->used, pa, val);
         address_space_cache_invalidate(&caches->used, pa, sizeof(val));
     }
 
@@ -388,8 +389,8 @@ static inline void vring_used_flags_set_bit(VirtQueue *vq, int mask)
         return;
     }
 
-    flags = virtio_lduw_phys_cached(vq->vdev, &caches->used, pa);
-    virtio_stw_phys_cached(vdev, &caches->used, pa, flags | mask);
+    flags = virtio_lduw_phys_cached_aligned(vq->vdev, &caches->used, pa);
+    virtio_stw_phys_cached_aligned(vdev, &caches->used, pa, flags | mask);
     address_space_cache_invalidate(&caches->used, pa, sizeof(flags));
 }
 
-- 
2.26.3



  parent reply	other threads:[~2021-05-18 18:57 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-18 18:36 [RFC PATCH 00/25] exec: Add load/store API for aligned pointers Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 01/25] exec/memory_ldst_cached: Sort declarations Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 02/25] exec/memory_ldst_phys: " Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 03/25] exec/memory_ldst: Use correct type sizes Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 04/25] exec/memory_ldst_phys: " Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 05/25] exec/memory_ldst_cached: Use correct type size Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 06/25] exec/memory: " Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 07/25] qemu/bswap: Introduce ST_CONVERT() macro Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 08/25] qemu/bswap: Use ST_CONVERT() macro to emit 16-bit load/store functions Philippe Mathieu-Daudé
2021-05-18 20:08   ` Peter Maydell
2021-05-19 17:30     ` Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 09/25] qemu/bswap: Introduce LD_CONVERT() macro Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 10/25] qemu/bswap: Use LD_CONVERT macro to emit 16-bit signed load/store code Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 11/25] qemu/bswap: Use LD_CONVERT macro to emit 16-bit unsigned " Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 12/25] qemu/bswap: Use LDST_CONVERT macro to emit 32-bit load/store functions Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 13/25] qemu/bswap: Use LDST_CONVERT macro to emit 64-bit " Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 14/25] qemu/bswap: Introduce load/store for aligned pointer Philippe Mathieu-Daudé
2021-05-18 20:15   ` Peter Maydell
2021-05-19  5:56     ` Philippe Mathieu-Daudé
2021-05-19  9:08       ` Stefan Hajnoczi
2021-05-18 18:36 ` [RFC PATCH 15/25] exec/memory: Add methods for aligned pointer access (address space) Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 16/25] exec/memory: Add methods for aligned pointer access (physical memory) Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 17/25] hw/virtio: Use correct type sizes Philippe Mathieu-Daudé
2021-05-20 16:27   ` Richard Henderson
2021-05-20 19:13     ` Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 18/25] hw/virtio: Introduce VIRTIO_LD_CONVERT() macro Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 19/25] hw/virtio: Use LD_CONVERT macro to emit 16-bit unsigned load/store code Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 20/25] hw/virtio: Introduce VIRTIO_ST_CONVERT() macro Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 21/25] hw/virtio: Use ST_CONVERT() macro to emit 16-bit load/store functions Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 22/25] hw/virtio: Use LDST_CONVERT macro to emit 32-bit " Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 23/25] hw/virtio: Use LDST_CONVERT macro to emit 64-bit " Philippe Mathieu-Daudé
2021-05-18 18:36 ` [RFC PATCH 24/25] hw/virtio: Add methods for aligned pointer access Philippe Mathieu-Daudé
2021-05-18 18:36 ` Philippe Mathieu-Daudé [this message]
2021-05-18 18:42 ` [RFC PATCH 00/25] exec: Add load/store API for aligned pointers no-reply
2021-05-19 19:20 ` Richard Henderson
2021-05-19 19:40   ` Philippe Mathieu-Daudé
2021-05-20 16:28     ` Richard Henderson

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=20210518183655.1711377-26-philmd@redhat.com \
    --to=philmd@redhat.com \
    --cc=chenhuacai@kernel.org \
    --cc=maobibo@loongson.cn \
    --cc=mst@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=peter.maydell@linaro.org \
    --cc=peterx@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=richard.henderson@linaro.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).