qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Eric Blake <eblake@redhat.com>
To: qemu-devel@nongnu.org
Cc: kwolf@redhat.com, famz@redhat.com, qemu-block@nongnu.org,
	stefanha@redhat.com, Max Reitz <mreitz@redhat.com>
Subject: [Qemu-devel] [PATCH v3 1/6] block: Fragment reads to max transfer length
Date: Fri, 15 Jul 2016 12:31:59 -0600	[thread overview]
Message-ID: <1468607524-19021-2-git-send-email-eblake@redhat.com> (raw)
In-Reply-To: <1468607524-19021-1-git-send-email-eblake@redhat.com>

Drivers should be able to rely on the block layer honoring the
max transfer length, rather than needing to return -EINVAL
(iscsi) or manually fragment things (nbd).  This patch adds
the fragmentation in the block layer, after requests have been
aligned (fragmenting before alignment would lead to multiple
unaligned requests, rather than just the head and tail).

The return value was previously nebulous on success on whether
it was zero or the length read; and fragmenting may introduce
yet other non-zero values if we use the last length read.  But
as at least some callers are sloppy and expect only zero on
success, it is easiest to just guarantee 0.

Signed-off-by: Eric Blake <eblake@redhat.com>

---
v3: Slam 'ret' to 0, not 'bytes', as some callers choke otherwise
v2: Fix uninitialized use of 'ret' for an all-zero read beyond eof
---
 block/io.c | 55 ++++++++++++++++++++++++++++++++++---------------------
 1 file changed, 34 insertions(+), 21 deletions(-)

diff --git a/block/io.c b/block/io.c
index 2887394..ceff694 100644
--- a/block/io.c
+++ b/block/io.c
@@ -971,8 +971,8 @@ err:

 /*
  * Forwards an already correctly aligned request to the BlockDriver. This
- * handles copy on read and zeroing after EOF; any other features must be
- * implemented by the caller.
+ * handles copy on read, zeroing after EOF, and fragmentation of large
+ * reads; any other features must be implemented by the caller.
  */
 static int coroutine_fn bdrv_aligned_preadv(BlockDriverState *bs,
     BdrvTrackedRequest *req, int64_t offset, unsigned int bytes,
@@ -980,12 +980,16 @@ static int coroutine_fn bdrv_aligned_preadv(BlockDriverState *bs,
 {
     int64_t total_bytes, max_bytes;
     int ret;
+    uint64_t bytes_remaining = bytes;
+    int max_transfer;

     assert(is_power_of_2(align));
     assert((offset & (align - 1)) == 0);
     assert((bytes & (align - 1)) == 0);
     assert(!qiov || bytes == qiov->size);
     assert((bs->open_flags & BDRV_O_NO_IO) == 0);
+    max_transfer = QEMU_ALIGN_DOWN(MIN_NON_ZERO(bs->bl.max_transfer, INT_MAX),
+                                   align);

     /* TODO: We would need a per-BDS .supported_read_flags and
      * potential fallback support, if we ever implement any read flags
@@ -1024,7 +1028,7 @@ static int coroutine_fn bdrv_aligned_preadv(BlockDriverState *bs,
         }
     }

-    /* Forward the request to the BlockDriver */
+    /* Forward the request to the BlockDriver, possibly fragmenting it */
     total_bytes = bdrv_getlength(bs);
     if (total_bytes < 0) {
         ret = total_bytes;
@@ -1032,30 +1036,39 @@ static int coroutine_fn bdrv_aligned_preadv(BlockDriverState *bs,
     }

     max_bytes = ROUND_UP(MAX(0, total_bytes - offset), align);
-    if (bytes <= max_bytes) {
+    if (bytes <= max_bytes && bytes <= max_transfer) {
         ret = bdrv_driver_preadv(bs, offset, bytes, qiov, 0);
-    } else if (max_bytes > 0) {
-        QEMUIOVector local_qiov;
-
-        qemu_iovec_init(&local_qiov, qiov->niov);
-        qemu_iovec_concat(&local_qiov, qiov, 0, max_bytes);
-
-        ret = bdrv_driver_preadv(bs, offset, max_bytes, &local_qiov, 0);
-
-        qemu_iovec_destroy(&local_qiov);
-    } else {
-        ret = 0;
+        goto out;
     }

-    /* Reading beyond end of file is supposed to produce zeroes */
-    if (ret == 0 && total_bytes < offset + bytes) {
-        uint64_t zero_offset = MAX(0, total_bytes - offset);
-        uint64_t zero_bytes = offset + bytes - zero_offset;
-        qemu_iovec_memset(qiov, zero_offset, 0, zero_bytes);
+    while (bytes_remaining) {
+        int num;
+
+        if (max_bytes) {
+            QEMUIOVector local_qiov;
+
+            num = MIN(bytes_remaining, MIN(max_bytes, max_transfer));
+            assert(num);
+            qemu_iovec_init(&local_qiov, qiov->niov);
+            qemu_iovec_concat(&local_qiov, qiov, bytes - bytes_remaining, num);
+
+            ret = bdrv_driver_preadv(bs, offset + bytes - bytes_remaining,
+                                     num, &local_qiov, 0);
+            max_bytes -= num;
+            qemu_iovec_destroy(&local_qiov);
+        } else {
+            num = bytes_remaining;
+            ret = qemu_iovec_memset(qiov, bytes - bytes_remaining, 0,
+                                    bytes_remaining);
+        }
+        if (ret < 0) {
+            goto out;
+        }
+        bytes_remaining -= num;
     }

 out:
-    return ret;
+    return ret < 0 ? ret : 0;
 }

 /*
-- 
2.5.5

  reply	other threads:[~2016-07-15 18:32 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-07-15 18:31 [Qemu-devel] [PATCH v3 0/6] Auto-fragment large transactions at the block layer Eric Blake
2016-07-15 18:31 ` Eric Blake [this message]
2016-07-15 18:32 ` [Qemu-devel] [PATCH v3 2/6] raw_bsd: Don't advertise flags not supported by protocol layer Eric Blake
2016-07-15 18:32 ` [Qemu-devel] [PATCH v3 3/6] block: Fragment writes to max transfer length Eric Blake
2016-07-15 18:32 ` [Qemu-devel] [PATCH v3 4/6] nbd: Rely on block layer to break up large requests Eric Blake
2016-07-18  8:16   ` Paolo Bonzini
2016-07-15 18:32 ` [Qemu-devel] [PATCH v3 5/6] nbd: Drop unused offset parameter Eric Blake
2016-07-18  8:16   ` Paolo Bonzini
2016-07-15 18:32 ` [Qemu-devel] [PATCH v3 6/6] iscsi: Rely on block layer to break up large requests Eric Blake
2016-07-18  8:16   ` Paolo Bonzini
2016-07-19 16:12 ` [Qemu-devel] [PATCH v3 0/6] Auto-fragment large transactions at the block layer Stefan Hajnoczi

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=1468607524-19021-2-git-send-email-eblake@redhat.com \
    --to=eblake@redhat.com \
    --cc=famz@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=mreitz@redhat.com \
    --cc=qemu-block@nongnu.org \
    --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).