qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH 18/21] parallels: naive implementation of parallels_co_pdisca


From: Denis V. Lunev
Subject: Re: [PATCH 18/21] parallels: naive implementation of parallels_co_pdiscard
Date: Mon, 18 Sep 2023 16:05:55 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.15.1

On 9/18/23 15:57, Alexander Ivanov wrote:
On 9/15/23 20:41, Denis V. Lunev wrote:
* Discarding with backing stores is not supported by the format.
* There is no buffering/queueing of the discard operation.
* Only operations aligned to the cluster are supported.

Signed-off-by: Denis V. Lunev <den@openvz.org>
---
  block/parallels.c | 47 +++++++++++++++++++++++++++++++++++++++++++++++
  1 file changed, 47 insertions(+)

diff --git a/block/parallels.c b/block/parallels.c
index 76aedfd7c4..83cb8d6722 100644
--- a/block/parallels.c
+++ b/block/parallels.c
@@ -537,6 +537,52 @@ parallels_co_readv(BlockDriverState *bs, int64_t sector_num, int nb_sectors,
      return ret;
  }
  +
+static int coroutine_fn GRAPH_RDLOCK_PTR
+parallels_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes)
+{
+    int ret = 0;
+    uint32_t cluster, count;
+    BDRVParallelsState *s = bs->opaque;
+
+    /*
+     * The image does not support ZERO mark inside the BAT, which means that
+     * stale data could be exposed from the backing file.
+     */
+    if (bs->backing) {
+        return -ENOTSUP;
+    }
+
+    if (!QEMU_IS_ALIGNED(offset, s->cluster_size)) {
+        return -ENOTSUP;
+    } else if (!QEMU_IS_ALIGNED(bytes, s->cluster_size)) {
+        return -ENOTSUP;
+    }
+
+    cluster = offset / s->cluster_size;
+    count = bytes / s->cluster_size;
+
+    qemu_co_mutex_lock(&s->lock);
+    for (; count > 0; cluster++, count--) {
+        int64_t host_off = bat2sect(s, cluster) << BDRV_SECTOR_BITS;
+        if (host_off == 0) {
+            continue;
+        }
+
+        ret = bdrv_co_pdiscard(bs->file, cluster * s->cluster_size,
+                               s->cluster_size);
It seems, bdrv_co_pdiscard() should be called with a host offset, but there is a guest one.

correct. On top of that unit test should be modified to catch
this problem.

Den



reply via email to

[Prev in Thread] Current Thread [Next in Thread]