Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging

Block pull request

 * Support -drive cache.direct=off live migration for POSIX files

# gpg: Signature made Sat 12 May 2018 10:27:51 BST
# gpg:                using RSA key 9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35  775A 9CA4 ABB3 81AB 73C8

* remotes/stefanha/tags/block-pull-request:
  block/file-posix: add x-check-page-cache=on|off option
  block/file-posix: implement bdrv_co_invalidate_cache() on Linux
  checkpatch: reduce MAINTAINERS update message frequency
  checkpatch: emit a warning on file add/move/delete
  checkpatch: ignore email headers better
  checkpatch: check utf-8 content from a commit log when it's missing from charset
  checkpatch: add a --strict check for utf-8 in commit logs
  blockjob: drop block_job_pause/resume_all()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

1 files changed, 144 insertions, 2 deletions

diff --git a/block/file-posix.c b/block/file-posix.c
index 3794c0007a..5a602cfe37 100644
--- a/block/file-posix.c
+++ b/block/file-posix.c
@@ -161,6 +161,7 @@ typedef struct BDRVRawState {
     bool page_cache_inconsistent:1;
     bool has_fallocate;
     bool needs_alignment;
+    bool check_cache_dropped;
 
     PRManager *pr_mgr;
 } BDRVRawState;
@@ -168,6 +169,7 @@ typedef struct BDRVRawState {
 typedef struct BDRVRawReopenState {
     int fd;
     int open_flags;
+    bool check_cache_dropped;
 } BDRVRawReopenState;
 
 static int fd_open(BlockDriverState *bs);
@@ -415,6 +417,11 @@ static QemuOptsList raw_runtime_opts = {
             .type = QEMU_OPT_STRING,
             .help = "id of persistent reservation manager object (default: none)",
         },
+        {
+            .name = "x-check-cache-dropped",
+            .type = QEMU_OPT_BOOL,
+            .help = "check that page cache was dropped on live migration (default: off)"
+        },
         { /* end of list */ }
     },
 };
@@ -500,6 +507,9 @@ static int raw_open_common(BlockDriverState *bs, QDict *options,
         }
     }
 
+    s->check_cache_dropped = qemu_opt_get_bool(opts, "x-check-cache-dropped",
+                                               false);
+
     s->open_flags = open_flags;
     raw_parse_flags(bdrv_flags, &s->open_flags);
 
@@ -777,6 +787,7 @@ static int raw_reopen_prepare(BDRVReopenState *state,
 {
     BDRVRawState *s;
     BDRVRawReopenState *rs;
+    QemuOpts *opts;
     int ret = 0;
     Error *local_err = NULL;
 
@@ -787,6 +798,19 @@ static int raw_reopen_prepare(BDRVReopenState *state,
 
     state->opaque = g_new0(BDRVRawReopenState, 1);
     rs = state->opaque;
+    rs->fd = -1;
+
+    /* Handle options changes */
+    opts = qemu_opts_create(&raw_runtime_opts, NULL, 0, &error_abort);
+    qemu_opts_absorb_qdict(opts, state->options, &local_err);
+    if (local_err) {
+        error_propagate(errp, local_err);
+        ret = -EINVAL;
+        goto out;
+    }
+
+    rs->check_cache_dropped = qemu_opt_get_bool(opts, "x-check-cache-dropped",
+                                                s->check_cache_dropped);
 
     if (s->type == FTYPE_CD) {
         rs->open_flags |= O_NONBLOCK;
@@ -794,8 +818,6 @@ static int raw_reopen_prepare(BDRVReopenState *state,
 
     raw_parse_flags(state->flags, &rs->open_flags);
 
-    rs->fd = -1;
-
     int fcntl_flags = O_APPEND | O_NONBLOCK;
 #ifdef O_NOATIME
     fcntl_flags |= O_NOATIME;
@@ -850,6 +872,8 @@ static int raw_reopen_prepare(BDRVReopenState *state,
         }
     }
 
+out:
+    qemu_opts_del(opts);
     return ret;
 }
 
@@ -858,6 +882,7 @@ static void raw_reopen_commit(BDRVReopenState *state)
     BDRVRawReopenState *rs = state->opaque;
     BDRVRawState *s = state->bs->opaque;
 
+    s->check_cache_dropped = rs->check_cache_dropped;
     s->open_flags = rs->open_flags;
 
     qemu_close(s->fd);
@@ -2236,6 +2261,120 @@ static int coroutine_fn raw_co_block_status(BlockDriverState *bs,
     return ret | BDRV_BLOCK_OFFSET_VALID;
 }
 
+#if defined(__linux__)
+/* Verify that the file is not in the page cache */
+static void check_cache_dropped(BlockDriverState *bs, Error **errp)
+{
+    const size_t window_size = 128 * 1024 * 1024;
+    BDRVRawState *s = bs->opaque;
+    void *window = NULL;
+    size_t length = 0;
+    unsigned char *vec;
+    size_t page_size;
+    off_t offset;
+    off_t end;
+
+    /* mincore(2) page status information requires 1 byte per page */
+    page_size = sysconf(_SC_PAGESIZE);
+    vec = g_malloc(DIV_ROUND_UP(window_size, page_size));
+
+    end = raw_getlength(bs);
+
+    for (offset = 0; offset < end; offset += window_size) {
+        void *new_window;
+        size_t new_length;
+        size_t vec_end;
+        size_t i;
+        int ret;
+
+        /* Unmap previous window if size has changed */
+        new_length = MIN(end - offset, window_size);
+        if (new_length != length) {
+            munmap(window, length);
+            window = NULL;
+            length = 0;
+        }
+
+        new_window = mmap(window, new_length, PROT_NONE, MAP_PRIVATE,
+                          s->fd, offset);
+        if (new_window == MAP_FAILED) {
+            error_setg_errno(errp, errno, "mmap failed");
+            break;
+        }
+
+        window = new_window;
+        length = new_length;
+
+        ret = mincore(window, length, vec);
+        if (ret < 0) {
+            error_setg_errno(errp, errno, "mincore failed");
+            break;
+        }
+
+        vec_end = DIV_ROUND_UP(length, page_size);
+        for (i = 0; i < vec_end; i++) {
+            if (vec[i] & 0x1) {
+                error_setg(errp, "page cache still in use!");
+                break;
+            }
+        }
+    }
+
+    if (window) {
+        munmap(window, length);
+    }
+
+    g_free(vec);
+}
+#endif /* __linux__ */
+
+static void coroutine_fn raw_co_invalidate_cache(BlockDriverState *bs,
+                                                 Error **errp)
+{
+    BDRVRawState *s = bs->opaque;
+    int ret;
+
+    ret = fd_open(bs);
+    if (ret < 0) {
+        error_setg_errno(errp, -ret, "The file descriptor is not open");
+        return;
+    }
+
+    if (s->open_flags & O_DIRECT) {
+        return; /* No host kernel page cache */
+    }
+
+#if defined(__linux__)
+    /* This sets the scene for the next syscall... */
+    ret = bdrv_co_flush(bs);
+    if (ret < 0) {
+        error_setg_errno(errp, -ret, "flush failed");
+        return;
+    }
+
+    /* Linux does not invalidate pages that are dirty, locked, or mmapped by a
+     * process.  These limitations are okay because we just fsynced the file,
+     * we don't use mmap, and the file should not be in use by other processes.
+     */
+    ret = posix_fadvise(s->fd, 0, 0, POSIX_FADV_DONTNEED);
+    if (ret != 0) { /* the return value is a positive errno */
+        error_setg_errno(errp, ret, "fadvise failed");
+        return;
+    }
+
+    if (s->check_cache_dropped) {
+        check_cache_dropped(bs, errp);
+    }
+#else /* __linux__ */
+    /* Do nothing.  Live migration to a remote host with cache.direct=off is
+     * unsupported on other host operating systems.  Cache consistency issues
+     * may occur but no error is reported here, partly because that's the
+     * historical behavior and partly because it's hard to differentiate valid
+     * configurations that should not cause errors.
+     */
+#endif /* !__linux__ */
+}
+
 static coroutine_fn BlockAIOCB *raw_aio_pdiscard(BlockDriverState *bs,
     int64_t offset, int bytes,
     BlockCompletionFunc *cb, void *opaque)
@@ -2328,6 +2467,7 @@ BlockDriver bdrv_file = {
     .bdrv_co_create_opts = raw_co_create_opts,
     .bdrv_has_zero_init = bdrv_has_zero_init_1,
     .bdrv_co_block_status = raw_co_block_status,
+    .bdrv_co_invalidate_cache = raw_co_invalidate_cache,
     .bdrv_co_pwrite_zeroes = raw_co_pwrite_zeroes,
 
     .bdrv_co_preadv         = raw_co_preadv,
@@ -2805,6 +2945,7 @@ static BlockDriver bdrv_host_device = {
     .bdrv_reopen_abort   = raw_reopen_abort,
     .bdrv_co_create_opts = hdev_co_create_opts,
     .create_opts         = &raw_create_opts,
+    .bdrv_co_invalidate_cache = raw_co_invalidate_cache,
     .bdrv_co_pwrite_zeroes = hdev_co_pwrite_zeroes,
 
     .bdrv_co_preadv         = raw_co_preadv,
@@ -2927,6 +3068,7 @@ static BlockDriver bdrv_host_cdrom = {
     .bdrv_reopen_abort   = raw_reopen_abort,
     .bdrv_co_create_opts = hdev_co_create_opts,
     .create_opts         = &raw_create_opts,
+    .bdrv_co_invalidate_cache = raw_co_invalidate_cache,
 
 
     .bdrv_co_preadv         = raw_co_preadv,