block: add a CoMutex to synchronous read drivers

The big conversion of bdrv_read/write to coroutines caused the two
homonymous callbacks in BlockDriver to become reentrant.  It goes
like this:

1) bdrv_read is now called in a coroutine, and calls bdrv_read or
bdrv_pread.

2) the nested bdrv_read goes through the fast path in bdrv_rw_co_entry;

3) in the common case when the protocol is file, bdrv_co_do_readv calls
bdrv_co_readv_em (and from here goes to bdrv_co_io_em), which yields
until the AIO operation is complete;

4) if bdrv_read had been called from a bottom half, the main loop
is free to iterate again: a device model or another bottom half
can then come and call bdrv_read again.

This applies to all four of read/write/flush/discard.  It would also
apply to is_allocated, but it is not used from within coroutines:
besides qemu-img.c and qemu-io.c, which operate synchronously, the
only user is the monitor.  Copy-on-read will introduce a use in the
block layer, and will require converting it.

The solution is "simply" to convert all drivers to coroutines!  We
just need to add a CoMutex that is taken around affected operations.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block/bochs.c

@@ -80,6 +80,7 @@ struct bochs_header {
 };
 
 typedef struct BDRVBochsState {
+    CoMutex lock;
     uint32_t *catalog_bitmap;
     int catalog_size;
 
@@ -150,6 +151,7 @@ static int bochs_open(BlockDriverState *bs, int flags)
 
     s->extent_size = le32_to_cpu(bochs.extra.redolog.extent);
 
+    qemu_co_mutex_init(&s->lock);
     return 0;
  fail:
     return -1;

block/cloop.c

@@ -27,6 +27,7 @@
 #include <zlib.h>
 
 typedef struct BDRVCloopState {
+    CoMutex lock;
     uint32_t block_size;
     uint32_t n_blocks;
     uint64_t* offsets;
@@ -93,6 +94,7 @@ static int cloop_open(BlockDriverState *bs, int flags)
 
     s->sectors_per_block = s->block_size/512;
     bs->total_sectors = s->n_blocks*s->sectors_per_block;
+    qemu_co_mutex_init(&s->lock);
     return 0;
 
 cloop_close:

block/cow.c

@@ -42,6 +42,7 @@ struct cow_header_v2 {
 };
 
 typedef struct BDRVCowState {
+    CoMutex lock;
     int64_t cow_sectors_offset;
 } BDRVCowState;
 
@@ -84,6 +85,7 @@ static int cow_open(BlockDriverState *bs, int flags)
 
     bitmap_size = ((bs->total_sectors + 7) >> 3) + sizeof(cow_header);
     s->cow_sectors_offset = (bitmap_size + 511) & ~511;
+    qemu_co_mutex_init(&s->lock);
     return 0;
  fail:
     return -1;

block/dmg.c

@@ -28,6 +28,7 @@
 #include <zlib.h>
 
 typedef struct BDRVDMGState {
+    CoMutex lock;
     /* each chunk contains a certain number of sectors,
      * offsets[i] is the offset in the .dmg file,
      * lengths[i] is the length of the compressed chunk,
@@ -177,6 +178,7 @@ static int dmg_open(BlockDriverState *bs, int flags)
 
     s->current_chunk = s->n_chunks;
 
+    qemu_co_mutex_init(&s->lock);
     return 0;
 fail:
     return -1;

block/nbd.c

@@ -47,6 +47,7 @@
 #endif
 
 typedef struct BDRVNBDState {
+    CoMutex lock;
     int sock;
     uint32_t nbdflags;
     off_t size;
@@ -175,6 +176,7 @@ static int nbd_open(BlockDriverState *bs, const char* filename, int flags)
      */
     result = nbd_establish_connection(bs);
 
+    qemu_co_mutex_init(&s->lock);
     return result;
 }
 

block/parallels.c

@@ -46,6 +46,7 @@ struct parallels_header {
 } QEMU_PACKED;
 
 typedef struct BDRVParallelsState {
+    CoMutex lock;
 
     uint32_t *catalog_bitmap;
     int catalog_size;
@@ -95,6 +96,7 @@ static int parallels_open(BlockDriverState *bs, int flags)
     for (i = 0; i < s->catalog_size; i++)
 	le32_to_cpus(&s->catalog_bitmap[i]);
 
+    qemu_co_mutex_init(&s->lock);
     return 0;
 fail:
     if (s->catalog_bitmap)

block/vmdk.c

@@ -90,6 +90,7 @@ typedef struct VmdkExtent {
 } VmdkExtent;
 
 typedef struct BDRVVmdkState {
+    CoMutex lock;
     int desc_offset;
     bool cid_updated;
     uint32_t parent_cid;
@@ -646,6 +647,7 @@ static int vmdk_open(BlockDriverState *bs, int flags)
         goto fail;
     }
     s->parent_cid = vmdk_read_cid(bs, 1);
+    qemu_co_mutex_init(&s->lock);
     return ret;
 
 fail:

block/vpc.c

@@ -110,6 +110,7 @@ struct vhd_dyndisk_header {
 };
 
 typedef struct BDRVVPCState {
+    CoMutex lock;
     uint8_t footer_buf[HEADER_SIZE];
     uint64_t free_data_block_offset;
     int max_table_entries;
@@ -226,6 +227,7 @@ static int vpc_open(BlockDriverState *bs, int flags)
     s->last_pagetable = -1;
 #endif
 
+    qemu_co_mutex_init(&s->lock);
     return 0;
  fail:
     return err;

block/vvfat.c

@@ -317,6 +317,7 @@ static void print_mapping(const struct mapping_t* mapping);
 /* here begins the real VVFAT driver */
 
 typedef struct BDRVVVFATState {
+    CoMutex lock;
     BlockDriverState* bs; /* pointer to parent */
     unsigned int first_sectors_number; /* 1 for a single partition, 0x40 for a disk with partition table */
     unsigned char first_sectors[0x40*0x200];
@@ -1065,6 +1066,7 @@ DLOG(if (stderr == NULL) {
 	bs->heads = bs->cyls = bs->secs = 0;
 
     //    assert(is_consistent(s));
+    qemu_co_mutex_init(&s->lock);
     return 0;
 }