This removes some hardcoded 0x200 making them far clearer.
This also renames some BDRV_SECTOR_* constants that were introduced
during the transitions for sector-based to bytes-based block interfaces.
While they have the same values, the BDRV_SECTOR_* constants are
unrelated to the VVFAT sectors.
Signed-off-by: Clément Chigot <chigot@adacore.com>
---
block/vvfat.c | 119 ++++++++++++++++++++++++++++----------------------
1 file changed, 66 insertions(+), 53 deletions(-)
diff --git a/block/vvfat.c b/block/vvfat.c
index d8c8d44f16..1c51dfa561 100644
--- a/block/vvfat.c
+++ b/block/vvfat.c
@@ -75,6 +75,9 @@ static void checkpoint(void);
*/
#define BOOTSECTOR_OEM_NAME "MSWIN4.1"
+#define VVFAT_SECTOR_BITS 9
+#define VVFAT_SECTOR_SIZE (1ULL << VVFAT_SECTOR_BITS)
+
#define DIR_DELETED 0xe5
#define DIR_KANJI DIR_DELETED
#define DIR_KANJI_FAKE 0x05
@@ -300,7 +303,7 @@ static void print_mapping(const struct mapping_t* mapping);
typedef struct BDRVVVFATState {
CoMutex lock;
BlockDriverState* bs; /* pointer to parent */
- unsigned char first_sectors[0x40*0x200];
+ unsigned char first_sectors[0x40 * VVFAT_SECTOR_SIZE];
int fat_type; /* 16 or 32 */
array_t fat,directory,mapping;
@@ -690,11 +693,11 @@ static inline void init_fat(BDRVVVFATState* s)
if (s->fat_type == 12) {
array_init(&(s->fat),1);
array_ensure_allocated(&(s->fat),
- s->sectors_per_fat * 0x200 * 3 / 2 - 1);
+ s->sectors_per_fat * VVFAT_SECTOR_SIZE * 3 / 2 - 1);
} else {
array_init(&(s->fat),(s->fat_type==32?4:2));
array_ensure_allocated(&(s->fat),
- s->sectors_per_fat * 0x200 / s->fat.item_size - 1);
+ s->sectors_per_fat * VVFAT_SECTOR_SIZE / s->fat.item_size - 1);
}
memset(s->fat.pointer,0,s->fat.size);
@@ -902,19 +905,19 @@ static int init_directories(BDRVVVFATState* s,
unsigned int i;
unsigned int cluster;
- memset(&(s->first_sectors[0]),0,0x40*0x200);
+ memset(&(s->first_sectors[0]), 0 , 0x40 * VVFAT_SECTOR_SIZE);
- s->cluster_size=s->sectors_per_cluster*0x200;
+ s->cluster_size = s->sectors_per_cluster * VVFAT_SECTOR_SIZE;
s->cluster_buffer=g_malloc(s->cluster_size);
/*
- * The formula: sc = spf+1+spf*spc*(512*8/fat_type),
+ * The formula: sc = spf+1+spf*spc*(VVFAT_SECTOR_SIZE*8/fat_type),
* where sc is sector_count,
* spf is sectors_per_fat,
* spc is sectors_per_clusters, and
* fat_type = 12, 16 or 32.
*/
- i = 1+s->sectors_per_cluster*0x200*8/s->fat_type;
+ i = 1 + s->sectors_per_cluster * VVFAT_SECTOR_SIZE * 8 / s->fat_type;
s->sectors_per_fat=(s->sector_count+i)/i; /* round up */
s->offset_to_fat = s->offset_to_bootsector + 1;
@@ -1012,12 +1015,13 @@ static int init_directories(BDRVVVFATState* s,
s->current_mapping = NULL;
bootsector = (bootsector_t *)(s->first_sectors
- + s->offset_to_bootsector * 0x200);
+ + s->offset_to_bootsector
+ * VVFAT_SECTOR_SIZE);
bootsector->jump[0]=0xeb;
bootsector->jump[1]=0x3e;
bootsector->jump[2]=0x90;
memcpy(bootsector->name, BOOTSECTOR_OEM_NAME, 8);
- bootsector->sector_size=cpu_to_le16(0x200);
+ bootsector->sector_size = cpu_to_le16(VVFAT_SECTOR_SIZE);
bootsector->sectors_per_cluster=s->sectors_per_cluster;
bootsector->reserved_sectors=cpu_to_le16(1);
bootsector->number_of_fats=0x2; /* number of FATs */
@@ -1313,7 +1317,7 @@ fail:
static void vvfat_refresh_limits(BlockDriverState *bs, Error **errp)
{
- bs->bl.request_alignment = BDRV_SECTOR_SIZE; /* No sub-sector I/O */
+ bs->bl.request_alignment = VVFAT_SECTOR_SIZE; /* No sub-sector I/O */
}
static inline void vvfat_close_current_file(BDRVVVFATState *s)
@@ -1499,21 +1503,23 @@ vvfat_read(BlockDriverState *bs, int64_t sector_num, uint8_t *buf, int nb_sector
if (s->qcow) {
int64_t n;
int ret;
- ret = bdrv_co_is_allocated(s->qcow->bs, sector_num * BDRV_SECTOR_SIZE,
- (nb_sectors - i) * BDRV_SECTOR_SIZE, &n);
+ ret = bdrv_co_is_allocated(s->qcow->bs,
+ sector_num * VVFAT_SECTOR_SIZE,
+ (nb_sectors - i) * VVFAT_SECTOR_SIZE,
+ &n);
if (ret < 0) {
return ret;
}
if (ret) {
DLOG(fprintf(stderr, "sectors %" PRId64 "+%" PRId64
" allocated\n", sector_num,
- n >> BDRV_SECTOR_BITS));
- if (bdrv_co_pread(s->qcow, sector_num * BDRV_SECTOR_SIZE, n,
- buf + i * 0x200, 0) < 0) {
+ n >> VVFAT_SECTOR_BITS));
+ if (bdrv_co_pread(s->qcow, sector_num * VVFAT_SECTOR_SIZE,
+ n, buf + i * VVFAT_SECTOR_SIZE, 0) < 0) {
return -1;
}
- i += (n >> BDRV_SECTOR_BITS) - 1;
- sector_num += (n >> BDRV_SECTOR_BITS) - 1;
+ i += (n >> VVFAT_SECTOR_BITS) - 1;
+ sector_num += (n >> VVFAT_SECTOR_BITS) - 1;
continue;
}
DLOG(fprintf(stderr, "sector %" PRId64 " not allocated\n",
@@ -1521,19 +1527,20 @@ vvfat_read(BlockDriverState *bs, int64_t sector_num, uint8_t *buf, int nb_sector
}
if (sector_num < s->offset_to_root_dir) {
if (sector_num < s->offset_to_fat) {
- memcpy(buf + i * 0x200,
- &(s->first_sectors[sector_num * 0x200]),
- 0x200);
+ memcpy(buf + i * VVFAT_SECTOR_SIZE,
+ &(s->first_sectors[sector_num * VVFAT_SECTOR_SIZE]),
+ VVFAT_SECTOR_SIZE);
} else if (sector_num < s->offset_to_fat + s->sectors_per_fat) {
- memcpy(buf + i * 0x200,
- &(s->fat.pointer[(sector_num
- - s->offset_to_fat) * 0x200]),
- 0x200);
+ memcpy(buf + i * VVFAT_SECTOR_SIZE,
+ &(s->fat.pointer[(sector_num - s->offset_to_fat)
+ * VVFAT_SECTOR_SIZE]),
+ VVFAT_SECTOR_SIZE);
} else if (sector_num < s->offset_to_root_dir) {
- memcpy(buf + i * 0x200,
+ memcpy(buf + i * VVFAT_SECTOR_SIZE,
&(s->fat.pointer[(sector_num - s->offset_to_fat
- - s->sectors_per_fat) * 0x200]),
- 0x200);
+ - s->sectors_per_fat)
+ * VVFAT_SECTOR_SIZE]),
+ VVFAT_SECTOR_SIZE);
}
} else {
uint32_t sector = sector_num - s->offset_to_root_dir,
@@ -1541,10 +1548,12 @@ vvfat_read(BlockDriverState *bs, int64_t sector_num, uint8_t *buf, int nb_sector
cluster_num=sector/s->sectors_per_cluster;
if(cluster_num > s->cluster_count || read_cluster(s, cluster_num) != 0) {
/* LATER TODO: strict: return -1; */
- memset(buf+i*0x200,0,0x200);
+ memset(buf + i * VVFAT_SECTOR_SIZE, 0, VVFAT_SECTOR_SIZE);
continue;
}
- memcpy(buf+i*0x200,s->cluster+sector_offset_in_cluster*0x200,0x200);
+ memcpy(buf + i * VVFAT_SECTOR_SIZE,
+ s->cluster + sector_offset_in_cluster * VVFAT_SECTOR_SIZE,
+ VVFAT_SECTOR_SIZE);
}
}
return 0;
@@ -1556,12 +1565,12 @@ vvfat_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes,
{
int ret;
BDRVVVFATState *s = bs->opaque;
- uint64_t sector_num = offset >> BDRV_SECTOR_BITS;
- int nb_sectors = bytes >> BDRV_SECTOR_BITS;
+ uint64_t sector_num = offset >> VVFAT_SECTOR_BITS;
+ int nb_sectors = bytes >> VVFAT_SECTOR_BITS;
void *buf;
- assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
- assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
+ assert(QEMU_IS_ALIGNED(offset, VVFAT_SECTOR_SIZE));
+ assert(QEMU_IS_ALIGNED(bytes, VVFAT_SECTOR_SIZE));
buf = g_try_malloc(bytes);
if (bytes && buf == NULL) {
@@ -1827,8 +1836,8 @@ cluster_was_modified(BDRVVVFATState *s, uint32_t cluster_num)
for (i = 0; !was_modified && i < s->sectors_per_cluster; i++) {
was_modified = bdrv_co_is_allocated(s->qcow->bs,
(cluster2sector(s, cluster_num) +
- i) * BDRV_SECTOR_SIZE,
- BDRV_SECTOR_SIZE, NULL);
+ i) * VVFAT_SECTOR_SIZE,
+ VVFAT_SECTOR_SIZE, NULL);
}
/*
@@ -1982,8 +1991,8 @@ get_cluster_count_for_direntry(BDRVVVFATState* s, direntry_t* direntry, const ch
int res;
res = bdrv_co_is_allocated(s->qcow->bs,
- (offs + i) * BDRV_SECTOR_SIZE,
- BDRV_SECTOR_SIZE, NULL);
+ (offs + i) * VVFAT_SECTOR_SIZE,
+ VVFAT_SECTOR_SIZE, NULL);
if (res < 0) {
return -1;
}
@@ -1992,9 +2001,9 @@ get_cluster_count_for_direntry(BDRVVVFATState* s, direntry_t* direntry, const ch
if (res) {
return -1;
}
- res = bdrv_co_pwrite(s->qcow, offs * BDRV_SECTOR_SIZE,
- BDRV_SECTOR_SIZE, s->cluster_buffer,
- 0);
+ res = bdrv_co_pwrite(s->qcow, offs * VVFAT_SECTOR_SIZE,
+ VVFAT_SECTOR_SIZE,
+ s->cluster_buffer, 0);
if (res < 0) {
return -2;
}
@@ -2172,7 +2181,7 @@ DLOG(checkpoint());
* - if all is fine, return number of used clusters
*/
if (s->fat2 == NULL) {
- int size = 0x200 * s->sectors_per_fat;
+ int size = VVFAT_SECTOR_SIZE * s->sectors_per_fat;
s->fat2 = g_malloc(size);
memcpy(s->fat2, s->fat.pointer, size);
}
@@ -2569,7 +2578,8 @@ commit_one_file(BDRVVVFATState* s, int dir_index, uint32_t offset)
(size > offset && c >=2 && !fat_eof(s, c)));
ret = vvfat_read(s->bs, cluster2sector(s, c),
- (uint8_t*)cluster, DIV_ROUND_UP(rest_size, 0x200));
+ (uint8_t *)cluster,
+ DIV_ROUND_UP(rest_size, VVFAT_SECTOR_SIZE));
if (ret < 0) {
qemu_close(fd);
@@ -2948,7 +2958,7 @@ static int coroutine_fn GRAPH_RDLOCK do_commit(BDRVVVFATState* s)
}
/* copy FAT (with bdrv_pread) */
- memcpy(s->fat.pointer, s->fat2, 0x200 * s->sectors_per_fat);
+ memcpy(s->fat.pointer, s->fat2, VVFAT_SECTOR_SIZE * s->sectors_per_fat);
/* recurse direntries from root (using bs->bdrv_pread) */
ret = commit_direntries(s, 0, -1);
@@ -3012,14 +3022,15 @@ DLOG(checkpoint());
* used to mark volume dirtiness
*/
unsigned char *bootsector = s->first_sectors
- + s->offset_to_bootsector * 0x200;
+ + s->offset_to_bootsector
+ * VVFAT_SECTOR_SIZE;
/*
* LATER TODO: if FAT32, this is wrong (see init_directories(),
* which always creates a FAT16 bootsector)
*/
const int reserved1_offset = offsetof(bootsector_t, u.fat16.reserved1);
- for (i = 0; i < 0x200; i++) {
+ for (i = 0; i < VVFAT_SECTOR_SIZE; i++) {
if (i != reserved1_offset && bootsector[i] != buf[i]) {
fprintf(stderr, "Tried to write to protected bootsector\n");
return -1;
@@ -3074,7 +3085,9 @@ DLOG(checkpoint());
end = sector_num + nb_sectors;
dir_index = mapping->dir_index +
0x10 * (begin - mapping->begin * s->sectors_per_cluster);
- direntries = (direntry_t*)(buf + 0x200 * (begin - sector_num));
+ direntries =
+ (direntry_t *)(buf + VVFAT_SECTOR_SIZE
+ * (begin - sector_num));
for (k = 0; k < (end - begin) * 0x10; k++) {
/* no access to the direntry of a read-only file */
@@ -3100,8 +3113,8 @@ DLOG(checkpoint());
* Use qcow backend. Commit later.
*/
DLOG(fprintf(stderr, "Write to qcow backend: %d + %d\n", (int)sector_num, nb_sectors));
- ret = bdrv_co_pwrite(s->qcow, sector_num * BDRV_SECTOR_SIZE,
- nb_sectors * BDRV_SECTOR_SIZE, buf, 0);
+ ret = bdrv_co_pwrite(s->qcow, sector_num * VVFAT_SECTOR_SIZE,
+ nb_sectors * VVFAT_SECTOR_SIZE, buf, 0);
if (ret < 0) {
fprintf(stderr, "Error writing to qcow backend\n");
return ret;
@@ -3127,12 +3140,12 @@ vvfat_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes,
{
int ret;
BDRVVVFATState *s = bs->opaque;
- uint64_t sector_num = offset >> BDRV_SECTOR_BITS;
- int nb_sectors = bytes >> BDRV_SECTOR_BITS;
+ uint64_t sector_num = offset >> VVFAT_SECTOR_BITS;
+ int nb_sectors = bytes >> VVFAT_SECTOR_BITS;
void *buf;
- assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
- assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
+ assert(QEMU_IS_ALIGNED(offset, VVFAT_SECTOR_SIZE));
+ assert(QEMU_IS_ALIGNED(bytes, VVFAT_SECTOR_SIZE));
buf = g_try_malloc(bytes);
if (bytes && buf == NULL) {
@@ -3198,7 +3211,7 @@ static int enable_write_target(BlockDriverState *bs, Error **errp)
opts = qemu_opts_create(bdrv_qcow->create_opts, NULL, 0, &error_abort);
qemu_opt_set_number(opts, BLOCK_OPT_SIZE,
- bs->total_sectors * BDRV_SECTOR_SIZE, &error_abort);
+ bs->total_sectors * VVFAT_SECTOR_SIZE, &error_abort);
qemu_opt_set(opts, BLOCK_OPT_BACKING_FILE, "fat:", &error_abort);
ret = bdrv_create(bdrv_qcow, s->qcow_filename, opts, errp);
--
2.43.0
© 2016 - 2025 Red Hat, Inc.