* [PATCH v3 2/5] erofs-utils: dump: add feature for collecting filesystem statistics.
2021-10-29 9:12 [PATCH v3 1/5] erofs-utils: add support for the full decompressed length Guo Xuenan
@ 2021-10-29 9:12 ` Guo Xuenan
2021-10-29 9:12 ` [PATCH v3 3/5] erofs-utils: dump: add option to print specified file infomation Guo Xuenan
` (3 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Guo Xuenan @ 2021-10-29 9:12 UTC (permalink / raw)
To: xiang, linux-erofs; +Cc: mpiglet
From: Wang Qi <mpiglet@outlook.com>
Add -S option, for printing statistics of the overall disk, including
file type(by file extension)/size statistics and distribution, number
of compressed and uncompressed files, whole compression ratio of image.
Signed-off-by: Guo Xuenan <guoxuenan@huawei.com>
Signed-off-by: Wang Qi <mpiglet@outlook.com>
---
dump/Makefile.am | 2 +-
dump/main.c | 357 ++++++++++++++++++++++++++++++++++++++-
include/erofs/internal.h | 1 +
lib/namei.c | 2 +-
4 files changed, 359 insertions(+), 3 deletions(-)
diff --git a/dump/Makefile.am b/dump/Makefile.am
index f0246d7..4759901 100644
--- a/dump/Makefile.am
+++ b/dump/Makefile.am
@@ -6,4 +6,4 @@ bin_PROGRAMS = dump.erofs
AM_CPPFLAGS = ${libuuid_CFLAGS}
dump_erofs_SOURCES = main.c
dump_erofs_CFLAGS = -Wall -Werror -I$(top_srcdir)/include
-dump_erofs_LDADD = $(top_builddir)/lib/liberofs.la ${libselinux_LIBS} ${libuuid_LIBS}
+dump_erofs_LDADD = $(top_builddir)/lib/liberofs.la ${libselinux_LIBS} ${libuuid_LIBS} ${liblz4_LIBS}
diff --git a/dump/main.c b/dump/main.c
index 5b7ac5c..c5e82d1 100644
--- a/dump/main.c
+++ b/dump/main.c
@@ -18,9 +18,51 @@
struct erofsdump_cfg {
unsigned int totalshow;
bool show_superblock;
+ bool show_statistics;
};
static struct erofsdump_cfg dumpcfg;
+static const char chart_format[] = "%-16s %-11d %8.2f%% |%-50s|\n";
+static const char header_format[] = "%-16s %11s %16s |%-50s|\n";
+static char *file_types[] = {
+ ".txt", ".so", ".xml", ".apk",
+ ".odex", ".vdex", ".oat", ".rc",
+ ".otf", ".txt", "others",
+};
+#define OTHERFILETYPE ARRAY_SIZE(file_types)
+/* (1 << FILE_MAX_SIZE_BITS)KB */
+#define FILE_MAX_SIZE_BITS 16
+
+static const char * const file_category_types[] = {
+ [EROFS_FT_UNKNOWN] = "unknown type",
+ [EROFS_FT_REG_FILE] = "regular file",
+ [EROFS_FT_DIR] = "directory",
+ [EROFS_FT_CHRDEV] = "char dev",
+ [EROFS_FT_BLKDEV] = "block dev",
+ [EROFS_FT_FIFO] = "FIFO file",
+ [EROFS_FT_SOCK] = "SOCK file",
+ [EROFS_FT_SYMLINK] = "symlink file",
+};
+
+struct erofs_statistics {
+ unsigned long files;
+ unsigned long compressed_files;
+ unsigned long uncompressed_files;
+ unsigned long files_total_size;
+ unsigned long files_total_origin_size;
+ double compress_rate;
+
+ /* [statistics] # of files based on inode_info->flags */
+ unsigned long file_category_stat[EROFS_FT_MAX];
+ /* [statistics] # of files based on file name extensions */
+ unsigned int file_type_stat[OTHERFILETYPE];
+ /* [statistics] # of files based on the original size of files */
+ unsigned int file_original_size[FILE_MAX_SIZE_BITS + 1];
+ /* [statistics] # of files based on the compressed size of files */
+ unsigned int file_comp_size[FILE_MAX_SIZE_BITS + 1];
+};
+static struct erofs_statistics stats;
+
static struct option long_options[] = {
{"help", no_argument, 0, 1},
{0, 0, 0, 0},
@@ -39,10 +81,13 @@ static struct erofsdump_feature feature_lists[] = {
{ false, EROFS_FEATURE_INCOMPAT_CHUNKED_FILE, "chunked_file" },
};
+static int erofs_read_dir(erofs_nid_t nid, erofs_nid_t parent_nid);
+
static void usage(void)
{
fputs("usage: [options] IMAGE\n\n"
"Dump erofs layout from IMAGE, and [options] are:\n"
+ " -S show statistic information of the image\n"
" -V print the version number of dump.erofs and exit.\n"
" -s show information about superblock\n"
" --help display this help and exit.\n",
@@ -58,13 +103,17 @@ static int erofsdump_parse_options_cfg(int argc, char **argv)
{
int opt;
- while ((opt = getopt_long(argc, argv, "Vs",
+ while ((opt = getopt_long(argc, argv, "SVs",
long_options, NULL)) != -1) {
switch (opt) {
case 's':
dumpcfg.show_superblock = true;
++dumpcfg.totalshow;
break;
+ case 'S':
+ dumpcfg.show_statistics = true;
+ ++dumpcfg.totalshow;
+ break;
case 'V':
erofsdump_print_version();
exit(0);
@@ -90,6 +139,309 @@ static int erofsdump_parse_options_cfg(int argc, char **argv)
return 0;
}
+static int erofs_get_occupied_size(struct erofs_inode *inode,
+ erofs_off_t *size)
+{
+ *size = 0;
+ switch (inode->datalayout) {
+ case EROFS_INODE_FLAT_INLINE:
+ case EROFS_INODE_FLAT_PLAIN:
+ case EROFS_INODE_CHUNK_BASED:
+ stats.uncompressed_files++;
+ *size = inode->i_size;
+ break;
+ case EROFS_INODE_FLAT_COMPRESSION_LEGACY:
+ case EROFS_INODE_FLAT_COMPRESSION:
+ stats.compressed_files++;
+ *size = inode->u.i_blocks * EROFS_BLKSIZ;
+ break;
+ default:
+ erofs_err("unknown datalayout");
+ return -1;
+ }
+ return 0;
+}
+
+static int erofs_getfile_extension(const char *filename)
+{
+ char *postfix = strrchr(filename, '.');
+ int type = 0;
+
+ if (!postfix)
+ return OTHERFILETYPE - 1;
+ for (type = 0; type < OTHERFILETYPE - 1; ++type) {
+ if (strcmp(postfix, file_types[type]) == 0)
+ break;
+ }
+ return type;
+}
+
+static void update_file_size_statatics(erofs_off_t occupied_size,
+ erofs_off_t original_size)
+{
+ int occupied_size_mark, original_size_mark;
+
+ original_size_mark = 0;
+ occupied_size_mark = 0;
+ occupied_size >>= 10;
+ original_size >>= 10;
+
+ while (occupied_size || original_size) {
+ if (occupied_size) {
+ occupied_size >>= 1;
+ occupied_size_mark++;
+ }
+ if (original_size) {
+ original_size >>= 1;
+ original_size_mark++;
+ }
+ }
+
+ if (original_size_mark >= FILE_MAX_SIZE_BITS)
+ stats.file_original_size[FILE_MAX_SIZE_BITS]++;
+ else
+ stats.file_original_size[original_size_mark]++;
+
+ if (occupied_size_mark >= FILE_MAX_SIZE_BITS)
+ stats.file_comp_size[FILE_MAX_SIZE_BITS]++;
+ else
+ stats.file_comp_size[occupied_size_mark]++;
+}
+
+static inline int erofs_checkdirent(struct erofs_dirent *de,
+ struct erofs_dirent *last_de,
+ u32 maxsize, const char *dname)
+{
+ int dname_len;
+ unsigned int nameoff = le16_to_cpu(de->nameoff);
+
+ if (nameoff < sizeof(struct erofs_dirent) ||
+ nameoff >= PAGE_SIZE) {
+ erofs_err("invalid de[0].nameoff %u @ nid %llu",
+ nameoff, de->nid | 0ULL);
+ return -EFSCORRUPTED;
+ }
+
+ dname_len = (de + 1 >= last_de) ? strnlen(dname, maxsize - nameoff) :
+ le16_to_cpu(de[1].nameoff) - nameoff;
+ /* a corrupted entry is found */
+ if (nameoff + dname_len > maxsize ||
+ dname_len > EROFS_NAME_LEN) {
+ erofs_err("bogus dirent @ nid %llu",
+ le64_to_cpu(de->nid) | 0ULL);
+ DBG_BUGON(1);
+ return -EFSCORRUPTED;
+ }
+ if (de->file_type >= EROFS_FT_MAX) {
+ erofs_err("invalid file type %llu", de->nid);
+ return -EFSCORRUPTED;
+ }
+ return dname_len;
+}
+
+static int erofs_read_dirent(struct erofs_dirent *de,
+ erofs_nid_t nid, erofs_nid_t parent_nid,
+ const char *dname)
+{
+ int err;
+ erofs_off_t occupied_size = 0;
+ struct erofs_inode inode = { .nid = de->nid };
+
+ stats.files++;
+ stats.file_category_stat[de->file_type]++;
+ err = erofs_read_inode_from_disk(&inode);
+ if (err) {
+ erofs_err("read file inode from disk failed!");
+ return err;
+ }
+
+ err = erofs_get_occupied_size(&inode, &occupied_size);
+ if (err) {
+ erofs_err("get file size failed\n");
+ return err;
+ }
+
+ if (de->file_type == EROFS_FT_REG_FILE) {
+ stats.files_total_origin_size += inode.i_size;
+ stats.file_type_stat[erofs_getfile_extension(dname)]++;
+ stats.files_total_size += occupied_size;
+ update_file_size_statatics(occupied_size, inode.i_size);
+ }
+
+ if ((de->file_type == EROFS_FT_DIR)
+ && de->nid != nid && de->nid != parent_nid) {
+ err = erofs_read_dir(de->nid, nid);
+ if (err) {
+ erofs_err("parse dir nid %llu error occurred\n",
+ de->nid);
+ return err;
+ }
+ }
+ return 0;
+}
+
+
+static int erofs_read_dir(erofs_nid_t nid, erofs_nid_t parent_nid)
+{
+ int err;
+ erofs_off_t offset;
+ char buf[EROFS_BLKSIZ];
+ struct erofs_inode vi = { .nid = nid };
+
+ err = erofs_read_inode_from_disk(&vi);
+ if (err)
+ return err;
+
+ offset = 0;
+ while (offset < vi.i_size) {
+ erofs_off_t maxsize = min_t(erofs_off_t,
+ vi.i_size - offset, EROFS_BLKSIZ);
+ struct erofs_dirent *de = (void *)buf;
+ struct erofs_dirent *end;
+ unsigned int nameoff;
+
+ err = erofs_pread(&vi, buf, maxsize, offset);
+ if (err)
+ return err;
+
+ nameoff = le16_to_cpu(de->nameoff);
+ end = (void *)buf + nameoff;
+ while (de < end) {
+ const char *dname;
+ int ret;
+
+ /* skip "." and ".." dentry */
+ if (de->nid == nid || de->nid == parent_nid) {
+ de++;
+ continue;
+ }
+
+ dname = (char *)buf + nameoff;
+ ret = erofs_checkdirent(de, end, maxsize, dname);
+ if (ret < 0)
+ return ret;
+ ret = erofs_read_dirent(de, nid, parent_nid, dname);
+ if (ret < 0)
+ return ret;
+ ++de;
+ }
+ offset += maxsize;
+ }
+ return 0;
+}
+
+static void erofsdump_print_chart_row(char *col1, unsigned int col2,
+ double col3, char *col4)
+{
+ char row[500] = {0};
+
+ sprintf(row, chart_format, col1, col2, col3, col4);
+ fprintf(stdout, row);
+}
+
+static void erofsdump_filesize_distribution(const char *title,
+ unsigned int *file_counts, unsigned int len)
+{
+ char col1[30];
+ unsigned int col2;
+ double col3;
+ char col4[400];
+ unsigned int lowerbound = 0;
+ unsigned int upperbound = 1;
+
+ fprintf(stdout, "\n%s file size distribution:\n", title);
+ fprintf(stdout, header_format, ">=(KB) .. <(KB) ", "count",
+ "ratio", "distribution");
+ for (int i = 0; i < len; i++) {
+ memset(col1, 0, sizeof(col1));
+ memset(col4, 0, sizeof(col4));
+ if (i == len - 1)
+ sprintf(col1, "%6d ..", lowerbound);
+ else if (i <= 6)
+ sprintf(col1, "%6d .. %-6d", lowerbound, upperbound);
+ else
+
+ sprintf(col1, "%6d .. %-6d", lowerbound, upperbound);
+ col2 = file_counts[i];
+ if (stats.file_category_stat[EROFS_FT_REG_FILE])
+ col3 = (double)(100 * col2) /
+ stats.file_category_stat[EROFS_FT_REG_FILE];
+ else
+ col3 = 0.0;
+ memset(col4, '#', col3 / 2);
+ erofsdump_print_chart_row(col1, col2, col3, col4);
+ lowerbound = upperbound;
+ upperbound <<= 1;
+ }
+}
+
+static void erofsdump_filetype_distribution(char **file_types, unsigned int len)
+{
+ char col1[30];
+ unsigned int col2;
+ double col3;
+ char col4[401];
+
+ fprintf(stdout, "\nFile type distribution:\n");
+ fprintf(stdout, header_format, "type", "count", "ratio",
+ "distribution");
+ for (int i = 0; i < len; i++) {
+ memset(col1, 0, sizeof(col1));
+ memset(col4, 0, sizeof(col4));
+ sprintf(col1, "%-17s", file_types[i]);
+ col2 = stats.file_type_stat[i];
+ if (stats.file_category_stat[EROFS_FT_REG_FILE])
+ col3 = (double)(100 * col2) /
+ stats.file_category_stat[EROFS_FT_REG_FILE];
+ else
+ col3 = 0.0;
+ memset(col4, '#', col3 / 2);
+ erofsdump_print_chart_row(col1, col2, col3, col4);
+ }
+}
+
+static void erofsdump_file_statistic(void)
+{
+ fprintf(stdout, "Filesystem total file count: %lu\n",
+ stats.files);
+ for (int i = 0; i < EROFS_FT_MAX; i++)
+ fprintf(stdout, "Filesystem %s count: %lu\n",
+ file_category_types[i], stats.file_category_stat[i]);
+
+ stats.compress_rate = (double)(100 * stats.files_total_size) /
+ (double)(stats.files_total_origin_size);
+ fprintf(stdout, "Filesystem compressed files: %lu\n",
+ stats.compressed_files);
+ fprintf(stdout, "Filesystem uncompressed files: %lu\n",
+ stats.uncompressed_files);
+ fprintf(stdout, "Filesystem total original file size: %lu Bytes\n",
+ stats.files_total_origin_size);
+ fprintf(stdout, "Filesystem total file size: %lu Bytes\n",
+ stats.files_total_size);
+ fprintf(stdout, "Filesystem compress rate: %.2f%%\n",
+ stats.compress_rate);
+}
+
+static void erofsdump_print_statistic(void)
+{
+ int err;
+
+ err = erofs_read_dir(sbi.root_nid, sbi.root_nid);
+ if (err) {
+ erofs_err("read dir failed");
+ return;
+ }
+
+ erofsdump_file_statistic();
+ erofsdump_filesize_distribution("Original",
+ stats.file_original_size,
+ ARRAY_SIZE(stats.file_original_size));
+ erofsdump_filesize_distribution("On-disk",
+ stats.file_comp_size,
+ ARRAY_SIZE(stats.file_comp_size));
+ erofsdump_filetype_distribution(file_types, OTHERFILETYPE);
+}
+
static void erofsdump_show_superblock(void)
{
time_t time = sbi.build_time;
@@ -156,6 +508,9 @@ int main(int argc, char **argv)
if (dumpcfg.show_superblock)
erofsdump_show_superblock();
+ if (dumpcfg.show_statistics)
+ erofsdump_print_statistic();
+
exit:
erofs_exit_configure();
return err;
diff --git a/include/erofs/internal.h b/include/erofs/internal.h
index a487871..139656c 100644
--- a/include/erofs/internal.h
+++ b/include/erofs/internal.h
@@ -269,6 +269,7 @@ struct erofs_map_blocks {
int erofs_read_superblock(void);
/* namei.c */
+int erofs_read_inode_from_disk(struct erofs_inode *vi);
int erofs_ilookup(const char *path, struct erofs_inode *vi);
/* data.c */
diff --git a/lib/namei.c b/lib/namei.c
index b4bdabf..56f199a 100644
--- a/lib/namei.c
+++ b/lib/namei.c
@@ -22,7 +22,7 @@ static dev_t erofs_new_decode_dev(u32 dev)
return makedev(major, minor);
}
-static int erofs_read_inode_from_disk(struct erofs_inode *vi)
+int erofs_read_inode_from_disk(struct erofs_inode *vi)
{
int ret, ifmt;
char buf[sizeof(struct erofs_inode_extended)];
--
2.31.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v3 3/5] erofs-utils: dump: add option to print specified file infomation.
2021-10-29 9:12 [PATCH v3 1/5] erofs-utils: add support for the full decompressed length Guo Xuenan
2021-10-29 9:12 ` [PATCH v3 2/5] erofs-utils: dump: add feature for collecting filesystem statistics Guo Xuenan
@ 2021-10-29 9:12 ` Guo Xuenan
2021-10-29 9:12 ` [PATCH v3 4/5] erofs-utils: dump: add support for showing file extents Guo Xuenan
` (2 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Guo Xuenan @ 2021-10-29 9:12 UTC (permalink / raw)
To: xiang, linux-erofs; +Cc: mpiglet
From: Wang Qi <mpiglet@outlook.com>
add option --nid to print information of specific nid, including
file name/file links/file size/data layout/compression ratio etc..
Signed-off-by: Guo Xuenan <guoxuenan@huawei.com>
Signed-off-by: Wang Qi <mpiglet@outlook.com>
---
dump/main.c | 141 ++++++++++++++++++++++++++++++++++++++++++++++++++--
1 file changed, 138 insertions(+), 3 deletions(-)
diff --git a/dump/main.c b/dump/main.c
index c5e82d1..78e952e 100644
--- a/dump/main.c
+++ b/dump/main.c
@@ -17,8 +17,10 @@
struct erofsdump_cfg {
unsigned int totalshow;
+ bool show_inode;
bool show_superblock;
bool show_statistics;
+ erofs_nid_t nid;
};
static struct erofsdump_cfg dumpcfg;
@@ -64,7 +66,8 @@ struct erofs_statistics {
static struct erofs_statistics stats;
static struct option long_options[] = {
- {"help", no_argument, 0, 1},
+ {"help", no_argument, NULL, 1},
+ {"nid", required_argument, NULL, 2},
{0, 0, 0, 0},
};
@@ -82,14 +85,18 @@ static struct erofsdump_feature feature_lists[] = {
};
static int erofs_read_dir(erofs_nid_t nid, erofs_nid_t parent_nid);
+static inline int erofs_checkdirent(struct erofs_dirent *de,
+ struct erofs_dirent *last_de,
+ u32 maxsize, const char *dname);
static void usage(void)
{
fputs("usage: [options] IMAGE\n\n"
"Dump erofs layout from IMAGE, and [options] are:\n"
- " -S show statistic information of the image\n"
" -V print the version number of dump.erofs and exit.\n"
" -s show information about superblock\n"
+ " -S show statistic information of the image\n"
+ " --nid=# show the target inode info of nid #\n"
" --help display this help and exit.\n",
stderr);
}
@@ -103,7 +110,7 @@ static int erofsdump_parse_options_cfg(int argc, char **argv)
{
int opt;
- while ((opt = getopt_long(argc, argv, "SVs",
+ while ((opt = getopt_long(argc, argv, "SV:s",
long_options, NULL)) != -1) {
switch (opt) {
case 's':
@@ -117,6 +124,11 @@ static int erofsdump_parse_options_cfg(int argc, char **argv)
case 'V':
erofsdump_print_version();
exit(0);
+ case 2:
+ dumpcfg.show_inode = true;
+ dumpcfg.nid = (erofs_nid_t)atoll(optarg);
+ ++dumpcfg.totalshow;
+ break;
case 1:
usage();
exit(0);
@@ -330,6 +342,126 @@ static int erofs_read_dir(erofs_nid_t nid, erofs_nid_t parent_nid)
return 0;
}
+static int erofs_get_pathname(erofs_nid_t nid, erofs_nid_t parent_nid,
+ erofs_nid_t target, char *path, unsigned int pos)
+{
+ int err;
+ erofs_off_t offset;
+ char buf[EROFS_BLKSIZ];
+ struct erofs_inode inode = { .nid = nid };
+
+ path[pos++] = '/';
+ if (target == sbi.root_nid)
+ return 0;
+
+ err = erofs_read_inode_from_disk(&inode);
+ if (err) {
+ erofs_err("read inode %lu failed", nid);
+ return err;
+ }
+
+ offset = 0;
+ while (offset < inode.i_size) {
+ erofs_off_t maxsize = min_t(erofs_off_t,
+ inode.i_size - offset, EROFS_BLKSIZ);
+ struct erofs_dirent *de = (void *)buf;
+ struct erofs_dirent *end;
+ unsigned int nameoff;
+
+ err = erofs_pread(&inode, buf, maxsize, offset);
+ if (err)
+ return err;
+
+ nameoff = le16_to_cpu(de->nameoff);
+ end = (void *)buf + nameoff;
+ while (de < end) {
+ const char *dname;
+ int len;
+
+ nameoff = le16_to_cpu(de->nameoff);
+ dname = (char *)buf + nameoff;
+ len = erofs_checkdirent(de, end, maxsize, dname);
+ if (len < 0)
+ return len;
+
+ if (de->nid == target) {
+ memcpy(path + pos, dname, len);
+ path[pos + len] = '\0';
+ return 0;
+ }
+
+ if (de->file_type == EROFS_FT_DIR &&
+ de->nid != parent_nid &&
+ de->nid != nid) {
+ memcpy(path + pos, dname, len);
+ err = erofs_get_pathname(de->nid, nid,
+ target, path, pos + len);
+ if (!err)
+ return 0;
+ memset(path + pos, 0, len);
+ }
+ ++de;
+ }
+ offset += maxsize;
+ }
+ return -1;
+}
+
+static void erofsdump_show_fileinfo(void)
+{
+ int err;
+ erofs_off_t size;
+ u16 access_mode;
+ time_t time_sec;
+ u32 time_nsec;
+ erofs_nid_t nid = dumpcfg.nid;
+ struct erofs_inode inode = { .nid = nid };
+ char path[PATH_MAX + 1] = {0};
+ char access_mode_str[] = "rwxrwxrwx";
+ char timebuf[128] = {0};
+
+ err = erofs_read_inode_from_disk(&inode);
+ if (err) {
+ erofs_err("failed to find inode %lu from disk", nid);
+ return;
+ }
+
+ err = erofs_get_occupied_size(&inode, &size);
+ if (err) {
+ erofs_err("get file size failed\n");
+ return;
+ }
+
+ err = erofs_get_pathname(sbi.root_nid, sbi.root_nid, nid, path, 0);
+ if (err < 0) {
+ fprintf(stderr, "File: Path not found.\n");
+ return;
+ }
+
+ time_sec = inode.i_ctime;
+ time_nsec = inode.i_ctime_nsec;
+ strftime(timebuf, sizeof(timebuf),
+ "%Y-%m-%d %H:%M:%S", localtime(&time_sec));
+ access_mode = inode.i_mode & 0777;
+ for (int i = 8; i >= 0; i--)
+ if (((access_mode >> i) & 1) == 0)
+ access_mode_str[8 - i] = '-';
+ fprintf(stdout, "File : %s\n", path);
+ fprintf(stdout, "Inode: %lu ", inode.nid);
+ fprintf(stdout, "Links: %u ", inode.i_nlink);
+ fprintf(stdout, "Layout: %d\n", inode.datalayout);
+ fprintf(stdout, "Inode size: %d ", inode.inode_isize);
+ fprintf(stdout, "Extent size: %u ", inode.extent_isize);
+ fprintf(stdout, "Xattr size: %u\n", inode.xattr_isize);
+ fprintf(stdout, "File size: %lu ", inode.i_size);
+ fprintf(stdout, "On-disk size: %lu ", size);
+ fprintf(stdout, "Compression ratio: %.2f%%\n",
+ (double)(100 * size) / (double)(inode.i_size));
+ fprintf(stdout, "Uid: %u Gid: %u ", inode.i_uid, inode.i_gid);
+ fprintf(stdout, "Access: %04o/%s\n", access_mode, access_mode_str);
+ fprintf(stdout, "Timestamp: %s.%09d\n", timebuf, time_nsec);
+}
+
static void erofsdump_print_chart_row(char *col1, unsigned int col2,
double col3, char *col4)
{
@@ -511,6 +643,9 @@ int main(int argc, char **argv)
if (dumpcfg.show_statistics)
erofsdump_print_statistic();
+ if (dumpcfg.show_inode)
+ erofsdump_show_fileinfo();
+
exit:
erofs_exit_configure();
return err;
--
2.31.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v3 4/5] erofs-utils: dump: add support for showing file extents.
2021-10-29 9:12 [PATCH v3 1/5] erofs-utils: add support for the full decompressed length Guo Xuenan
2021-10-29 9:12 ` [PATCH v3 2/5] erofs-utils: dump: add feature for collecting filesystem statistics Guo Xuenan
2021-10-29 9:12 ` [PATCH v3 3/5] erofs-utils: dump: add option to print specified file infomation Guo Xuenan
@ 2021-10-29 9:12 ` Guo Xuenan
2021-10-29 9:12 ` [PATCH v3 5/5] erofs-utils: manpage: add dump.erofs manpage Guo Xuenan
2021-11-09 11:21 ` [PATCH v3 1/5] erofs-utils: add support for the full decompressed length Gao Xiang
4 siblings, 0 replies; 6+ messages in thread
From: Guo Xuenan @ 2021-10-29 9:12 UTC (permalink / raw)
To: xiang, linux-erofs; +Cc: mpiglet
Add option -e to show file extents info, this option needs
specify nid as well. (eg. dump.erofs --nid # -e erofs.img)
Signed-off-by: Guo Xuenan <guoxuenan@huawei.com>
Signed-off-by: Wang Qi <mpiglet@outlook.com>
---
dump/main.c | 50 +++++++++++++++++++++++++++++++++++++---
include/erofs/internal.h | 2 ++
lib/data.c | 4 ++--
3 files changed, 51 insertions(+), 5 deletions(-)
diff --git a/dump/main.c b/dump/main.c
index 78e952e..4acb742 100644
--- a/dump/main.c
+++ b/dump/main.c
@@ -18,6 +18,7 @@
struct erofsdump_cfg {
unsigned int totalshow;
bool show_inode;
+ bool show_extent;
bool show_superblock;
bool show_statistics;
erofs_nid_t nid;
@@ -94,6 +95,7 @@ static void usage(void)
fputs("usage: [options] IMAGE\n\n"
"Dump erofs layout from IMAGE, and [options] are:\n"
" -V print the version number of dump.erofs and exit.\n"
+ " -e show extent info (require --nid set)\n"
" -s show information about superblock\n"
" -S show statistic information of the image\n"
" --nid=# show the target inode info of nid #\n"
@@ -110,9 +112,13 @@ static int erofsdump_parse_options_cfg(int argc, char **argv)
{
int opt;
- while ((opt = getopt_long(argc, argv, "SV:s",
+ while ((opt = getopt_long(argc, argv, "SVes",
long_options, NULL)) != -1) {
switch (opt) {
+ case 'e':
+ dumpcfg.show_extent = true;
+ ++dumpcfg.totalshow;
+ break;
case 's':
dumpcfg.show_superblock = true;
++dumpcfg.totalshow;
@@ -407,7 +413,15 @@ static int erofs_get_pathname(erofs_nid_t nid, erofs_nid_t parent_nid,
return -1;
}
-static void erofsdump_show_fileinfo(void)
+static int erofsdump_map_blocks(struct erofs_inode *inode,
+ struct erofs_map_blocks *map, int flags)
+{
+ if (erofs_inode_is_data_compressed(inode->datalayout))
+ return z_erofs_map_blocks_iter(inode, map, flags);
+ return erofs_map_blocks(inode, map, flags);
+}
+
+static void erofsdump_show_fileinfo(bool show_extent)
{
int err;
erofs_off_t size;
@@ -419,6 +433,11 @@ static void erofsdump_show_fileinfo(void)
char path[PATH_MAX + 1] = {0};
char access_mode_str[] = "rwxrwxrwx";
char timebuf[128] = {0};
+ unsigned int extent_count = 0;
+ struct erofs_map_blocks map = {
+ .index = UINT_MAX,
+ .m_la = 0,
+ };
err = erofs_read_inode_from_disk(&inode);
if (err) {
@@ -460,6 +479,26 @@ static void erofsdump_show_fileinfo(void)
fprintf(stdout, "Uid: %u Gid: %u ", inode.i_uid, inode.i_gid);
fprintf(stdout, "Access: %04o/%s\n", access_mode, access_mode_str);
fprintf(stdout, "Timestamp: %s.%09d\n", timebuf, time_nsec);
+
+ if (!dumpcfg.show_extent)
+ return;
+
+ fprintf(stdout, "\nExt: logical_offset/length: physical_offset/length\n");
+ while (map.m_la < inode.i_size) {
+ err = erofsdump_map_blocks(&inode, &map,
+ EROFS_GET_BLOCKS_FIEMAP);
+ if (err) {
+ erofs_err("get file blocks range failed");
+ return;
+ }
+ fprintf(stdout, "%u: ", extent_count++);
+ fprintf(stdout, "%6lu..%-6lu / %-lu: ",
+ map.m_la, map.m_la + map.m_llen, map.m_llen);
+ fprintf(stdout, "%lu..%lu / %-lu\n",
+ map.m_pa, map.m_pa + map.m_plen, map.m_plen);
+ map.m_la += map.m_llen;
+ }
+ fprintf(stdout, "%s:%d extents found\n", cfg.c_img_path, extent_count);
}
static void erofsdump_print_chart_row(char *col1, unsigned int col2,
@@ -643,8 +682,13 @@ int main(int argc, char **argv)
if (dumpcfg.show_statistics)
erofsdump_print_statistic();
+ if (dumpcfg.show_extent && !dumpcfg.show_inode) {
+ usage();
+ goto exit;
+ }
+
if (dumpcfg.show_inode)
- erofsdump_show_fileinfo();
+ erofsdump_show_fileinfo(dumpcfg.show_extent);
exit:
erofs_exit_configure();
diff --git a/include/erofs/internal.h b/include/erofs/internal.h
index 139656c..8282f03 100644
--- a/include/erofs/internal.h
+++ b/include/erofs/internal.h
@@ -275,6 +275,8 @@ int erofs_ilookup(const char *path, struct erofs_inode *vi);
/* data.c */
int erofs_pread(struct erofs_inode *inode, char *buf,
erofs_off_t count, erofs_off_t offset);
+int erofs_map_blocks(struct erofs_inode *inode,
+ struct erofs_map_blocks *map, int flags);
/* zmap.c */
int z_erofs_fill_inode(struct erofs_inode *vi);
int z_erofs_map_blocks_iter(struct erofs_inode *vi,
diff --git a/lib/data.c b/lib/data.c
index 30e7312..9048707 100644
--- a/lib/data.c
+++ b/lib/data.c
@@ -61,8 +61,8 @@ err_out:
return err;
}
-static int erofs_map_blocks(struct erofs_inode *inode,
- struct erofs_map_blocks *map, int flags)
+int erofs_map_blocks(struct erofs_inode *inode,
+ struct erofs_map_blocks *map, int flags)
{
struct erofs_inode *vi = inode;
struct erofs_inode_chunk_index *idx;
--
2.31.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v3 5/5] erofs-utils: manpage: add dump.erofs manpage.
2021-10-29 9:12 [PATCH v3 1/5] erofs-utils: add support for the full decompressed length Guo Xuenan
` (2 preceding siblings ...)
2021-10-29 9:12 ` [PATCH v3 4/5] erofs-utils: dump: add support for showing file extents Guo Xuenan
@ 2021-10-29 9:12 ` Guo Xuenan
2021-11-09 11:21 ` [PATCH v3 1/5] erofs-utils: add support for the full decompressed length Gao Xiang
4 siblings, 0 replies; 6+ messages in thread
From: Guo Xuenan @ 2021-10-29 9:12 UTC (permalink / raw)
To: xiang, linux-erofs; +Cc: mpiglet
This patch adds dump.erofs manpage.
Signed-off-by: Guo Xuenan <guoxuenan@huawei.com>
---
man/Makefile.am | 2 +-
man/dump.erofs.1 | 58 ++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 59 insertions(+), 1 deletion(-)
create mode 100644 man/dump.erofs.1
diff --git a/man/Makefile.am b/man/Makefile.am
index d62d6e2..769b557 100644
--- a/man/Makefile.am
+++ b/man/Makefile.am
@@ -1,6 +1,6 @@
# SPDX-License-Identifier: GPL-2.0+
-dist_man_MANS = mkfs.erofs.1
+dist_man_MANS = mkfs.erofs.1 dump.erofs.1
EXTRA_DIST = erofsfuse.1
if ENABLE_FUSE
diff --git a/man/dump.erofs.1 b/man/dump.erofs.1
new file mode 100644
index 0000000..98be560
--- /dev/null
+++ b/man/dump.erofs.1
@@ -0,0 +1,58 @@
+.\" Copyright (c) 2021 Guo Xuenan <guoxuenan@huawei.com>
+.\"
+.TH DUMP.EROFS 1
+.SH NAME
+dump.erofs \- retrieve directory and file entries, show specific file
+or overall disk statistics information from an EROFS-formatted image.
+.SH SYNOPSIS
+.B dump.erofs
+[
+.B \--nid
+.I inode number
+]
+[
+.B \-e
+]
+[
+.B \-s
+]
+[
+.B \-S
+]
+[
+.B \-V
+]
+.I IMAGE
+.SH DESCRIPTION
+.B dump.erofs
+is used to retrieve erofs metadata from \fIIMAGE\fP and demonstrate
+1) overall disk statistics,
+2) erofs superblock information,
+3) file information of given inode number,
+4) file extent information.
+.SH OPTIONS
+.TP
+.BI \--nid " inode number"
+Specify an inode number to print its file information.
+.TP
+.BI \-e
+Show the file extent information, the option depends on option --nid to specify nid.
+.TP
+.BI \-V
+Print the version number and exit.
+.TP
+.BI \-s
+Show superblock information of the an EROFS-formated image.
+.TP
+.BI \-S
+Show EROFS disk statistics, including file type/size distribution, number of (un)compressed files, compression ratio of the whole image, etc.
+.SH AUTHOR
+Initial code was written by Wang Qi <mpiglet@outlook.com>, Guo Xuenan <guoxuenan@huawei.com>.
+.PP
+This manual page was written by Guo Xuenan <guoxuenan@huawei.com>
+.SH AVAILABILITY
+.B dump.erofs
+is part of erofs-utils package and is available from git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs-utils.git.
+.SH SEE ALSO
+.BR mkfs.erofs(1),
+.BR fsck.erofs(1)
--
2.31.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH v3 1/5] erofs-utils: add support for the full decompressed length
2021-10-29 9:12 [PATCH v3 1/5] erofs-utils: add support for the full decompressed length Guo Xuenan
` (3 preceding siblings ...)
2021-10-29 9:12 ` [PATCH v3 5/5] erofs-utils: manpage: add dump.erofs manpage Guo Xuenan
@ 2021-11-09 11:21 ` Gao Xiang
4 siblings, 0 replies; 6+ messages in thread
From: Gao Xiang @ 2021-11-09 11:21 UTC (permalink / raw)
To: Guo Xuenan; +Cc: linux-erofs, mpiglet
On Fri, Oct 29, 2021 at 05:12:40PM +0800, Guo Xuenan wrote:
> From: Huang Jianan <huangjianan@oppo.com>
>
> Previously, there is no need to get the full decompressed length since
> EROFS supports partial decompression. However for some other cases
> such as fiemap, the full decompressed length is necessary for iomap to
> make it work properly.
>
> This patch adds a way to get the full decompressed length. Note that
> it takes more metadata overhead and it'd be avoided if possible in the
> performance sensitive scenario.
>
> Signed-off-by: Huang Jianan <huangjianan@oppo.com>
> Signed-off-by: Gao Xiang <xiang@kernel.org>
The series has been applied with some clean-up work.
^ permalink raw reply [flat|nested] 6+ messages in thread