mdadm: Use _FILE_OFFSET_BITS to use largefile support
PATCH 1/6 mdadm: Use _FILE_OFFSET_BITS to use largefile support
--- ...FILE_OFFSET_BITS-to-enable-largefile.patch | 147 ++++++++++++++++++ meta/recipes-extended/mdadm/mdadm_4.2.bb | 1 + 2 files changed, 148 insertions(+) create mode 100644 meta/recipes-extended/mdadm/files/0001-restripe.c-Use-_FILE_OFFSET_BITS-to-enable-largefile.patch
meta/recipes-extended/mdadm/files/0001-restripe.c-Use-_FILE_OFFSET_BITS-to-enable-largefile.patch new file +147 -0
@@ -0,0 +1,147 @@ +From 6b861a267a6ef6f60f6cc21e4c8e6d7cdd2451dc Mon Sep 17 00:00:00 2001 +From: Khem Raj <[email protected]> +Date: Thu, 10 Nov 2022 12:31:22 -0800 +Subject: [PATCH] restripe.c: Use _FILE_OFFSET_BITS to enable largefile support + +Instead of using the lseek64 and friends, its better to enable it via +the feature macro _FILE_OFFSET_BITS = 64 and let the C library deal with +the width of types + +Upstream-Status: Submitted [https://lore.kernel.org/linux-raid/[email protected]/] +Signed-off-by: Khem Raj <[email protected]> +--- + restripe.c | 13 ++++++++----- + 1 file changed, 8 insertions(+), 5 deletions(-) + +--- a/restripe.c ++++ b/restripe.c +@@ -22,6 +22,9 @@ + * Email: <[email protected]> + */ + ++/* Enable largefile support */ ++#define _FILE_OFFSET_BITS 64 ++ + #include "mdadm.h" + #include <stdint.h> + +@@ -581,7 +584,7 @@ int save_stripes(int *source, unsigned l + raid_disks, level, layout); + if (dnum < 0) abort(); + if (source[dnum] < 0 || +- lseek64(source[dnum], ++ lseek(source[dnum], + offsets[dnum] + offset, 0) < 0 || + read(source[dnum], buf+disk * chunk_size, + chunk_size) != chunk_size) { +@@ -754,8 +757,8 @@ int restore_stripes(int *dest, unsigned + raid_disks, level, layout); + if (src_buf == NULL) { + /* read from file */ +- if (lseek64(source, read_offset, 0) != +- (off64_t)read_offset) { ++ if (lseek(source, read_offset, 0) != ++ (off_t)read_offset) { + rv = -1; + goto abort; + } +@@ -816,7 +819,7 @@ int restore_stripes(int *dest, unsigned + } + for (i=0; i < raid_disks ; i++) + if (dest[i] >= 0) { +- if (lseek64(dest[i], ++ if (lseek(dest[i], + offsets[i]+offset, 0) < 0) { + rv = -1; + goto abort; +@@ -866,7 +869,7 @@ int test_stripes(int *source, unsigned l + int disk; + + for (i = 0 ; i < raid_disks ; i++) { +- if ((lseek64(source[i], offsets[i]+start, 0) < 0) || ++ if ((lseek(source[i], offsets[i]+start, 0) < 0) || + (read(source[i], stripes[i], chunk_size) != + chunk_size)) { + free(q); +--- a/raid6check.c ++++ b/raid6check.c +@@ -22,6 +22,9 @@ + * Based on "restripe.c" from "mdadm" codebase + */ + ++/* Enable largefile support */ ++#define _FILE_OFFSET_BITS 64 ++ + #include "mdadm.h" + #include <stdint.h> + #include <signal.h> +@@ -279,9 +282,9 @@ int manual_repair(int chunk_size, int sy + } + + int write_res1, write_res2; +- off64_t seek_res; ++ off_t seek_res; + +- seek_res = lseek64(source[fd1], ++ seek_res = lseek(source[fd1], + offsets[fd1] + start * chunk_size, SEEK_SET); + if (seek_res < 0) { + fprintf(stderr, "lseek failed for failed_disk1\n"); +@@ -289,7 +292,7 @@ int manual_repair(int chunk_size, int sy + } + write_res1 = write(source[fd1], blocks[failed_slot1], chunk_size); + +- seek_res = lseek64(source[fd2], ++ seek_res = lseek(source[fd2], + offsets[fd2] + start * chunk_size, SEEK_SET); + if (seek_res < 0) { + fprintf(stderr, "lseek failed for failed_disk2\n"); +@@ -374,7 +377,7 @@ int check_stripes(struct mdinfo *info, i + goto exitCheck; + } + for (i = 0 ; i < raid_disks ; i++) { +- off64_t seek_res = lseek64(source[i], offsets[i] + start * chunk_size, ++ off_t seek_res = lseek(source[i], offsets[i] + start * chunk_size, + SEEK_SET); + if (seek_res < 0) { + fprintf(stderr, "lseek to source %d failed\n", i); +--- a/swap_super.c ++++ b/swap_super.c +@@ -1,3 +1,6 @@ ++/* Enable largefile support */ ++#define _FILE_OFFSET_BITS 64 ++ + #include <unistd.h> + #include <stdlib.h> + #include <fcntl.h> +@@ -16,8 +19,6 @@ + + #define MD_NEW_SIZE_SECTORS(x) ((x & ~(MD_RESERVED_SECTORS - 1)) - MD_RESERVED_SECTORS) + +-extern long long lseek64(int, long long, int); +- + int main(int argc, char *argv[]) + { + int fd, i; +@@ -38,8 +39,8 @@ int main(int argc, char *argv[]) + exit(1); + } + offset = MD_NEW_SIZE_SECTORS(size) * 512LL; +- if (lseek64(fd, offset, 0) < 0LL) { +- perror("lseek64"); ++ if (lseek(fd, offset, 0) < 0LL) { ++ perror("lseek"); + exit(1); + } + if (read(fd, super, 4096) != 4096) { +@@ -68,8 +69,8 @@ int main(int argc, char *argv[]) + super[32*4+10*4 +i] = t; + } + +- if (lseek64(fd, offset, 0) < 0LL) { +- perror("lseek64"); ++ if (lseek(fd, offset, 0) < 0LL) { ++ perror("lseek"); + exit(1); + } + if (write(fd, super, 4096) != 4096) {
meta/recipes-extended/mdadm/mdadm_4.2.bb +1 -0
@@ -24,6 +24,7 @@ SRC_URI = "${KERNELORG_MIRROR}/linux/utils/raid/mdadm/${BPN}-${PV}.tar.xz \ file://0001-mdadm-skip-test-11spare-migration.patch \ file://0001-Fix-parsing-of-r-in-monitor-manager-mode.patch \ file://0001-Makefile-install-mdcheck.patch \ + file://0001-restripe.c-Use-_FILE_OFFSET_BITS-to-enable-largefile.patch \ " SRC_URI[sha256sum] = "461c215670864bb74a4d1a3620684aa2b2f8296dffa06743f26dda5557acf01d"
Signature · 1 line
2.38.1
PATCH 2/6 acl: Enable largefile support by default
This ensures right flags are passed so enable it. Musl and decoupled these APIs from _GNU_SOURCE and now are only available when _LARGEFILE64_SOURCE macro is enabled
--- ...rtable-version-of-dirent-and-readdir.patch | 39 +++++++++++++++++++ meta/recipes-support/attr/acl_2.3.1.bb | 3 ++ 2 files changed, 42 insertions(+) create mode 100644 meta/recipes-support/attr/acl/0001-chacl-Use-portable-version-of-dirent-and-readdir.patch
meta/recipes-support/attr/acl/0001-chacl-Use-portable-version-of-dirent-and-readdir.patch new file +39 -0
@@ -0,0 +1,39 @@ +From 1a5671a4eb13b81b98c3e71f00370781563f66d8 Mon Sep 17 00:00:00 2001 +From: Khem Raj <[email protected]> +Date: Thu, 10 Nov 2022 17:57:24 -0800 +Subject: [PATCH] chacl: Use portable version of dirent and readdir + +Using 64bit versions on 32bit architectures should be enabled with +--enable-largefile, this makes it portable across musl and glibc + +Upstream-Status: Submitted [https://lists.nongnu.org/archive/html/acl-devel/2022-11/msg00005.html] +Signed-off-by: Khem Raj <[email protected]> +--- + tools/chacl.c | 4 ++-- + 1 file changed, 2 insertions(+), 2 deletions(-) + +diff --git a/tools/chacl.c b/tools/chacl.c +index 525a7ff..8fff875 100644 +--- a/tools/chacl.c ++++ b/tools/chacl.c +@@ -320,7 +320,7 @@ walk_dir(acl_t acl, acl_t dacl, const char *fname) + { + int failed = 0; + DIR *dir; +- struct dirent64 *d; ++ struct dirent *d; + char *name; + + if ((dir = opendir(fname)) == NULL) { +@@ -332,7 +332,7 @@ walk_dir(acl_t acl, acl_t dacl, const char *fname) + return(0); /* got a file, not an error */ + } + +- while ((d = readdir64(dir)) != NULL) { ++ while ((d = readdir(dir)) != NULL) { + /* skip "." and ".." entries */ + if (strcmp(d->d_name, ".") == 0 || strcmp(d->d_name, "..") == 0) + continue; +-- +2.38.1 +
meta/recipes-support/attr/acl_2.3.1.bb +3 -0
@@ -19,12 +19,15 @@ SRC_URI = "${SAVANNAH_GNU_MIRROR}/acl/${BP}.tar.gz \ file://run-ptest \ file://0001-tests-do-not-hardcode-the-build-path-into-a-helper-l.patch \ file://0001-test-patch-out-failing-bits.patch \ + file://0001-chacl-Use-portable-version-of-dirent-and-readdir.patch \ " SRC_URI[sha256sum] = "760c61c68901b37fdd5eefeeaf4c0c7a26bdfdd8ac747a1edff1ce0e243c11af" inherit autotools gettext ptest +EXTRA_OECONF += "--enable-largefile" + PACKAGES =+ "lib${BPN}" FILES:lib${BPN} = "${libdir}/lib*${SOLIBS}"
Signature · 1 line
2.38.1
Re: [OE-core] [PATCH 2/6] acl: Enable largefile support by default
Quoted 17 lines · Khem Raj
This ensures right flags are passed so enable it. Musl and decoupled these APIs from
_GNU_SOURCEand now are only available when_LARGEFILE64_SOURCEmacro is enabledSigned-off-by Khem Raj <[email protected]>--- ...rtable-version-of-dirent-and-readdir.patch | 39 +++++++++++++++++++ meta/recipes-support/attr/acl_2.3.1.bb | 3 ++ 2 files changed, 42 insertions(+) create mode 100644 meta/recipes-support/attr/acl/0001-chacl-Use-portable-version-of-dirent-and-readdir.patchmeta/recipes-support/attr/acl/0001-chacl-Use-portable-version-of-dirent-and-readdir.patch new file +0
-0
meta/recipes-support/attr/acl/0001-chacl-Use-portable-version-of-dirent-and-readdir.patch line 10
Quoted 10 lines
meta/recipes-support/attr/acl/0001-chacl-Use-portable-version-of-dirent-and-readdir.patch new file +10
-0@@ -0,0 +1,39 @@ +From 1a5671a4eb13b81b98c3e71f00370781563f66d8 Mon Sep 17 00:00:00 2001 +From: Khem Raj <[email protected]> +Date: Thu, 10 Nov 2022 17:57:24 -0800 +Subject: [PATCH] chacl: Use portable version of dirent and readdir + +Using 64bit versions on 32bit architectures should be enabled with +--enable-largefile, this makes it portable across musl and glibc + +Upstream-Status: Submitted [https://lists.nongnu.org/archive/html/acl-devel/2022-11/msg00005.html] +Signed-off-by: Khem Raj <[email protected]>
Now a backport:
https://git.savannah.nongnu.org/cgit/acl.git/commit/?id=2b42f64737adf6a2ddd491213580d6e9cdd2f5af
:)
Cheers,
Richard
PATCH 3/6 btrfs-tools: Do not use 64bit functions for largefile support
In reply to Khem RajUsing -D_FILE_OFFSET_BITS=64 already makes the normal function behave like 64bit variants. Moreover it makes them portable across libc
--- ...-ftruncate-stat-instead-of-64bit-equ.patch | 903 ++++++++++++++++++ .../btrfs-tools/btrfs-tools_6.0.1.bb | 1 + 2 files changed, 904 insertions(+) create mode 100644 meta/recipes-devtools/btrfs-tools/btrfs-tools/0001-Use-pread-pwrite-ftruncate-stat-instead-of-64bit-equ.patch
meta/recipes-devtools/btrfs-tools/btrfs-tools/0001-Use-pread-pwrite-ftruncate-stat-instead-of-64bit-equ.patch new file +903 -0
@@ -0,0 +1,903 @@ +From 37425f422e46003ed623f8e6387bed1aa5778b96 Mon Sep 17 00:00:00 2001 +From: Khem Raj <[email protected]> +Date: Fri, 11 Nov 2022 11:26:05 -0800 +Subject: [PATCH] Use pread/pwrite/ftruncate/stat instead of 64bit equivalents + +64bit functions are aliases to original functions when largefile feature +is enabled via autoconf or right macro is passed on compiler cmdline + +-D_FILE_OFFSET_BITS=64 + +Upstream-Status: Pending +Signed-off-by: Khem Raj <[email protected]> +--- + cmds/rescue-chunk-recover.c | 4 +- + image/main.c | 14 +-- + kernel-shared/zoned.c | 6 +- + kernel-shared/zoned.h | 4 +- + mkfs/main.c | 4 +- + mkfs/rootdir.c | 10 +- + tests/fsstress.c | 192 ++++++++++++++++++------------------ + tests/fssum.c | 8 +- + 8 files changed, 121 insertions(+), 121 deletions(-) + +--- a/cmds/rescue-chunk-recover.c ++++ b/cmds/rescue-chunk-recover.c +@@ -755,7 +755,7 @@ static int scan_one_device(void *dev_sca + if (is_super_block_address(bytenr)) + bytenr += rc->sectorsize; + +- if (pread64(fd, buf->data, rc->nodesize, bytenr) < ++ if (pread(fd, buf->data, rc->nodesize, bytenr) < + rc->nodesize) + break; + +@@ -1875,7 +1875,7 @@ static int check_one_csum(int fd, u64 st + data = malloc(len); + if (!data) + return -1; +- ret = pread64(fd, data, len, start); ++ ret = pread(fd, data, len, start); + if (ret < 0 || ret != len) { + ret = -1; + goto out; +--- a/image/main.c ++++ b/image/main.c +@@ -691,7 +691,7 @@ static int flush_pending(struct metadump + if (start == BTRFS_SUPER_INFO_OFFSET) { + int fd = get_dev_fd(md->root); + +- ret = pread64(fd, async->buffer, size, start); ++ ret = pread(fd, async->buffer, size, start); + if (ret < size) { + free(async->buffer); + free(async); +@@ -1366,7 +1366,7 @@ static void write_backup_supers(int fd, + break; + btrfs_set_super_bytenr(super, bytenr); + csum_block(buf, BTRFS_SUPER_INFO_SIZE); +- ret = pwrite64(fd, buf, BTRFS_SUPER_INFO_SIZE, bytenr); ++ ret = pwrite(fd, buf, BTRFS_SUPER_INFO_SIZE, bytenr); + if (ret < BTRFS_SUPER_INFO_SIZE) { + if (ret < 0) + error( +@@ -1487,12 +1487,12 @@ static int restore_one_work(struct mdres + else + bytenr = logical; + +- ret = pwrite64(outfd, buffer + offset, chunk_size, bytenr); ++ ret = pwrite(outfd, buffer + offset, chunk_size, bytenr); + if (ret != chunk_size) + goto write_error; + + if (physical_dup) +- ret = pwrite64(outfd, buffer + offset, ++ ret = pwrite(outfd, buffer + offset, + chunk_size, physical_dup); + if (ret != chunk_size) + goto write_error; +@@ -2454,7 +2454,7 @@ static int fixup_device_size(struct btrf + } + if (S_ISREG(buf.st_mode)) { + /* Don't forget to enlarge the real file */ +- ret = ftruncate64(out_fd, dev_size); ++ ret = ftruncate(out_fd, dev_size); + if (ret < 0) { + error("failed to enlarge result image: %m"); + return -errno; +@@ -2913,7 +2913,7 @@ static int restore_metadump(const char * + goto out; + } + if (S_ISREG(st.st_mode) && st.st_size < dev_size) { +- ret = ftruncate64(fileno(out), dev_size); ++ ret = ftruncate(fileno(out), dev_size); + if (ret < 0) { + error( + "failed to enlarge result image file from %llu to %llu: %m", +@@ -3010,7 +3010,7 @@ static int update_disk_super_on_device(s + memcpy(dev_item->fsid, fs_uuid, BTRFS_UUID_SIZE); + csum_block((u8 *)&disk_super, BTRFS_SUPER_INFO_SIZE); + +- ret = pwrite64(fp, &disk_super, BTRFS_SUPER_INFO_SIZE, BTRFS_SUPER_INFO_OFFSET); ++ ret = pwrite(fp, &disk_super, BTRFS_SUPER_INFO_SIZE, BTRFS_SUPER_INFO_OFFSET); + if (ret != BTRFS_SUPER_INFO_SIZE) { + if (ret < 0) { + errno = ret; +--- a/kernel-shared/zoned.c ++++ b/kernel-shared/zoned.c +@@ -194,7 +194,7 @@ static int sb_write_pointer(int fd, stru + bytenr = ((zones[i].start + zones[i].len) + << SECTOR_SHIFT) - BTRFS_SUPER_INFO_SIZE; + +- ret = pread64(fd, buf[i], BTRFS_SUPER_INFO_SIZE, bytenr); ++ ret = pread(fd, buf[i], BTRFS_SUPER_INFO_SIZE, bytenr); + if (ret != BTRFS_SUPER_INFO_SIZE) + return -EIO; + super[i] = (struct btrfs_super_block *)&buf[i]; +@@ -515,8 +515,8 @@ size_t btrfs_sb_io(int fd, void *buf, of + /* We can call pread/pwrite if 'fd' is non-zoned device/file */ + if (zone_size_sector == 0) { + if (rw == READ) +- return pread64(fd, buf, count, offset); +- return pwrite64(fd, buf, count, offset); ++ return pread(fd, buf, count, offset); ++ return pwrite(fd, buf, count, offset); + } + + ASSERT(IS_ALIGNED(zone_size_sector, sb_size_sector)); +--- a/kernel-shared/zoned.h ++++ b/kernel-shared/zoned.h +@@ -150,9 +150,9 @@ int btrfs_wipe_temporary_sb(struct btrfs + #else + + #define sbread(fd, buf, offset) \ +- pread64(fd, buf, BTRFS_SUPER_INFO_SIZE, offset) ++ pread(fd, buf, BTRFS_SUPER_INFO_SIZE, offset) + #define sbwrite(fd, buf, offset) \ +- pwrite64(fd, buf, BTRFS_SUPER_INFO_SIZE, offset) ++ pwrite(fd, buf, BTRFS_SUPER_INFO_SIZE, offset) + + static inline int btrfs_reset_dev_zone(int fd, struct blk_zone *zone) + { +--- a/mkfs/main.c ++++ b/mkfs/main.c +@@ -453,14 +453,14 @@ static int zero_output_file(int out_fd, + /* Only zero out the first 1M */ + loop_num = SZ_1M / SZ_4K; + for (i = 0; i < loop_num; i++) { +- written = pwrite64(out_fd, buf, SZ_4K, location); ++ written = pwrite(out_fd, buf, SZ_4K, location); + if (written != SZ_4K) + ret = -EIO; + location += SZ_4K; + } + + /* Then enlarge the file to size */ +- written = pwrite64(out_fd, buf, 1, size - 1); ++ written = pwrite(out_fd, buf, 1, size - 1); + if (written < 1) + ret = -EIO; + return ret; +--- a/mkfs/rootdir.c ++++ b/mkfs/rootdir.c +@@ -340,7 +340,7 @@ static int add_file_items(struct btrfs_t + goto end; + } + +- ret_read = pread64(fd, buffer, st->st_size, bytes_read); ++ ret_read = pread(fd, buffer, st->st_size, bytes_read); + if (ret_read == -1) { + error("cannot read %s at offset %llu length %llu: %m", + path_name, bytes_read, (unsigned long long)st->st_size); +@@ -386,7 +386,7 @@ again: + + memset(eb->data, 0, sectorsize); + +- ret_read = pread64(fd, eb->data, sectorsize, file_pos + ++ ret_read = pread(fd, eb->data, sectorsize, file_pos + + bytes_read); + if (ret_read == -1) { + error("cannot read %s at offset %llu length %u: %m", +@@ -929,7 +929,7 @@ int btrfs_mkfs_shrink_fs(struct btrfs_fs + u64 new_size; + struct btrfs_device *device; + struct list_head *cur; +- struct stat64 file_stat; ++ struct stat file_stat; + int nr_devs = 0; + int ret; + +@@ -963,14 +963,14 @@ int btrfs_mkfs_shrink_fs(struct btrfs_fs + *new_size_ret = new_size; + + if (shrink_file_size) { +- ret = fstat64(device->fd, &file_stat); ++ ret = fstat(device->fd, &file_stat); + if (ret < 0) { + error("failed to stat devid %llu: %m", device->devid); + return ret; + } + if (!S_ISREG(file_stat.st_mode)) + return ret; +- ret = ftruncate64(device->fd, new_size); ++ ret = ftruncate(device->fd, new_size); + if (ret < 0) { + error("failed to truncate device file of devid %llu: %m", + device->devid); +--- a/tests/fsstress.c ++++ b/tests/fsstress.c +@@ -458,7 +458,7 @@ int get_fname(int, long, pathname_t *, f + void init_pathname(pathname_t *); + int lchown_path(pathname_t *, uid_t, gid_t); + int link_path(pathname_t *, pathname_t *); +-int lstat64_path(pathname_t *, struct stat64 *); ++int lstat_path(pathname_t *, struct stat *); + void make_freq_table(void); + int mkdir_path(pathname_t *, mode_t); + int mknod_path(pathname_t *, mode_t, dev_t); +@@ -472,9 +472,9 @@ int rename_path(pathname_t *, pathname_t + int rmdir_path(pathname_t *); + void separate_pathname(pathname_t *, char *, pathname_t *); + void show_ops(int, char *); +-int stat64_path(pathname_t *, struct stat64 *); ++int stat_path(pathname_t *, struct stat *); + int symlink_path(const char *, pathname_t *); +-int truncate64_path(pathname_t *, off64_t); ++int truncate_path(pathname_t *, off64_t); + int unlink_path(pathname_t *); + void usage(void); + void write_freq(void); +@@ -998,12 +998,12 @@ void + check_cwd(void) + { + #ifdef DEBUG +- struct stat64 statbuf; ++ struct stat statbuf; + int ret; + +- ret = stat64(".", &statbuf); ++ ret = stat(".", &statbuf); + if (ret != 0) { +- fprintf(stderr, "fsstress: check_cwd stat64() returned %d with errno: %d (%m)\n", ++ fprintf(stderr, "fsstress: check_cwd stat() returned %d with errno: %d (%m)\n", + ret, errno); + goto out; + } +@@ -1171,7 +1171,7 @@ again: + void + doproc(void) + { +- struct stat64 statbuf; ++ struct stat statbuf; + char buf[10]; + char cmd[64]; + opnum_t opno; +@@ -1182,7 +1182,7 @@ doproc(void) + dividend = (operations + execute_freq) / (execute_freq + 1); + sprintf(buf, "p%x", procid); + (void)mkdir(buf, 0777); +- if (chdir(buf) < 0 || stat64(".", &statbuf) < 0) { ++ if (chdir(buf) < 0 || stat(".", &statbuf) < 0) { + perror(buf); + _exit(1); + } +@@ -1214,7 +1214,7 @@ doproc(void) + * the forced shutdown happened. + */ + if (errtag != 0 && opno % 100 == 0) { +- rval = stat64(".", &statbuf); ++ rval = stat(".", &statbuf); + if (rval == EIO) { + fprintf(stderr, "Detected EIO\n"); + goto errout; +@@ -1537,18 +1537,18 @@ link_path(pathname_t *name1, pathname_t + } + + int +-lstat64_path(pathname_t *name, struct stat64 *sbuf) ++lstat_path(pathname_t *name, struct stat *sbuf) + { + char buf[NAME_MAX + 1]; + pathname_t newname; + int rval; + +- rval = lstat64(name->path, sbuf); ++ rval = lstat(name->path, sbuf); + if (rval >= 0 || errno != ENAMETOOLONG) + return rval; + separate_pathname(name, buf, &newname); + if (chdir(buf) == 0) { +- rval = lstat64_path(&newname, sbuf); ++ rval = lstat_path(&newname, sbuf); + assert(chdir("..") == 0); + } + free_pathname(&newname); +@@ -1870,18 +1870,18 @@ show_ops(int flag, char *lead_str) + } + + int +-stat64_path(pathname_t *name, struct stat64 *sbuf) ++stat_path(pathname_t *name, struct stat *sbuf) + { + char buf[NAME_MAX + 1]; + pathname_t newname; + int rval; + +- rval = stat64(name->path, sbuf); ++ rval = stat(name->path, sbuf); + if (rval >= 0 || errno != ENAMETOOLONG) + return rval; + separate_pathname(name, buf, &newname); + if (chdir(buf) == 0) { +- rval = stat64_path(&newname, sbuf); ++ rval = stat_path(&newname, sbuf); + assert(chdir("..") == 0); + } + free_pathname(&newname); +@@ -1913,18 +1913,18 @@ symlink_path(const char *name1, pathname + } + + int +-truncate64_path(pathname_t *name, off64_t length) ++truncate_path(pathname_t *name, off64_t length) + { + char buf[NAME_MAX + 1]; + pathname_t newname; + int rval; + +- rval = truncate64(name->path, length); ++ rval = truncate(name->path, length); + if (rval >= 0 || errno != ENAMETOOLONG) + return rval; + separate_pathname(name, buf, &newname); + if (chdir(buf) == 0) { +- rval = truncate64_path(&newname, length); ++ rval = truncate_path(&newname, length); + assert(chdir("..") == 0); + } + free_pathname(&newname); +@@ -2026,7 +2026,7 @@ non_btrfs_freq(const char *path) + ops[btrfs_ops[i]].freq = 0; + } + +-void inode_info(char *str, size_t sz, struct stat64 *s, int verbose) ++void inode_info(char *str, size_t sz, struct stat *s, int verbose) + { + if (verbose) + snprintf(str, sz, "[%ld %ld %d %d %lld %lld]", +@@ -2101,7 +2101,7 @@ allocsp_f(opnum_t opno, long r) + struct xfs_flock64 fl; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + +@@ -2122,9 +2122,9 @@ allocsp_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: allocsp - fstat64 %s failed %d\n", ++ printf("%d/%lld: allocsp - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -2160,7 +2160,7 @@ do_aio_rw(opnum_t opno, long r, int flag + size_t len; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + char *dio_env; +@@ -2184,9 +2184,9 @@ do_aio_rw(opnum_t opno, long r, int flag + procid, opno, f.path, e); + goto aio_out; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: do_aio_rw - fstat64 %s failed %d\n", ++ printf("%d/%lld: do_aio_rw - fstat %s failed %d\n", + procid, opno, f.path, errno); + goto aio_out; + } +@@ -2278,7 +2278,7 @@ do_uring_rw(opnum_t opno, long r, int fl + size_t len; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + struct io_uring_sqe *sqe; +@@ -2304,9 +2304,9 @@ do_uring_rw(opnum_t opno, long r, int fl + procid, opno, f.path, e); + goto uring_out; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: do_uring_rw - fstat64 %s failed %d\n", ++ printf("%d/%lld: do_uring_rw - fstat %s failed %d\n", + procid, opno, f.path, errno); + goto uring_out; + } +@@ -2522,7 +2522,7 @@ bulkstat1_f(opnum_t opno, long r) + int fd; + int good; + __u64 ino; +- struct stat64 s; ++ struct stat s; + struct xfs_bstat t; + int v; + struct xfs_fsop_bulkreq bsr; +@@ -2534,7 +2534,7 @@ bulkstat1_f(opnum_t opno, long r) + init_pathname(&f); + if (!get_fname(FT_ANYm, r, &f, NULL, NULL, &v)) + append_pathname(&f, "."); +- ino = stat64_path(&f, &s) < 0 ? (ino64_t)r : s.st_ino; ++ ino = stat_path(&f, &s) < 0 ? (ino64_t)r : s.st_ino; + check_cwd(); + free_pathname(&f); + } else { +@@ -2605,8 +2605,8 @@ clonerange_f( + struct file_clone_range fcr; + struct pathname fpath1; + struct pathname fpath2; +- struct stat64 stat1; +- struct stat64 stat2; ++ struct stat stat1; ++ struct stat stat2; + char inoinfo1[1024]; + char inoinfo2[1024]; + off64_t lr; +@@ -2660,17 +2660,17 @@ clonerange_f( + } + + /* Get file stats */ +- if (fstat64(fd1, &stat1) < 0) { ++ if (fstat(fd1, &stat1) < 0) { + if (v1) +- printf("%d/%lld: clonerange read - fstat64 %s failed %d\n", ++ printf("%d/%lld: clonerange read - fstat %s failed %d\n", + procid, opno, fpath1.path, errno); + goto out_fd2; + } + inode_info(inoinfo1, sizeof(inoinfo1), &stat1, v1); + +- if (fstat64(fd2, &stat2) < 0) { ++ if (fstat(fd2, &stat2) < 0) { + if (v2) +- printf("%d/%lld: clonerange write - fstat64 %s failed %d\n", ++ printf("%d/%lld: clonerange write - fstat %s failed %d\n", + procid, opno, fpath2.path, errno); + goto out_fd2; + } +@@ -2743,8 +2743,8 @@ copyrange_f( + #ifdef HAVE_COPY_FILE_RANGE + struct pathname fpath1; + struct pathname fpath2; +- struct stat64 stat1; +- struct stat64 stat2; ++ struct stat stat1; ++ struct stat stat2; + char inoinfo1[1024]; + char inoinfo2[1024]; + loff_t lr; +@@ -2802,17 +2802,17 @@ copyrange_f( + } + + /* Get file stats */ +- if (fstat64(fd1, &stat1) < 0) { ++ if (fstat(fd1, &stat1) < 0) { + if (v1) +- printf("%d/%lld: copyrange read - fstat64 %s failed %d\n", ++ printf("%d/%lld: copyrange read - fstat %s failed %d\n", + procid, opno, fpath1.path, errno); + goto out_fd2; + } + inode_info(inoinfo1, sizeof(inoinfo1), &stat1, v1); + +- if (fstat64(fd2, &stat2) < 0) { ++ if (fstat(fd2, &stat2) < 0) { + if (v2) +- printf("%d/%lld: copyrange write - fstat64 %s failed %d\n", ++ printf("%d/%lld: copyrange write - fstat %s failed %d\n", + procid, opno, fpath2.path, errno); + goto out_fd2; + } +@@ -2900,7 +2900,7 @@ deduperange_f( + #define INFO_SZ 1024 + struct file_dedupe_range *fdr; + struct pathname *fpath; +- struct stat64 *stat; ++ struct stat *stat; + char *info; + off64_t *off; + int *v; +@@ -2938,7 +2938,7 @@ deduperange_f( + goto out_fdr; + } + +- stat = calloc(nr, sizeof(struct stat64)); ++ stat = calloc(nr, sizeof(struct stat)); + if (!stat) { + printf("%d/%lld: line %d error %d\n", + procid, opno, __LINE__, errno); +@@ -3017,9 +3017,9 @@ deduperange_f( + } + + /* Get file stats */ +- if (fstat64(fd[0], &stat[0]) < 0) { ++ if (fstat(fd[0], &stat[0]) < 0) { + if (v[0]) +- printf("%d/%lld: deduperange read - fstat64 %s failed %d\n", ++ printf("%d/%lld: deduperange read - fstat %s failed %d\n", + procid, opno, fpath[0].path, errno); + goto out_fds; + } +@@ -3027,9 +3027,9 @@ deduperange_f( + inode_info(&info[0], INFO_SZ, &stat[0], v[0]); + + for (i = 1; i < nr; i++) { +- if (fstat64(fd[i], &stat[i]) < 0) { ++ if (fstat(fd[i], &stat[i]) < 0) { + if (v[i]) +- printf("%d/%lld: deduperange write - fstat64 %s failed %d\n", ++ printf("%d/%lld: deduperange write - fstat %s failed %d\n", + procid, opno, fpath[i].path, errno); + goto out_fds; + } +@@ -3179,8 +3179,8 @@ splice_f(opnum_t opno, long r) + { + struct pathname fpath1; + struct pathname fpath2; +- struct stat64 stat1; +- struct stat64 stat2; ++ struct stat stat1; ++ struct stat stat2; + char inoinfo1[1024]; + char inoinfo2[1024]; + loff_t lr; +@@ -3237,17 +3237,17 @@ splice_f(opnum_t opno, long r) + } + + /* Get file stats */ +- if (fstat64(fd1, &stat1) < 0) { ++ if (fstat(fd1, &stat1) < 0) { + if (v1) +- printf("%d/%lld: splice read - fstat64 %s failed %d\n", ++ printf("%d/%lld: splice read - fstat %s failed %d\n", + procid, opno, fpath1.path, errno); + goto out_fd2; + } + inode_info(inoinfo1, sizeof(inoinfo1), &stat1, v1); + +- if (fstat64(fd2, &stat2) < 0) { ++ if (fstat(fd2, &stat2) < 0) { + if (v2) +- printf("%d/%lld: splice write - fstat64 %s failed %d\n", ++ printf("%d/%lld: splice write - fstat %s failed %d\n", + procid, opno, fpath2.path, errno); + goto out_fd2; + } +@@ -3432,7 +3432,7 @@ dread_f(opnum_t opno, long r) + size_t len; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + char *dio_env; +@@ -3454,9 +3454,9 @@ dread_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: dread - fstat64 %s failed %d\n", ++ printf("%d/%lld: dread - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -3522,7 +3522,7 @@ dwrite_f(opnum_t opno, long r) + size_t len; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + char *dio_env; +@@ -3544,9 +3544,9 @@ dwrite_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: dwrite - fstat64 %s failed %d\n", ++ printf("%d/%lld: dwrite - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -3620,7 +3620,7 @@ do_fallocate(opnum_t opno, long r, int m + int64_t lr; + off64_t off; + off64_t len; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + +@@ -3640,9 +3640,9 @@ do_fallocate(opnum_t opno, long r, int m + return; + } + check_cwd(); +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: do_fallocate - fstat64 %s failed %d\n", ++ printf("%d/%lld: do_fallocate - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -3734,7 +3734,7 @@ fiemap_f(opnum_t opno, long r) + int fd; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + int blocks_to_map; +@@ -3757,9 +3757,9 @@ fiemap_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: fiemap - fstat64 %s failed %d\n", ++ printf("%d/%lld: fiemap - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -3807,7 +3807,7 @@ freesp_f(opnum_t opno, long r) + struct xfs_flock64 fl; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + +@@ -3828,9 +3828,9 @@ freesp_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: freesp - fstat64 %s failed %d\n", ++ printf("%d/%lld: freesp - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -4226,7 +4226,7 @@ do_mmap(opnum_t opno, long r, int prot) + int64_t lr; + off64_t off; + int flags; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + sigjmp_buf sigbus_jmpbuf; +@@ -4248,9 +4248,9 @@ do_mmap(opnum_t opno, long r, int prot) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: do_mmap - fstat64 %s failed %d\n", ++ printf("%d/%lld: do_mmap - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -4370,7 +4370,7 @@ read_f(opnum_t opno, long r) + size_t len; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + +@@ -4391,9 +4391,9 @@ read_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: read - fstat64 %s failed %d\n", ++ printf("%d/%lld: read - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -4454,7 +4454,7 @@ readv_f(opnum_t opno, long r) + size_t len; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + struct iovec *iov = NULL; +@@ -4480,9 +4480,9 @@ readv_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: readv - fstat64 %s failed %d\n", ++ printf("%d/%lld: readv - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -4739,7 +4739,7 @@ resvsp_f(opnum_t opno, long r) + struct xfs_flock64 fl; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + +@@ -4760,9 +4760,9 @@ resvsp_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: resvsp - fstat64 %s failed %d\n", ++ printf("%d/%lld: resvsp - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -4971,7 +4971,7 @@ stat_f(opnum_t opno, long r) + { + int e; + pathname_t f; +- struct stat64 stb; ++ struct stat stb; + int v; + + init_pathname(&f); +@@ -4981,7 +4981,7 @@ stat_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- e = lstat64_path(&f, &stb) < 0 ? errno : 0; ++ e = lstat_path(&f, &stb) < 0 ? errno : 0; + check_cwd(); + if (v) + printf("%d/%lld: stat %s %d\n", procid, opno, f.path, e); +@@ -5133,7 +5133,7 @@ truncate_f(opnum_t opno, long r) + pathname_t f; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + +@@ -5144,11 +5144,11 @@ truncate_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- e = stat64_path(&f, &stb) < 0 ? errno : 0; ++ e = stat_path(&f, &stb) < 0 ? errno : 0; + check_cwd(); + if (e > 0) { + if (v) +- printf("%d/%lld: truncate - stat64 %s failed %d\n", ++ printf("%d/%lld: truncate - stat %s failed %d\n", + procid, opno, f.path, e); + free_pathname(&f); + return; +@@ -5157,7 +5157,7 @@ truncate_f(opnum_t opno, long r) + lr = ((int64_t)random() << 32) + random(); + off = (off64_t)(lr % MIN(stb.st_size + (1024 * 1024), MAXFSIZE)); + off %= maxfsize; +- e = truncate64_path(&f, off) < 0 ? errno : 0; ++ e = truncate_path(&f, off) < 0 ? errno : 0; + check_cwd(); + if (v) + printf("%d/%lld: truncate %s%s %lld %d\n", procid, opno, f.path, +@@ -5209,7 +5209,7 @@ unresvsp_f(opnum_t opno, long r) + struct xfs_flock64 fl; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + +@@ -5230,9 +5230,9 @@ unresvsp_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: unresvsp - fstat64 %s failed %d\n", ++ printf("%d/%lld: unresvsp - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -5281,7 +5281,7 @@ write_f(opnum_t opno, long r) + size_t len; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + +@@ -5302,9 +5302,9 @@ write_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: write - fstat64 %s failed %d\n", ++ printf("%d/%lld: write - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +@@ -5337,7 +5337,7 @@ writev_f(opnum_t opno, long r) + size_t len; + int64_t lr; + off64_t off; +- struct stat64 stb; ++ struct stat stb; + int v; + char st[1024]; + struct iovec *iov = NULL; +@@ -5363,9 +5363,9 @@ writev_f(opnum_t opno, long r) + free_pathname(&f); + return; + } +- if (fstat64(fd, &stb) < 0) { ++ if (fstat(fd, &stb) < 0) { + if (v) +- printf("%d/%lld: writev - fstat64 %s failed %d\n", ++ printf("%d/%lld: writev - fstat %s failed %d\n", + procid, opno, f.path, errno); + free_pathname(&f); + close(fd); +--- a/tests/fssum.c ++++ b/tests/fssum.c +@@ -519,9 +519,9 @@ sum(int dirfd, int level, sum_t *dircs, + int excl; + sum_file_data_t sum_file_data = flags[FLAG_STRUCTURE] ? + sum_file_data_strict : sum_file_data_permissive; +- struct stat64 dir_st; ++ struct stat dir_st; + +- if (fstat64(dirfd, &dir_st)) { ++ if (fstat(dirfd, &dir_st)) { + perror("fstat"); + exit(-1); + } +@@ -552,7 +552,7 @@ sum(int dirfd, int level, sum_t *dircs, + } + qsort(namelist, entries, sizeof(*namelist), namecmp); + for (i = 0; i < entries; ++i) { +- struct stat64 st; ++ struct stat st; + sum_t cs; + sum_t meta; + char *path; +@@ -572,7 +572,7 @@ sum(int dirfd, int level, sum_t *dircs, + perror("fchdir"); + exit(-1); + } +- ret = lstat64(namelist[i], &st); ++ ret = lstat(namelist[i], &st); + if (ret) { + fprintf(stderr, "stat failed for %s/%s: %m\n", + path_prefix, path);
meta/recipes-devtools/btrfs-tools/btrfs-tools_6.0.1.bb +1 -0
@@ -17,6 +17,7 @@ DEPENDS = "util-linux zlib" SRC_URI = "git://git.kernel.org/pub/scm/linux/kernel/git/kdave/btrfs-progs.git;branch=master \ file://0001-Add-a-possibility-to-specify-where-python-modules-ar.patch \ + file://0001-Use-pread-pwrite-ftruncate-stat-instead-of-64bit-equ.patch \ " SRCREV = "441d01556873385d55fd4940f50ee7ae1fcfb13d" S = "${WORKDIR}/git"
Signature · 1 line
2.38.1
Re: [OE-core] [PATCH 3/6] btrfs-tools: Do not use 64bit functions for largefile support
Quoted 1 line · Khem Raj
+Upstream-Status: Pending
Please submit this and all other patches in this set upstream first.
Come on Khem, you should expect this response, and I should not have to tell you.
Alex
A message that is not in the archive (sent off-list or not yet received): [email protected]
Re: [OE-core] [PATCH 3/6] btrfs-tools: Do not use 64bit functions for largefile support
I want to once again request that the whole set is not merged into master until every patch has an upstream submission link. It's fine to test them in master-next, but please do not merge.
Alex
Quoted 19 lines · Alexander Kanavin via lists.openembedded.org
Quoted 1 line · Khem Raj
+Upstream-Status: Pending
Please submit this and all other patches in this set upstream first.
Come on Khem, you should expect this response, and I should not have to tell you.
Alex
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#174208): https://lists.openembedded.org/g/openembedded-core/message/174208
Mute This Topic: https://lists.openembedded.org/mt/95389895/1686489
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
Re: [OE-core] [PATCH 3/6] btrfs-tools: Do not use 64bit functions for largefile support
Quoted 3 lines · Alexander Kanavin
I want to once again request that the whole set is not merged into master until every patch has an upstream submission link. It's fine to test them in master-next, but please do not merge.
I did a quick check and the acl patch was merged so I think that is ok to go in. The pseudo one doesn't need any further work either.
I'm worried the mdadm upstream isn't very active so that one, whilst submitted, may take a while.
I do agree we need to at least get the others submitted. OE isn't in the business of carrying patches, particularly when this is near enough musl feature development :/
Cheers,
Richard
Re: [OE-core] [PATCH 3/6] btrfs-tools: Do not use 64bit functions for largefile support
Quoted 2 lines · Richard Purdie
I'm worried the mdadm upstream isn't very active so that one, whilst submitted, may take a while.
There's a steady amount of new commits here:
https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/
so I'd send directly to the person committing them (Jes Sorensen <[email protected]>), and cc: maillist.
Alex
PATCH 4/6 e2fsprogs: Do not use 64bit functions for largefile support
In reply to Khem RajUsing -D_FILE_OFFSET_BITS=64 already makes the normal function behave like 64bit variants. Moreover it makes them portable across libc
--- ...-to-enable-disable-largefile-support.patch | 57 +++++++++++++++++++ .../e2fsprogs/e2fsprogs_1.46.5.bb | 3 +- 2 files changed, 59 insertions(+), 1 deletion(-) create mode 100644 meta/recipes-devtools/e2fsprogs/e2fsprogs/0001-Add-option-to-enable-disable-largefile-support.patch
meta/recipes-devtools/e2fsprogs/e2fsprogs/0001-Add-option-to-enable-disable-largefile-support.patch new file +57 -0
@@ -0,0 +1,57 @@ +From 6fab3346d448298a24cee1faeb5fc8507a3b9712 Mon Sep 17 00:00:00 2001 +From: Khem Raj <[email protected]> +Date: Thu, 10 Nov 2022 20:34:54 -0800 +Subject: [PATCH] Add option to enable/disable largefile support + +fallocate can be used to have 64bit off_t provided its compiled with +_FILE_OFFSET_BITS=64 which will be added automatically when +--enable-largefile is used. + +Upstream-Status: Pending +Signed-off-by: Khem Raj <[email protected]> +--- + configure.ac | 3 +++ + misc/e4defrag.c | 6 +----- + 2 files changed, 4 insertions(+), 5 deletions(-) + +diff --git a/configure.ac b/configure.ac +index dff3d1ca..ec1c5b64 100644 +--- a/configure.ac ++++ b/configure.ac +@@ -1800,6 +1800,9 @@ OS_IO_FILE="" + ;; + esac] + AC_SUBST(OS_IO_FILE) ++ ++AC_SYS_LARGEFILE ++ + dnl + dnl Make our output files, being sure that we create the some miscellaneous + dnl directories +diff --git a/misc/e4defrag.c b/misc/e4defrag.c +index 86e97ee3..e7175c37 100644 +--- a/misc/e4defrag.c ++++ b/misc/e4defrag.c +@@ -195,10 +195,6 @@ static struct frag_statistic_ino frag_rank[SHOW_FRAG_FILES]; + #error posix_fadvise not available! + #endif + +-#ifndef HAVE_FALLOCATE64 +-#error fallocate64 not available! +-#endif /* ! HAVE_FALLOCATE64 */ +- + /* + * get_mount_point() - Get device's mount point. + * +@@ -1568,7 +1564,7 @@ static int file_defrag(const char *file, const struct stat64 *buf, + /* Allocate space for donor inode */ + orig_group_tmp = orig_group_head; + do { +- ret = fallocate64(donor_fd, 0, ++ ret = fallocate(donor_fd, 0, + (ext2_loff_t)orig_group_tmp->start->data.logical * block_size, + (ext2_loff_t)orig_group_tmp->len * block_size); + if (ret < 0) { +-- +2.38.1 +
meta/recipes-devtools/e2fsprogs/e2fsprogs_1.46.5.bb +2 -1
@@ -5,6 +5,7 @@ SRC_URI += "file://remove.ldconfig.call.patch \ file://ptest.patch \ file://mkdir_p.patch \ file://extents.patch \ + file://0001-Add-option-to-enable-disable-largefile-support.patch \ " SRC_URI:append:class-native = " \ file://e2fsprogs-fix-missing-check-for-permission-denied.patch \
@@ -17,7 +18,7 @@ UPSTREAM_CHECK_GITTAGREGEX = "v(?P<pver>\d+\.\d+(\.\d+)*)$" EXTRA_OECONF += "--libdir=${base_libdir} --sbindir=${base_sbindir} \ --enable-elf-shlibs --disable-libuuid --disable-uuidd \ --disable-libblkid --enable-verbose-makecmds \ - --with-crond-dir=no" + --enable-largefile --with-crond-dir=no" EXTRA_OECONF:darwin = "--libdir=${base_libdir} --sbindir=${base_sbindir} --enable-bsd-shlibs"
Signature · 1 line
2.38.1
PATCH 5/6 libpciaccess: Do not use 64bit functions for largefile support
In reply to Khem RajUsing -D_FILE_OFFSET_BITS=64 already makes the normal function behave like 64bit variants. Moreover it makes them portable across libc
--- ...pwrite-pread-instead-of-64bit-versio.patch | 39 +++++++++++++++++++ .../xorg-lib/libpciaccess_0.17.bb | 2 + 2 files changed, 41 insertions(+) create mode 100644 meta/recipes-graphics/xorg-lib/libpciaccess/0001-linux_sysfs-Use-pwrite-pread-instead-of-64bit-versio.patch
meta/recipes-graphics/xorg-lib/libpciaccess/0001-linux_sysfs-Use-pwrite-pread-instead-of-64bit-versio.patch new file +39 -0
@@ -0,0 +1,39 @@ +From d20ff4f2e4feaafa6a6191b9cae3815b85617943 Mon Sep 17 00:00:00 2001 +From: Khem Raj <[email protected]> +Date: Fri, 11 Nov 2022 11:15:58 -0800 +Subject: [PATCH] linux_sysfs: Use pwrite/pread instead of 64bit versions + +pread64/pwrite64 are aliased to pread/pwrite when largefile support is +enabled e.g. using _FILE_OFFSET_BITS=64 macro + +Upstream-Status: Pending +Signed-off-by: Khem Raj <[email protected]> +--- + src/linux_sysfs.c | 4 ++-- + 1 file changed, 2 insertions(+), 2 deletions(-) + +diff --git a/src/linux_sysfs.c b/src/linux_sysfs.c +index d62f48c..20c800e 100644 +--- a/src/linux_sysfs.c ++++ b/src/linux_sysfs.c +@@ -462,7 +462,7 @@ pci_device_linux_sysfs_read( struct pci_device * dev, void * data, + + + while ( temp_size > 0 ) { +- const ssize_t bytes = pread64( fd, data_bytes, temp_size, offset ); ++ const ssize_t bytes = pread( fd, data_bytes, temp_size, offset ); + + /* If zero bytes were read, then we assume it's the end of the + * config file. +@@ -522,7 +522,7 @@ pci_device_linux_sysfs_write( struct pci_device * dev, const void * data, + + + while ( temp_size > 0 ) { +- const ssize_t bytes = pwrite64( fd, data_bytes, temp_size, offset ); ++ const ssize_t bytes = pwrite( fd, data_bytes, temp_size, offset ); + + /* If zero bytes were written, then we assume it's the end of the + * config file. +-- +2.38.1 +
meta/recipes-graphics/xorg-lib/libpciaccess_0.17.bb +2 -0
@@ -5,6 +5,8 @@ PCI bus and devices in a platform-independent way." require xorg-lib-common.inc +SRC_URI += "file://0001-linux_sysfs-Use-pwrite-pread-instead-of-64bit-versio.patch" + XORG_EXT = "tar.xz" SRC_URI[sha256sum] = "74283ba3c974913029e7a547496a29145b07ec51732bbb5b5c58d5025ad95b73"
Signature · 1 line
2.38.1
PATCH 6/6 pseudo: Disable LFS on 32bit arches
In reply to Khem Rajpseudo overrides certain libc functions which are aliases when LFS64 is enabled. In anycase pseudo may not be much used on 32bit systems anyway
--- meta/recipes-devtools/pseudo/pseudo_git.bb | 2 ++ 1 file changed, 2 insertions(+)
meta/recipes-devtools/pseudo/pseudo_git.bb +2 -0
@@ -17,5 +17,7 @@ SRCREV = "c9670c27ff67ab899007ce749254b16091577e55" S = "${WORKDIR}/git" PV = "1.9.0+git${SRCPV}" +TARGET_CPPFLAGS:remove = "-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64" + # error: use of undeclared identifier '_STAT_VER' COMPATIBLE_HOST:libc-musl = 'null'
Signature · 1 line
2.38.1
Re: [OE-core] [PATCH 6/6] pseudo: Disable LFS on 32bit arches
Quoted 13 lines · Khem Raj
pseudo overrides certain libc functions which are aliases when LFS64 is enabled. In anycase pseudo may not be much used on 32bit systems anyway
Signed-off-by Khem Raj <[email protected]>--- meta/recipes-devtools/pseudo/pseudo_git.bb | 2 ++ 1 file changed, 2 insertions(+)meta/recipes-devtools/pseudo/pseudo_git.bb +0
-0
meta/recipes-devtools/pseudo/pseudo_git.bb SRCREV = "c9670c27ff67ab899007ce749254b16091577e55" · line 22
Quoted 6 lines
meta/recipes-devtools/pseudo/pseudo_git.bb +2
-0@@ -17,5 +17,7 @@ SRCREV = "c9670c27ff67ab899007ce749254b16091577e55" S = "${WORKDIR}/git" PV = "1.9.0+git${SRCPV}" +TARGET_CPPFLAGS:remove = "-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64" +
We need to put some comments in here about what this is doing and why.
Cheers,
Richard
Re: [OE-core] [PATCH 6/6] pseudo: Disable LFS on 32bit arches
Quoted 17 lines · Richard Purdie
Quoted 16 lines · Khem Raj
pseudo overrides certain libc functions which are aliases when LFS64 is enabled. In anycase pseudo may not be much used on 32bit systems anyway
Signed-off-by Khem Raj <[email protected]>--- meta/recipes-devtools/pseudo/pseudo_git.bb | 2 ++ 1 file changed, 2 insertions(+)meta/recipes-devtools/pseudo/pseudo_git.bb +0
-0@@ -17,5 +17,7 @@ SRCREV = "c9670c27ff67ab899007ce749254b16091577e55" S = "${WORKDIR}/git" PV = "1.9.0+git${SRCPV}"
Quoted 6 lines
meta/recipes-devtools/pseudo/pseudo_git.bb SRCREV = "c9670c27ff67ab899007ce749254b16091577e55"
Quoted 3 lines
meta/recipes-devtools/pseudo/pseudo_git.bb +2
-0+TARGET_CPPFLAGS:remove = "-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64" +We need to put some comments in here about what this is doing and why.
Note that this was one of the issues I had with the -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 build. My guess is that we also have to remove -D_TIME_BITS=64.
Signature · 3 lines
Alexandre Belloni, co-owner and COO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
Re: [OE-core] [PATCH 6/6] pseudo: Disable LFS on 32bit arches
Quoted 26 lines · Alexandre Belloni
Quoted 17 lines · Richard Purdie
Quoted 16 lines · Khem Raj
pseudo overrides certain libc functions which are aliases when LFS64 is enabled. In anycase pseudo may not be much used on 32bit systems anyway
Signed-off-by Khem Raj <[email protected]>--- meta/recipes-devtools/pseudo/pseudo_git.bb | 2 ++ 1 file changed, 2 insertions(+)meta/recipes-devtools/pseudo/pseudo_git.bb +0
-0@@ -17,5 +17,7 @@ SRCREV = "c9670c27ff67ab899007ce749254b16091577e55" S = "${WORKDIR}/git" PV = "1.9.0+git${SRCPV}"Quoted 6 lines
meta/recipes-devtools/pseudo/pseudo_git.bb SRCREV = "c9670c27ff67ab899007ce749254b16091577e55"
Quoted 3 lines
meta/recipes-devtools/pseudo/pseudo_git.bb +2
-0+TARGET_CPPFLAGS:remove = "-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64" +We need to put some comments in here about what this is doing and why.
Quoted 3 lines
Note that this was one of the issues I had with the -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 build. My guess is that we also have to remove -D_TIME_BITS=64.
Eventually yes, however lets do that incrementally. First I am trying to get LFS handled.
Quoted 6 lines
--
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
Re: [OE-core] [PATCH 6/6] pseudo: Disable LFS on 32bit arches
Quoted 28 lines · Khem Raj
Quoted 26 lines · Alexandre Belloni
Quoted 23 lines · Richard Purdie
Quoted 13 lines · Khem Raj
pseudo overrides certain libc functions which are aliases when LFS64 is enabled. In anycase pseudo may not be much used on 32bit systems anyway
Signed-off-by Khem Raj <[email protected]>--- meta/recipes-devtools/pseudo/pseudo_git.bb | 2 ++ 1 file changed, 2 insertions(+)meta/recipes-devtools/pseudo/pseudo_git.bb +0
-0meta/recipes-devtools/pseudo/pseudo_git.bb SRCREV = "c9670c27ff67ab899007ce749254b16091577e55" · line 22
Quoted 6 lines
meta/recipes-devtools/pseudo/pseudo_git.bb +2
-0@@ -17,5 +17,7 @@ SRCREV = "c9670c27ff67ab899007ce749254b16091577e55" S = "${WORKDIR}/git" PV = "1.9.0+git${SRCPV}" +TARGET_CPPFLAGS:remove = "-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64" +We need to put some comments in here about what this is doing and why.
Quoted 6 lines
Quoted 3 lines
Note that this was one of the issues I had with the -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 build. My guess is that we also have to remove -D_TIME_BITS=64.
Eventually yes, however lets do that incrementally. First I am trying to get LFS handled.
Pseudo is a bit of a special case so if we can see an issue here, lets add the other bit needed as well please! :)
I don't mind if it is one or two patches but we should do it.
Cheers,
Richard
Re: [OE-core] [PATCH 6/6] pseudo: Disable LFS on 32bit arches
Quoted 34 lines · Richard Purdie
Quoted 32 lines · Khem Raj
Quoted 29 lines · Alexandre Belloni
Quoted 17 lines · Richard Purdie
Quoted 16 lines · Khem Raj
pseudo overrides certain libc functions which are aliases when LFS64 is enabled. In anycase pseudo may not be much used on 32bit systems anyway
Signed-off-by Khem Raj <[email protected]>--- meta/recipes-devtools/pseudo/pseudo_git.bb | 2 ++ 1 file changed, 2 insertions(+)meta/recipes-devtools/pseudo/pseudo_git.bb +0
-0@@ -17,5 +17,7 @@ SRCREV = "c9670c27ff67ab899007ce749254b16091577e55" S = "${WORKDIR}/git" PV = "1.9.0+git${SRCPV}"Quoted 6 lines
meta/recipes-devtools/pseudo/pseudo_git.bb SRCREV = "c9670c27ff67ab899007ce749254b16091577e55" · line 22
Quoted 3 lines
meta/recipes-devtools/pseudo/pseudo_git.bb +2
-0+TARGET_CPPFLAGS:remove = "-D_LARGEFILE_SOURCE -D +We need to put some comments in here about what this is doing and why.
Note that this was one of the issues I had with the -D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 build. My guess is that we also have to remove -D_TIME_BITS=64.
Quoted 8 lines
Quoted 2 lines
Eventually yes, however lets do that incrementally. First I am trying to get LFS handled.
Pseudo is a bit of a special case so if we can see an issue here, lets add the other bit needed as well please! :)
I don't mind if it is one or two patches but we should do it.
OK, I wanted to first fail with that option because we are removing -D_FILE_OFFSET_BITS=64 it should be enough
for 64bit time_t as well.
Quoted 4 lines
Cheers,
Richard