Skip to contentSkip to conversation

f2fs-tools: prefer physical over logical block size

PATCH In review nl6720 via Linux-f2fs-devel 1 participant 5 messages Apr 25, 2025 – Apr 29, 2025 Library code

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 f2fs-tools: prefer physical over logical block size

Some drives operate in "512e" configuration with their logical block size set to 512 bytes for legacy compatibility reasons while providing a more optimal 4096 byte value as the physical block size.

Since the physical block size is the smallest unit a physical storage device can write atomically, prefer it over the logical block size.

---
 lib/libf2fs.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
lib/libf2fs.c +7 -2
index d2579d76a5c7eaf141dc8aad11c176eeadabad83..040ea4bc87288fa075cb4bf51df08b9db28a65b3 100644
@@ -933,10 +933,15 @@ void get_kernel_uname_version(__u8 *version)
 #define BLKSSZGET	_IO(0x12,104)
 #endif
 
+#if defined(__linux__) && defined(_IO) && !defined(BLKPBSZGET)
+#define BLKPBSZGET	_IO(0x12,123)
+#endif
+
 #if defined(__APPLE__)
 #include <sys/disk.h>
 #define BLKGETSIZE	DKIOCGETBLOCKCOUNT
 #define BLKSSZGET	DKIOCGETBLOCKCOUNT
+#define BLKPBSZGET	DKIOCGETBLOCKCOUNT
 #endif /* APPLE_DARWIN */
 
 #ifndef _WIN32
@@ -1050,8 +1055,8 @@ int get_device_info(int i)
 	} else if (S_ISREG(stat_buf->st_mode)) {
 		dev->total_sectors = stat_buf->st_size / dev->sector_size;
 	} else if (S_ISBLK(stat_buf->st_mode)) {
-#ifdef BLKSSZGET
-		if (ioctl(fd, BLKSSZGET, &sector_size) < 0)
+#if defined(BLKPBSZGET) && defined(BLKSSZGET)
+		if ((ioctl(fd, BLKPBSZGET, &sector_size) < 0) && (ioctl(fd, BLKSSZGET, &sector_size) < 0))
 			MSG(0, "\tError: Using the default sector size\n");
 		else if (dev->sector_size < sector_size)
 			dev->sector_size = sector_size;
Signature · 8 lines
2.49.0



_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
#1 lore raw
Quoted 8 lines · On 4/25/25 09:36, nl6720 via Linux-f2fs-devel wrote:

Some drives operate in "512e" configuration with their logical block size set to 512 bytes for legacy compatibility reasons while providing a more optimal 4096 byte value as the physical block size.

Since the physical block size is the smallest unit a physical storage device can write atomically, prefer it over the logical block size.

f2fs uses PAGE_SIZE as block size, commonly it's 4096 rather than 512?

Thanks,

Quoted 27 lines
---
 lib/libf2fs.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
lib/libf2fs.c +5 -0
index d2579d76a5c7eaf141dc8aad11c176eeadabad83..040ea4bc87288fa075cb4bf51df08b9db28a65b3 100644
@@ -933,10 +933,15 @@ void get_kernel_uname_version(__u8 *version)
 #define BLKSSZGET	_IO(0x12,104)
 #endif
 
+#if defined(__linux__) && defined(_IO) && !defined(BLKPBSZGET)
+#define BLKPBSZGET	_IO(0x12,123)
+#endif
+
 #if defined(__APPLE__)
 #include <sys/disk.h>
 #define BLKGETSIZE	DKIOCGETBLOCKCOUNT
 #define BLKSSZGET	DKIOCGETBLOCKCOUNT
+#define BLKPBSZGET	DKIOCGETBLOCKCOUNT
 #endif /* APPLE_DARWIN */
 
 #ifndef _WIN32

lib/libf2fs.c int get_device_info(int i) · line 1062

Quoted 10 lines
lib/libf2fs.c +2 -2
@@ -1050,8 +1055,8 @@ int get_device_info(int i)
 	} else if (S_ISREG(stat_buf->st_mode)) {
 		dev->total_sectors = stat_buf->st_size / dev->sector_size;
 	} else if (S_ISBLK(stat_buf->st_mode)) {
-#ifdef BLKSSZGET
-		if (ioctl(fd, BLKSSZGET, &sector_size) < 0)
+#if defined(BLKPBSZGET) && defined(BLKSSZGET)
+		if ((ioctl(fd, BLKPBSZGET, &sector_size) < 0) && (ioctl(fd, BLKSSZGET, &sector_size) < 0))
 			MSG(0, "\tError: Using the default sector size\n");
 		else if (dev->sector_size < sector_size)
 			dev->sector_size = sector_size;

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

#2 lore raw
Quoted 11 lines · On Sunday, 27 April 2025 11:45:59 EEST Chao Yu wrote:
Quoted 8 lines · On 4/25/25 09:36, nl6720 via Linux-f2fs-devel wrote:

Some drives operate in "512e" configuration with their logical block size set to 512 bytes for legacy compatibility reasons while providing a more optimal 4096 byte value as the physical block size.

Since the physical block size is the smallest unit a physical storage device can write atomically, prefer it over the logical block size.

f2fs uses PAGE_SIZE as block size, commonly it's 4096 rather than 512?

The thing mkfs.f2fs calls "sector size" in its output is 512 on 512e drives instead of 4096.
E.g.:

    F2FS-tools: mkfs.f2fs Ver: 1.16.0 (2023-04-11)

Info: Disable heap-based policy
Info: Debug level = 0
Info: Trim is enabled
Info: [/dev/disk/by-partlabel/512e] Disk Model: QEMU HARDDISK
Info: Segments per section = 1
Info: Sections per zone = 1
Info: sector size = 512
Info: total sectors = 124997632 (61034 MB)
Info: zone aligned segment0 blkaddr: 512
Info: format version with
"Linux version 6.13.8-arch1-1 (linux@archlinux) (gcc (GCC) 14.2.1 20250207, GNU ld (GNU Binutils) 2.44) #1 SMP PREEMPT_DYNAMIC Sun, 23 Mar 2025 17:17:30 +0000"
Info: [/dev/disk/by-partlabel/512e] Discarding device
Info: This device doesn't support BLKSECDISCARD
Info: Discarded 61034 MB
Info: Overprovision ratio = 0.590%
Info: Overprovision segments = 179 (GC reserved = 176)
Info: format successful

nl6720

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

#3 lore raw
Quoted 9 lines · On 4/27/25 17:30, nl6720 wrote:
Quoted 8 lines · On Sunday, 27 April 2025 11:45:59 EEST Chao Yu wrote:
Quoted 7 lines · On 4/25/25 09:36, nl6720 via Linux-f2fs-devel wrote:

Some drives operate in "512e" configuration with their logical block size set to 512 bytes for legacy compatibility reasons while providing a more optimal 4096 byte value as the physical block size.

Since the physical block size is the smallest unit a physical storage device can write atomically, prefer it over the logical block size.

Quoted 6 lines
Quoted 3 lines
Quoted 1 line

f2fs uses PAGE_SIZE as block size, commonly it's 4096 rather than 512?

The thing mkfs.f2fs calls "sector size" in its output is 512 on 512e drives instead of 4096.

How about printing block size of f2fs as below:

output of mkfs.f2fs:

Info: sector size = 512
Info: total sectors = 16777216 (8192 MB)
Info: block size = 4096

Thanks,

Quoted 23 lines

E.g.:

    F2FS-tools: mkfs.f2fs Ver: 1.16.0 (2023-04-11)

Info: Disable heap-based policy
Info: Debug level = 0
Info: Trim is enabled
Info: [/dev/disk/by-partlabel/512e] Disk Model: QEMU HARDDISK
Info: Segments per section = 1
Info: Sections per zone = 1
Info: sector size = 512
Info: total sectors = 124997632 (61034 MB)
Info: zone aligned segment0 blkaddr: 512
Info: format version with
"Linux version 6.13.8-arch1-1 (linux@archlinux) (gcc (GCC) 14.2.1 20250207, GNU ld (GNU Binutils) 2.44) #1 SMP PREEMPT_DYNAMIC Sun, 23 Mar 2025 17:17:30 +0000"
Info: [/dev/disk/by-partlabel/512e] Discarding device
Info: This device doesn't support BLKSECDISCARD
Info: Discarded 61034 MB
Info: Overprovision ratio = 0.590%
Info: Overprovision segments = 179 (GC reserved = 176)
Info: format successful

Quoted 3 lines

nl6720

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

#4 lore raw
Quoted 7 lines · On Monday, 28 April 2025 10:33:18 EEST Chao Yu wrote:

How about printing block size of f2fs as below:

output of mkfs.f2fs:

Info: sector size = 512
Info: total sectors = 16777216 (8192 MB)
Info: block size = 4096

mkfs.f2fs printing it would be nice.

I'm still confused by the "sector size" in mkfs.f2fs output. Is it simply the drive's logical sector size and not a fundamental part of the file system like with xfs?

nl6720

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel