Skip to contentSkip to conversation

fs/ufs: Use LUT for dir entry types

PATCH In review Luis Felipe Hernandez 3 participants 4 messages Jun 10, 2024 – Jun 19, 2024

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 fs/ufs: Use LUT for dir entry types

As per the original TODO, replacing the switch statement with a lookup table results in more concise mapping logic and improved performance.

ufs_mode_to_dt maps file mode bits to directory entry types. Indices are created by right shifting 12 bits to isolate the higher-order bits in order to use smaller indexing values.
e.g. S_IFSOCK >> FT_SHIFT => 0140000 >> 12 = 1100

---
 fs/ufs/util.h | 46 +++++++++++++++++++---------------------------
 1 file changed, 19 insertions(+), 27 deletions(-)
fs/ufs/util.h +19 -27
index 0ecd2ed792f5..caa843373191 100644
@@ -7,10 +7,13 @@
  * Charles University, Faculty of Mathematics and Physics
  */
 
+#include <linux/array_size.h>
 #include <linux/buffer_head.h>
 #include <linux/fs.h>
 #include "swab.h"
 
+#define FT_SHIFT 12
+
 /*
  * functions used for retyping
  */
@@ -146,40 +149,29 @@ ufs_set_de_namlen(struct super_block *sb, struct ufs_dir_entry *de, u16 value)
 		de->d_u.d_44.d_namlen = value; /* XXX this seems wrong */
 }
 
+static const unsigned char ufs_mode_to_dt[] = {
+	[S_IFSOCK >> FT_SHIFT] = DT_SOCK,
+	[S_IFLNK >> FT_SHIFT] = DT_LNK,
+	[S_IFREG >> FT_SHIFT] = DT_REG,
+	[S_IFBLK >> FT_SHIFT] = DT_BLK,
+	[S_IFDIR >> FT_SHIFT] = DT_DIR,
+	[S_IFCHR >> FT_SHIFT] = DT_CHR,
+	[S_IFIFO >> FT_SHIFT] = DT_FIFO,
+	[0] = DT_UNKNOWN
+};
+
 static inline void
 ufs_set_de_type(struct super_block *sb, struct ufs_dir_entry *de, int mode)
 {
 	if ((UFS_SB(sb)->s_flags & UFS_DE_MASK) != UFS_DE_44BSD)
 		return;
 
-	/*
-	 * TODO turn this into a table lookup
-	 */
-	switch (mode & S_IFMT) {
-	case S_IFSOCK:
-		de->d_u.d_44.d_type = DT_SOCK;
-		break;
-	case S_IFLNK:
-		de->d_u.d_44.d_type = DT_LNK;
-		break;
-	case S_IFREG:
-		de->d_u.d_44.d_type = DT_REG;
-		break;
-	case S_IFBLK:
-		de->d_u.d_44.d_type = DT_BLK;
-		break;
-	case S_IFDIR:
-		de->d_u.d_44.d_type = DT_DIR;
-		break;
-	case S_IFCHR:
-		de->d_u.d_44.d_type = DT_CHR;
-		break;
-	case S_IFIFO:
-		de->d_u.d_44.d_type = DT_FIFO;
-		break;
-	default:
+	unsigned int mode_index = (mode & S_IFMT) >> FT_SHIFT;
+
+	if (mode_index < ARRAY_SIZE(ufs_mode_to_dt))
+		de->d_u.d_44.d_type = ufs_mode_to_dt[mode_index];
+	else
 		de->d_u.d_44.d_type = DT_UNKNOWN;
-	}
 }
 
 static inline u32
Signature · 1 line
2.45.2
#1 lore raw
Quoted 2 lines

As per the original TODO, replacing the switch statement with a lookup table results in more concise mapping logic and improved performance.

…

Can imperative wordings be relevant for another improved change description?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.10-rc2#n94

…

Quoted 1 line

+++ b/fs/ufs/util.h

…

Quoted 2 lines

static inline void
ufs_set_de_type(struct super_block *sb, struct ufs_dir_entry *de, int mode)

…

fs/ufs/util.h ufs_set_de_namlen(struct super_block *sb, struct ufs_dir_entry *de, u16 value) · line 174

Quoted 4 lines
fs/ufs/util.h +3 -0
+	if (mode_index < ARRAY_SIZE(ufs_mode_to_dt))
+		de->d_u.d_44.d_type = ufs_mode_to_dt[mode_index];
+	else
 		de->d_u.d_44.d_type = DT_UNKNOWN;

…

May a conditional operator expression be applied at this source code place?

Regards,
Markus

#2 lore raw
Quoted 20 lines · Markus Elfring
Quoted 2 lines

As per the original TODO, replacing the switch statement with a lookup table results in more concise mapping logic and improved performance.

…

Can imperative wordings be relevant for another improved change description?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.10-rc2#n94

…

Quoted 1 line

+++ b/fs/ufs/util.h

…

Quoted 2 lines

static inline void
ufs_set_de_type(struct super_block *sb, struct ufs_dir_entry *de, int mode)

…

Quoted 4 lines
fs/ufs/util.h +3 -0
+	if (mode_index < ARRAY_SIZE(ufs_mode_to_dt))
+		de->d_u.d_44.d_type = ufs_mode_to_dt[mode_index];
+	else
 		de->d_u.d_44.d_type = DT_UNKNOWN;

…

Quoted 5 lines

May a conditional operator expression be applied at this source code place?

Regards,
Markus

Hi,

This is the semi-friendly patch-bot of Greg Kroah-Hartman.

Markus, you seem to have sent a nonsensical or otherwise pointless review comment to a patch submission on a Linux kernel developer mailing list. I strongly suggest that you not do this anymore. Please do not bother developers who are actively working to produce patches and features with comments that, in the end, are a waste of time.

Patch submitter, please ignore Markus's suggestion; you do not need to follow it at all. The person/bot/AI that sent it is being ignored by almost all Linux kernel maintainers for having a persistent pattern of behavior of producing distracting and pointless commentary, and inability to adapt to feedback. Please feel free to also ignore emails from them.

thanks,

greg k-h's patch email bot

#3 lore raw
Quoted 44 lines · Greg KH
Quoted 20 lines · Markus Elfring
Quoted 2 lines

As per the original TODO, replacing the switch statement with a lookup table results in more concise mapping logic and improved performance.

…

Can imperative wordings be relevant for another improved change description?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.10-rc2#n94

…

Quoted 1 line

+++ b/fs/ufs/util.h

…

Quoted 2 lines

static inline void
ufs_set_de_type(struct super_block *sb, struct ufs_dir_entry *de, int mode)

…

Quoted 4 lines
fs/ufs/util.h +3 -0
+	if (mode_index < ARRAY_SIZE(ufs_mode_to_dt))
+		de->d_u.d_44.d_type = ufs_mode_to_dt[mode_index];
+	else
 		de->d_u.d_44.d_type = DT_UNKNOWN;

…

Quoted 5 lines

May a conditional operator expression be applied at this source code place?

Regards,
Markus

Hi,

This is the semi-friendly patch-bot of Greg Kroah-Hartman.

Markus, you seem to have sent a nonsensical or otherwise pointless review comment to a patch submission on a Linux kernel developer mailing list. I strongly suggest that you not do this anymore. Please do not bother developers who are actively working to produce patches and features with comments that, in the end, are a waste of time.

Patch submitter, please ignore Markus's suggestion; you do not need to follow it at all. The person/bot/AI that sent it is being ignored by almost all Linux kernel maintainers for having a persistent pattern of behavior of producing distracting and pointless commentary, and inability to adapt to feedback. Please feel free to also ignore emails from them.

Quoted 3 lines

thanks,

greg k-h's patch email bot

Hi Evgeniy,

I am writing to follow up on my patch submission from June 9th. The patch addresses a TODO, "turn this (switch statement) into a table lookup", originally introduced on 2005-04-16. I believe the original commit message body may have not been written in the correct form and wanted to offer an improved version:

"Replace the switch statement with a lookup table (LUT) as suggested in the original TODO. This change results in more concise mapping logic and improves performance.

The ufs_mode_to_dt function now maps file mode bits to directory entry types using a LUT. Indices are created by right shifting the mode bits by 12 (FT_SHIFT), isolating the higher-order bits for indexing.

This enhancement simplifies the code and optimizes the mapping process."

Any feedback or guidance on the patch would be greatly appreciated.

Thank you for your time and consideration.

Best regards,

Felipe