Skip to contentSkip to conversation

kconfig: Print reverse dependencies on new line consistently

PATCH Superseded Eugeniu Rosca 3 participants 14 messages Feb 13, 2018 – Feb 18, 2018

Subsystems (2)

v1 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/2 kconfig: Print reverse dependencies on new line consistently

From: Eugeniu Rosca <[email protected]>

Commit 1ccb27143360 ("kconfig: make "Selected by:" and "Implied by:" readable") made an incredible improvement in how reverse dependencies are perceived by the user, by breaking down the single (often interminable) expression string into small readable chunks, each of them displayed on a separate line:

Selected by:
- A & B
- C & (D || E)

Unfortunately, what happens with the non-OR (either E_SYMBOL or E_AND) expressions is that they don't get a dedicated line:

Selected by: F & G

That's arguably a bug/misbehavior, but it makes the implementation of something like below more complicated:

Selected by:
- [m] F & G /* where (F & G) evaluates to '=m' */

Adding '[m]', '[y]' or '[ ]' to the left side of each reverse dependency is subject of a different commit. The purpose of current commit is to print the 'Selected by:' and 'Implied by:' expressions on a separate line _consistently_.

An example of change contributed by this commit is seen below.

│ Symbol: NEED_SG_DMA_LENGTH [=y]
│ ...
│   Selected by: IOMMU_DMA [=y] && IOMMU_SUPPORT [=y]

becomes:

│ Symbol: NEED_SG_DMA_LENGTH [=y]
│ ...
│   Selected by:
│   - IOMMU_DMA [=y] && IOMMU_SUPPORT [=y]

This patch has been tested using a custom variant of zconfdump which prints the reverse dependencies for each config symbol.

---
 scripts/kconfig/expr.c | 31 ++++++++++++++++++++++---------
 1 file changed, 22 insertions(+), 9 deletions(-)
scripts/kconfig/expr.c +22 -9
index d45381986ac7..0704bcdf4c78 100644
@@ -1179,6 +1179,16 @@ struct expr *expr_simplify_unmet_dep(struct expr *e1, struct expr *e2)
 	return expr_get_leftmost_symbol(ret);
 }
 
+static void
+expr_print_newline(struct expr *e,
+		   void (*fn)(void *, struct symbol *, const char *),
+		   void *data,
+		   int prevtoken)
+{
+	fn(data, NULL, "\n  - ");
+	expr_print(e, fn, data, prevtoken);
+}
+
 static void __expr_print(struct expr *e, void (*fn)(void *, struct symbol *, const char *), void *data, int prevtoken, bool revdep)
 {
 	if (!e) {
@@ -1191,7 +1201,10 @@ static void __expr_print(struct expr *e, void (*fn)(void *, struct symbol *, con
 	switch (e->type) {
 	case E_SYMBOL:
 		if (e->left.sym->name)
-			fn(data, e->left.sym, e->left.sym->name);
+			if (revdep)
+				expr_print_newline(e, fn, data, E_OR);
+			else
+				fn(data, e->left.sym, e->left.sym->name);
 		else
 			fn(data, NULL, "<choice>");
 		break;
@@ -1234,19 +1247,19 @@ static void __expr_print(struct expr *e, void (*fn)(void *, struct symbol *, con
 		fn(data, e->right.sym, e->right.sym->name);
 		break;
 	case E_OR:
-		if (revdep && e->left.expr->type != E_OR)
-			fn(data, NULL, "\n  - ");
 		__expr_print(e->left.expr, fn, data, E_OR, revdep);
-		if (revdep)
-			fn(data, NULL, "\n  - ");
-		else
+		if (!revdep)
 			fn(data, NULL, " || ");
 		__expr_print(e->right.expr, fn, data, E_OR, revdep);
 		break;
 	case E_AND:
-		expr_print(e->left.expr, fn, data, E_AND);
-		fn(data, NULL, " && ");
-		expr_print(e->right.expr, fn, data, E_AND);
+		if (revdep) {
+			expr_print_newline(e, fn, data, E_OR);
+		} else {
+			expr_print(e->left.expr, fn, data, E_AND);
+			fn(data, NULL, " && ");
+			expr_print(e->right.expr, fn, data, E_AND);
+		}
 		break;
 	case E_LIST:
 		fn(data, e->right.sym, e->right.sym->name);
Signature · 1 line
2.16.1
#1 lore raw

PATCH 2/2 kconfig: Print the value of each reverse dependency

From: Eugeniu Rosca <[email protected]>

Assuming commit 617aebe6a97e ("Merge tag 'usercopy-v4.16-rc1' of
git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux"), ARCH=arm64 and vanilla arm64 defconfig, here is the top 10 CONFIG options with the highest amount of OR sub-expressions that make up the final "{Selected,Implied} by" reverse dependency expression.

| Config                        | Revdep all | Revdep ![=n] |
|-------------------------------|------------|--------------|
| REGMAP_I2C                    | 212        | 9            |
| CRC32                         | 167        | 25           |
| FW_LOADER                     | 128        | 5            |
| MFD_CORE                      | 124        | 9            |
| FB_CFB_IMAGEBLIT              | 114        | 2            |
| FB_CFB_COPYAREA               | 111        | 2            |
| FB_CFB_FILLRECT               | 110        | 2            |
| SND_PCM                       | 103        | 2            |
| CRYPTO_HASH                   | 87         | 19           |
| WATCHDOG_CORE                 | 86         | 6            |

The story behind the above table is that the user needs to visually review/evaluate 212 expressions which *potentially* select REGMAP_I2C in order to identify the expressions which *actually* select REGMAP_I2C, for a particular ARCH and for a particular defconfig used.

To make this experience smoother, transform the way reverse dependencies are displayed to the user from [1] to [2].

[1] Before this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- FTL [=n] && MTD [=y] && BLOCK [=y]
- NFTL [=n] && MTD [=y] && BLOCK [=y]
- INFTL [=n] && MTD [=y] && BLOCK [=y]
- RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- SSFDC [=n] && MTD [=y] && BLOCK [=y]
- SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

[2] After this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- [y] MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- [ ] MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- [ ] FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] NFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] INFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] SSFDC [=n] && MTD [=y] && BLOCK [=y]
- [ ] SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

This patch has been tested using a custom variant of zconfdump which prints the reverse dependencies for each config symbol.

---
 scripts/kconfig/expr.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)
scripts/kconfig/expr.c +13 -1
index 0704bcdf4c78..173fc5b252d5 100644
@@ -1185,7 +1185,19 @@ expr_print_newline(struct expr *e,
 		   void *data,
 		   int prevtoken)
 {
-	fn(data, NULL, "\n  - ");
+	switch (expr_calc_value(e)) {
+	case yes:
+		fn(data, NULL, "\n  - [y] ");
+		break;
+	case mod:
+		fn(data, NULL, "\n  - [m] ");
+		break;
+	case no:
+		fn(data, NULL, "\n  - [ ] ");
+		break;
+	default:
+		break;
+	}
 	expr_print(e, fn, data, prevtoken);
 }
 
Signature · 1 line
2.16.1
#2 lore raw

Re: [PATCH 2/2] kconfig: Print the value of each reverse dependency

Quoted 78 lines · Eugeniu Rosca

From: Eugeniu Rosca <[email protected]>

Assuming commit 617aebe6a97e ("Merge tag 'usercopy-v4.16-rc1' of
git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux"), ARCH=arm64 and vanilla arm64 defconfig, here is the top 10 CONFIG options with the highest amount of OR sub-expressions that make up the final "{Selected,Implied} by" reverse dependency expression.

| Config                        | Revdep all | Revdep ![=n] |
|-------------------------------|------------|--------------|
| REGMAP_I2C                    | 212        | 9            |
| CRC32                         | 167        | 25           |
| FW_LOADER                     | 128        | 5            |
| MFD_CORE                      | 124        | 9            |
| FB_CFB_IMAGEBLIT              | 114        | 2            |
| FB_CFB_COPYAREA               | 111        | 2            |
| FB_CFB_FILLRECT               | 110        | 2            |
| SND_PCM                       | 103        | 2            |
| CRYPTO_HASH                   | 87         | 19           |
| WATCHDOG_CORE                 | 86         | 6            |

The story behind the above table is that the user needs to visually review/evaluate 212 expressions which *potentially* select REGMAP_I2C in order to identify the expressions which *actually* select REGMAP_I2C, for a particular ARCH and for a particular defconfig used.

To make this experience smoother, transform the way reverse dependencies are displayed to the user from [1] to [2].

[1] Before this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- FTL [=n] && MTD [=y] && BLOCK [=y]
- NFTL [=n] && MTD [=y] && BLOCK [=y]
- INFTL [=n] && MTD [=y] && BLOCK [=y]
- RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- SSFDC [=n] && MTD [=y] && BLOCK [=y]
- SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

[2] After this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- [y] MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- [ ] MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- [ ] FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] NFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] INFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] SSFDC [=n] && MTD [=y] && BLOCK [=y]
- [ ] SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

This patch has been tested using a custom variant of zconfdump which prints the reverse dependencies for each config symbol.

---
 scripts/kconfig/expr.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)
scripts/kconfig/expr.c +2 -1
index 0704bcdf4c78..173fc5b252d5 100644
@@ -1185,7 +1185,19 @@ expr_print_newline(struct expr *e,
                   void *data,
                   int prevtoken)
 {
-       fn(data, NULL, "\n  - ");
+       switch (expr_calc_value(e)) {
+       case yes:

scripts/kconfig/expr.c expr_print_newline(struct expr *e, · line 1199

Quoted 10 lines
scripts/kconfig/expr.c +10 -0
+               fn(data, NULL, "\n  - [y] ");
+               break;
+       case mod:
+               fn(data, NULL, "\n  - [m] ");
+               break;
+       case no:
+               fn(data, NULL, "\n  - [ ] ");
+               break;
+       default:
+               break;

The default case is redundant. expr_calc_value() returns a

        typedef enum tristate {
                no, mod, yes
        } tristate;

(There's a separate sym_get_string_value() that deals with "string values" of symbols.)

Quoted 7 lines
scripts/kconfig/expr.c +1 -1
+       }
        expr_print(e, fn, data, prevtoken);
 }
 
--

2.16.1

Looks good to me otherwise. I still like that earlier sorting idea, but this is a big improvement already.

Cheers,
Ulf

#3 lore raw

Re: [PATCH 2/2] kconfig: Print the value of each reverse dependency

Hi Ulf,

Quoted 92 lines · Ulf Magnusson
Quoted 78 lines · Eugeniu Rosca

From: Eugeniu Rosca <[email protected]>

Assuming commit 617aebe6a97e ("Merge tag 'usercopy-v4.16-rc1' of
git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux"), ARCH=arm64 and vanilla arm64 defconfig, here is the top 10 CONFIG options with the highest amount of OR sub-expressions that make up the final "{Selected,Implied} by" reverse dependency expression.

| Config                        | Revdep all | Revdep ![=n] |
|-------------------------------|------------|--------------|
| REGMAP_I2C                    | 212        | 9            |
| CRC32                         | 167        | 25           |
| FW_LOADER                     | 128        | 5            |
| MFD_CORE                      | 124        | 9            |
| FB_CFB_IMAGEBLIT              | 114        | 2            |
| FB_CFB_COPYAREA               | 111        | 2            |
| FB_CFB_FILLRECT               | 110        | 2            |
| SND_PCM                       | 103        | 2            |
| CRYPTO_HASH                   | 87         | 19           |
| WATCHDOG_CORE                 | 86         | 6            |

The story behind the above table is that the user needs to visually review/evaluate 212 expressions which *potentially* select REGMAP_I2C in order to identify the expressions which *actually* select REGMAP_I2C, for a particular ARCH and for a particular defconfig used.

To make this experience smoother, transform the way reverse dependencies are displayed to the user from [1] to [2].

[1] Before this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- FTL [=n] && MTD [=y] && BLOCK [=y]
- NFTL [=n] && MTD [=y] && BLOCK [=y]
- INFTL [=n] && MTD [=y] && BLOCK [=y]
- RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- SSFDC [=n] && MTD [=y] && BLOCK [=y]
- SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

[2] After this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- [y] MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- [ ] MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- [ ] FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] NFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] INFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] SSFDC [=n] && MTD [=y] && BLOCK [=y]
- [ ] SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

This patch has been tested using a custom variant of zconfdump which prints the reverse dependencies for each config symbol.

---
 scripts/kconfig/expr.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)
scripts/kconfig/expr.c +2 -1
index 0704bcdf4c78..173fc5b252d5 100644
@@ -1185,7 +1185,19 @@ expr_print_newline(struct expr *e,
                   void *data,
                   int prevtoken)
 {
-       fn(data, NULL, "\n  - ");
+       switch (expr_calc_value(e)) {
+       case yes:

scripts/kconfig/expr.c expr_print_newline(struct expr *e, · line 1199

Quoted 10 lines
scripts/kconfig/expr.c +10 -0
+               fn(data, NULL, "\n  - [y] ");
+               break;
+       case mod:
+               fn(data, NULL, "\n  - [m] ");
+               break;
+       case no:
+               fn(data, NULL, "\n  - [ ] ");
+               break;
+       default:
+               break;

The default case is redundant. expr_calc_value() returns a

Quoted 6 lines
        typedef enum tristate {
                no, mod, yes
        } tristate;

(There's a separate sym_get_string_value() that deals with "string values" of symbols.)

Is this variant better?

scripts/kconfig/expr.c +18 -1
index 0704bcdf4c78..4bf01269ce72 100644
@@ -1185,7 +1185,24 @@ expr_print_newline(struct expr *e,
 		   void *data,
 		   int prevtoken)
 {
-	fn(data, NULL, "\n  - ");
+	char buf[32];
+	const char *str;
+
+	switch (expr_calc_value(e)) {
+	case yes:
+		str = sym_get_string_value(&symbol_yes);
+		break;
+	case mod:
+		str = sym_get_string_value(&symbol_mod);
+		break;
+	case no:
+		/* Prefer '[ ]' to '[n]' for readability */
+		str = " ";
+		break;
+	}
+
+	sprintf(buf, "\n  - [%s] ", str);
+	fn(data, NULL, buf);
 	expr_print(e, fn, data, prevtoken);
 }
 
Quoted 10 lines
Quoted 7 lines
scripts/kconfig/expr.c +1 -1
+       }
        expr_print(e, fn, data, prevtoken);
 }
 
--

2.16.1

Looks good to me otherwise. I still like that earlier sorting idea, but this is a big improvement already.

To be honest, purely as a user, I would probably prefer something like below:

Selected by [y/m]:
- EXPR_01
- EXPR_02
Selected by [n]:
- EXPR_03
- EXPR_04

But (as a developer) I feel it can only come at the price of increased complexity of __expr_print() (at least, if not bigger than, [1]). What I like about the current approach suggested by Masahiro is that it keeps the impact low with still making a great improvement in readability.

If you think we should explore the second possibility of grouping the active/inactive dependencies, I'll try to come up with some solution in the next days.

[1] https://marc.info/?l=linux-kbuild&m=151777006005199&w=4

Quoted 2 lines

Cheers,
Ulf

Thanks for your review and comments!

Best regards,
Eugeniu.

#4 lore raw

Re: [PATCH 2/2] kconfig: Print the value of each reverse dependency

Hi Eugeniu,

Quoted 9 lines

To be honest, purely as a user, I would probably prefer something like below:

Selected by [y/m]:
- EXPR_01
- EXPR_02
Selected by [n]:
- EXPR_03
- EXPR_04

Sorting again, just with header? It looks more readable (although Masahiro's approach is quite readable as well). If you chose this, I'd suggest not displaying title, when there is empty list. i.e., don't show "Selected by [y/m]" in this:
Selected by [y/m]:
Selected by [n]:
- EXPR_03
- EXPR_04

Quoted 10 lines

But (as a developer) I feel it can only come at the price of increased complexity of __expr_print() (at least, if not bigger than, [1]). What I like about the current approach suggested by Masahiro is that it keeps the impact low with still making a great improvement in readability.

If you think we should explore the second possibility of grouping the active/inactive dependencies, I'll try to come up with some solution in the next days.

[1] https://marc.info/?l=linux-kbuild&m=151777006005199&w=4

Kind regards,
Petr

#5 lore raw

Re: [PATCH 2/2] kconfig: Print the value of each reverse dependency

Hello Petr,

Quoted 11 lines · Petr Vorel

Hi Eugeniu,

Quoted 9 lines

To be honest, purely as a user, I would probably prefer something like below:

Selected by [y/m]:
- EXPR_01
- EXPR_02
Selected by [n]:
- EXPR_03
- EXPR_04

Quoted 7 lines

Sorting again, just with header? It looks more readable (although Masahiro's approach is quite readable as well). If you chose this, I'd suggest not displaying title, when there is empty list. i.e., don't show "Selected by [y/m]" in this:
Selected by [y/m]:
Selected by [n]:
- EXPR_03
- EXPR_04

I took your preference into consideration in the v3 revision. Thanks!

Best regards,
Eugeniu.

#6 lore raw

Re: [PATCH 2/2] kconfig: Print the value of each reverse dependency

In reply to Eugeniu Rosca
Quoted 124 lines · Eugeniu Rosca

Hi Ulf,

Quoted 92 lines · Ulf Magnusson
Quoted 78 lines · Eugeniu Rosca

From: Eugeniu Rosca <[email protected]>

Assuming commit 617aebe6a97e ("Merge tag 'usercopy-v4.16-rc1' of
git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux"), ARCH=arm64 and vanilla arm64 defconfig, here is the top 10 CONFIG options with the highest amount of OR sub-expressions that make up the final "{Selected,Implied} by" reverse dependency expression.

| Config                        | Revdep all | Revdep ![=n] |
|-------------------------------|------------|--------------|
| REGMAP_I2C                    | 212        | 9            |
| CRC32                         | 167        | 25           |
| FW_LOADER                     | 128        | 5            |
| MFD_CORE                      | 124        | 9            |
| FB_CFB_IMAGEBLIT              | 114        | 2            |
| FB_CFB_COPYAREA               | 111        | 2            |
| FB_CFB_FILLRECT               | 110        | 2            |
| SND_PCM                       | 103        | 2            |
| CRYPTO_HASH                   | 87         | 19           |
| WATCHDOG_CORE                 | 86         | 6            |

The story behind the above table is that the user needs to visually review/evaluate 212 expressions which *potentially* select REGMAP_I2C in order to identify the expressions which *actually* select REGMAP_I2C, for a particular ARCH and for a particular defconfig used.

To make this experience smoother, transform the way reverse dependencies are displayed to the user from [1] to [2].

[1] Before this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- FTL [=n] && MTD [=y] && BLOCK [=y]
- NFTL [=n] && MTD [=y] && BLOCK [=y]
- INFTL [=n] && MTD [=y] && BLOCK [=y]
- RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- SSFDC [=n] && MTD [=y] && BLOCK [=y]
- SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

[2] After this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- [y] MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- [ ] MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- [ ] FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] NFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] INFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] SSFDC [=n] && MTD [=y] && BLOCK [=y]
- [ ] SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

This patch has been tested using a custom variant of zconfdump which prints the reverse dependencies for each config symbol.

---
 scripts/kconfig/expr.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)
scripts/kconfig/expr.c +2 -1
index 0704bcdf4c78..173fc5b252d5 100644
@@ -1185,7 +1185,19 @@ expr_print_newline(struct expr *e,
                   void *data,
                   int prevtoken)
 {
-       fn(data, NULL, "\n  - ");
+       switch (expr_calc_value(e)) {
+       case yes:

scripts/kconfig/expr.c expr_print_newline(struct expr *e, · line 1199

Quoted 10 lines
scripts/kconfig/expr.c +10 -0
+               fn(data, NULL, "\n  - [y] ");
+               break;
+       case mod:
+               fn(data, NULL, "\n  - [m] ");
+               break;
+       case no:
+               fn(data, NULL, "\n  - [ ] ");
+               break;
+       default:
+               break;

The default case is redundant. expr_calc_value() returns a

Quoted 6 lines
        typedef enum tristate {
                no, mod, yes
        } tristate;

(There's a separate sym_get_string_value() that deals with "string values" of symbols.)

Is this variant better?

scripts/kconfig/expr.c +11 -1
index 0704bcdf4c78..4bf01269ce72 100644
@@ -1185,7 +1185,24 @@ expr_print_newline(struct expr *e,
                   void *data,
                   int prevtoken)
 {
-       fn(data, NULL, "\n  - ");
+       char buf[32];
+       const char *str;
+
+       switch (expr_calc_value(e)) {
+       case yes:
+               str = sym_get_string_value(&symbol_yes);
+               break;
+       case mod:
+               str = sym_get_string_value(&symbol_mod);
+               break;
+       case no:

scripts/kconfig/expr.c expr_print_newline(struct expr *e, · line 1208

Quoted 10 lines
scripts/kconfig/expr.c +7 -0
+               /* Prefer '[ ]' to '[n]' for readability */
+               str = " ";
+               break;
+       }
+
+       sprintf(buf, "\n  - [%s] ", str);
+       fn(data, NULL, buf);
        expr_print(e, fn, data, prevtoken);
 }
 

The older version was simpler. sym_get_string_value(&symbol_yes/no) would never be anything other than "y"/"n" in practice. :)

Just removing the following two lines from the original version would be best I think:

default:
break

Quoted 26 lines
Quoted 10 lines
Quoted 7 lines
scripts/kconfig/expr.c +1 -1
+       }
        expr_print(e, fn, data, prevtoken);
 }
 
--

2.16.1

Looks good to me otherwise. I still like that earlier sorting idea, but this is a big improvement already.

To be honest, purely as a user, I would probably prefer something like below:

Selected by [y/m]:
- EXPR_01
- EXPR_02
Selected by [n]:
- EXPR_03
- EXPR_04

But (as a developer) I feel it can only come at the price of increased complexity of __expr_print() (at least, if not bigger than, [1]). What I like about the current approach suggested by Masahiro is that it keeps the impact low with still making a great improvement in readability.

Quoted 5 lines

If you think we should explore the second possibility of grouping the active/inactive dependencies, I'll try to come up with some solution in the next days.

[1] https://marc.info/?l=linux-kbuild&m=151777006005199&w=4

IMO, we could take this as is (after addressing other people's comments), since it's already a big improvement, and then add sorting later if we feel like it.

Don't have to do everything at once. :)

Think it should be pretty straightforward to add the kind of sorting you had in mind at least, by changing PRINT_ALL_SELECTS into PRINT_INACTIVE_SELECTS in [1], and tweaking some conditions.

Quoted 8 lines
Quoted 3 lines

Cheers,
Ulf

Thanks for your review and comments!

Best regards,
Eugeniu.

Cheers,
Ulf

#7 lore raw

Re: [PATCH 2/2] kconfig: Print the value of each reverse dependency

Hi Ulf,

Quoted 5 lines · Ulf Magnusson

IMO, we could take this as is (after addressing other people's comments), since it's already a big improvement, and then add sorting later if we feel like it.

Don't have to do everything at once. :)

I agree that it's not possible to do everything perfect right from the start. But since we are changing the user experience, I just thought that we would first like to see how both implementations (prefixed vs grouped tokens) look like, then make a decision (I'm still open minded about it), then affect the users.

Quoted 4 lines

Think it should be pretty straightforward to add the kind of sorting you had in mind at least, by changing PRINT_ALL_SELECTS into PRINT_INACTIVE_SELECTS in [1], and tweaking some conditions.

Yep. Thanks to your support, I've pushed v3, where you've already provided your review findings, which I'm going to fix soon :)

Thanks,
Eugeniu.

#8 lore raw

Re: [PATCH 2/2] kconfig: Print the value of each reverse dependency

Quoted 7 lines · Eugeniu Rosca

Hi Ulf,

Quoted 4 lines · Ulf Magnusson

IMO, we could take this as is (after addressing other people's comments), since it's already a big improvement, and then add sorting later if we feel like it.

Quoted 7 lines
Quoted 1 line

Don't have to do everything at once. :)

I agree that it's not possible to do everything perfect right from the start. But since we are changing the user experience, I just thought that we would first like to see how both implementations (prefixed vs grouped tokens) look like, then make a decision (I'm still open minded about it), then affect the users.

Isn't the latest patchset doing both? That seems like the best of both worlds.

And yeah, I just get some personal restlessness when good stuff floats around for a long time without getting in. :)

Cheers,
Ulf

#9 lore raw

Re: [PATCH 2/2] kconfig: Print the value of each reverse dependency

Quoted 10 lines · Ulf Magnusson
Quoted 9 lines · Eugeniu Rosca

Hi Ulf,

Quoted 5 lines · Ulf Magnusson

IMO, we could take this as is (after addressing other people's comments), since it's already a big improvement, and then add sorting later if we feel like it.

Don't have to do everything at once. :)

Quoted 7 lines
Quoted 5 lines

I agree that it's not possible to do everything perfect right from the start. But since we are changing the user experience, I just thought that we would first like to see how both implementations (prefixed vs grouped tokens) look like, then make a decision (I'm still open minded about it), then affect the users.

Isn't the latest patchset doing both? That seems like the best of both worlds.

Well, I call "prefixed tokens" solution [1] and "grouped tokens" solution [2]. Assuming we share the same view, [1] has definitely less of an impact on Kconfig (see [3]). That's the main argument to prefer it over [2]. My personal choice is still [2]. However if it happens that other reviewers prefer [1], I won't object on it.

[1] Selected by:
- [y] EXPR_Y
- [m] EXPR_M
- [ ] EXPR_N

[2] Selected by [y]:
- EXPR_Y
Selected by [m]:
- EXPR_M
Selected by [n]:
- EXPR_N

[3] https://marc.info/?l=linux-kbuild&m=151848343313335&w=4

Quoted 5 lines

And yeah, I just get some personal restlessness when good stuff floats around for a long time without getting in. :)

Cheers,
Ulf

Thanks,
Eugeniu.

#10 lore raw

Re: [PATCH 2/2] kconfig: Print the value of each reverse dependency

In reply to Eugeniu Rosca

Hi Eugeniu,

Quoted 43 lines

From: Eugeniu Rosca <[email protected]>

Assuming commit 617aebe6a97e ("Merge tag 'usercopy-v4.16-rc1' of
git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux"), ARCH=arm64 and vanilla arm64 defconfig, here is the top 10 CONFIG options with the highest amount of OR sub-expressions that make up the final "{Selected,Implied} by" reverse dependency expression.

| Config                        | Revdep all | Revdep ![=n] |
|-------------------------------|------------|--------------|
| REGMAP_I2C                    | 212        | 9            |
| CRC32                         | 167        | 25           |
| FW_LOADER                     | 128        | 5            |
| MFD_CORE                      | 124        | 9            |
| FB_CFB_IMAGEBLIT              | 114        | 2            |
| FB_CFB_COPYAREA               | 111        | 2            |
| FB_CFB_FILLRECT               | 110        | 2            |
| SND_PCM                       | 103        | 2            |
| CRYPTO_HASH                   | 87         | 19           |
| WATCHDOG_CORE                 | 86         | 6            |

The story behind the above table is that the user needs to visually review/evaluate 212 expressions which *potentially* select REGMAP_I2C in order to identify the expressions which *actually* select REGMAP_I2C, for a particular ARCH and for a particular defconfig used.

To make this experience smoother, transform the way reverse dependencies are displayed to the user from [1] to [2].

[1] Before this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- FTL [=n] && MTD [=y] && BLOCK [=y]
- NFTL [=n] && MTD [=y] && BLOCK [=y]
- INFTL [=n] && MTD [=y] && BLOCK [=y]
- RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- SSFDC [=n] && MTD [=y] && BLOCK [=y]
- SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

Quoted 13 lines

[2] After this commit
Symbol: MTD_BLKDEVS [=y]
...
Selected by:
- [y] MTD_BLOCK [=y] && MTD [=y] && BLOCK [=y]
- [ ] MTD_BLOCK_RO [=n] && MTD [=y] && MTD_BLOCK [=y]!=y && BLOCK [=y]
- [ ] FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] NFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] INFTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] RFD_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] SSFDC [=n] && MTD [=y] && BLOCK [=y]
- [ ] SM_FTL [=n] && MTD [=y] && BLOCK [=y]
- [ ] MTD_SWAP [=n] && MTD [=y] && SWAP [=y]

Note, Masahiro suggested to have ':' after first expression [1]:
Selected by:

    [y]: EEPROM_AT24 [=y] && I2C [=y] && SYSFS [=y]
    [ ]: NET_DSA_SMSC_LAN9303_I2C [=n] && NETDEVICES [=y] &&

HAVE_NET_DSA [=y] && NET_DSA [=n] && I2C [=y]
...

I think it's better, it's clearer that the expression is for whole line.

Kind regards,
Petr

[1]: https://marc.info/?l=linux-kbuild&m=151791725709825&w=4

Reviewed-by on v1 1/2
#11 lore raw
In reply to Eugeniu Rosca
Quoted 113 lines · Eugeniu Rosca

From: Eugeniu Rosca <[email protected]>

Commit 1ccb27143360 ("kconfig: make "Selected by:" and "Implied by:" readable") made an incredible improvement in how reverse dependencies are perceived by the user, by breaking down the single (often interminable) expression string into small readable chunks, each of them displayed on a separate line:

Selected by:
- A & B
- C & (D || E)

Unfortunately, what happens with the non-OR (either E_SYMBOL or E_AND) expressions is that they don't get a dedicated line:

Selected by: F & G

That's arguably a bug/misbehavior, but it makes the implementation of something like below more complicated:

Selected by:
- [m] F & G /* where (F & G) evaluates to '=m' */

Adding '[m]', '[y]' or '[ ]' to the left side of each reverse dependency is subject of a different commit. The purpose of current commit is to print the 'Selected by:' and 'Implied by:' expressions on a separate line _consistently_.

An example of change contributed by this commit is seen below.

│ Symbol: NEED_SG_DMA_LENGTH [=y]
│ ...
│   Selected by: IOMMU_DMA [=y] && IOMMU_SUPPORT [=y]

becomes:

│ Symbol: NEED_SG_DMA_LENGTH [=y]
│ ...
│   Selected by:
│   - IOMMU_DMA [=y] && IOMMU_SUPPORT [=y]

This patch has been tested using a custom variant of zconfdump which prints the reverse dependencies for each config symbol.

---
 scripts/kconfig/expr.c | 31 ++++++++++++++++++++++---------
 1 file changed, 22 insertions(+), 9 deletions(-)
scripts/kconfig/expr.c +22 -10
index d45381986ac7..0704bcdf4c78 100644
@@ -1179,6 +1179,16 @@ struct expr *expr_simplify_unmet_dep(struct expr *e1, struct expr *e2)
        return expr_get_leftmost_symbol(ret);
 }
 
+static void
+expr_print_newline(struct expr *e,
+                  void (*fn)(void *, struct symbol *, const char *),
+                  void *data,
+                  int prevtoken)
+{
+       fn(data, NULL, "\n  - ");
+       expr_print(e, fn, data, prevtoken);
+}
+
 static void __expr_print(struct expr *e, void (*fn)(void *, struct symbol *, const char *), void *data, int prevtoken, bool revdep)
 {
        if (!e) {
@@ -1191,7 +1201,10 @@ static void __expr_print(struct expr *e, void (*fn)(void *, struct symbol *, con
        switch (e->type) {
        case E_SYMBOL:
                if (e->left.sym->name)
-                       fn(data, e->left.sym, e->left.sym->name);
+                       if (revdep)
+                               expr_print_newline(e, fn, data, E_OR);
+                       else
+                               fn(data, e->left.sym, e->left.sym->name);
                else
                        fn(data, NULL, "<choice>");
                break;
@@ -1234,19 +1247,19 @@ static void __expr_print(struct expr *e, void (*fn)(void *, struct symbol *, con
                fn(data, e->right.sym, e->right.sym->name);
                break;
        case E_OR:
-               if (revdep && e->left.expr->type != E_OR)
-                       fn(data, NULL, "\n  - ");
                __expr_print(e->left.expr, fn, data, E_OR, revdep);
-               if (revdep)
-                       fn(data, NULL, "\n  - ");
-               else
+               if (!revdep)
                        fn(data, NULL, " || ");
                __expr_print(e->right.expr, fn, data, E_OR, revdep);
                break;
        case E_AND:
-               expr_print(e->left.expr, fn, data, E_AND);
-               fn(data, NULL, " && ");
-               expr_print(e->right.expr, fn, data, E_AND);
+               if (revdep) {
+                       expr_print_newline(e, fn, data, E_OR);
+               } else {
+                       expr_print(e->left.expr, fn, data, E_AND);
+                       fn(data, NULL, " && ");
+                       expr_print(e->right.expr, fn, data, E_AND);
+               }
                break;
        case E_LIST:
                fn(data, e->right.sym, e->right.sym->name);
--
Quoted 2 lines

2.16.1

Cheers,
Ulf

#12 lore raw

Hi Eugeniu,

Quoted 9 lines · Eugeniu Rosca
Quoted 7 lines

From: Eugeniu Rosca <[email protected]>

Commit 1ccb27143360 ("kconfig: make "Selected by:" and "Implied by:" readable") made an incredible improvement in how reverse dependencies are perceived by the user, by breaking down the single (often interminable) expression string into small readable chunks, each of them displayed on a separate line:

Quoted 8 lines
Quoted 8 lines

Selected by:
- A & B
- C & (D || E)

Unfortunately, what happens with the non-OR (either E_SYMBOL or E_AND) expressions is that they don't get a dedicated line:

Selected by: F & G

I deliberately choose it :-). It didn't make much sense for me to add new line for just one line. I thought someone wouldn't like it for the inconsistency :-).

Kind regards,
Petr

#13 lore raw

Hi Petr,

Quoted 15 lines · Petr Vorel

Hi Eugeniu,

Quoted 12 lines · Eugeniu Rosca
Quoted 11 lines

From: Eugeniu Rosca <[email protected]>

Commit 1ccb27143360 ("kconfig: make "Selected by:" and "Implied by:" readable") made an incredible improvement in how reverse dependencies are perceived by the user, by breaking down the single (often interminable) expression string into small readable chunks, each of them displayed on a separate line:

Selected by:
- A & B
- C & (D || E)

Quoted 6 lines
Quoted 4 lines
Quoted 4 lines

Unfortunately, what happens with the non-OR (either E_SYMBOL or E_AND) expressions is that they don't get a dedicated line:

Selected by: F & G

I deliberately choose it :-). It didn't make much sense for me to add new line for just one line. I thought someone wouldn't like it for the inconsistency :-).

I have to say I do like how Selected by: F && G looks on a single row. However, when the expression is prefixed by its value (i.e. Selected by: [m] F && G), imho it doesn't look as crisp and clear as we would like it to (feel free to disagree here). Fwiw, this patch doesn't see itself as a fix, but rather prepares the ground for the next commit implementing prefixing reverse dependencies by their expression value.

Quoted 3 lines

Kind regards,
Petr

Thanks!
Eugeniu.