Skip to contentSkip to conversation

cmd: ufetch improvements

PATCH Superseded J. Neuschäfer 2 participants 10 messages Dec 5, 2024 – Dec 9, 2024

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover cmd: ufetch improvements

Following the recent addition of the "ufetch" command, this patchset improves it in a few ways.

---
J. Neuschäfer (3):
cmd: ufetch: Fix type mismatch on 32-bit
cmd: Allow building ufetch without CONFIG_BLK
cmd: ufetch: Show CPU architecture under "CPU"

 cmd/Kconfig  |  1 -
 cmd/ufetch.c | 14 +++++++++-----
 2 files changed, 9 insertions(+), 6 deletions(-)

---
base-commit: e5c1c2c6d9da79d874015f6d33fa2c7a34899733
change-id: 20241203-ufetch-326e771b26ad

Best regards,

Signature · 1 line
J. Neuschäfer <[email protected]>
#1 lore raw

PATCH 1/3 cmd: ufetch: Fix type mismatch on 32-bit

On 32-bit architectures, LAST_LINE (_LAST_LINE - 1UL) is 64 bits long, but size_t (from ARRAY_SIZE(...)) is 32 bits. This results in a warning because the max() macro expects the same type on both sides:

cmd/ufetch.c: In function ‘do_ufetch’:
include/linux/kernel.h:179:24: warning: comparison of distinct pointer types lacks a cast [-Wcompare-distinct-pointer-types]
179 | (void) (&_max1 == &_max2); \
| ^~
cmd/ufetch.c:92:25: note: in expansion of macro ‘max’
92 | int num_lines = max(LAST_LINE + 1, ARRAY_SIZE(logo_lines));
| ^~~

Fix this by casting LAST_LINE to size_t.

---
 cmd/ufetch.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
cmd/ufetch.c +1 -1
index 0b825d7e8c75f3b18933d3e3f77e5f40f2c7b658..5f3ef847b268dc384271fc6774720e5fd2337157 100644
@@ -89,7 +89,7 @@ enum output_lines {
 static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 		     char *const argv[])
 {
-	int num_lines = max(LAST_LINE + 1, ARRAY_SIZE(logo_lines));
+	int num_lines = max((size_t)LAST_LINE + 1, ARRAY_SIZE(logo_lines));
 	const char *model, *compatible;
 	char *ipaddr;
 	int n_cmds, n_cpus = 0, ret, compatlen;
Signature · 1 line
2.45.2
Reviewed-by on v1 1/3
#2 lore raw

Re: [PATCH 1/3] cmd: ufetch: Fix type mismatch on 32-bit

Hi J,

Thanks for the patch!

Quoted 14 lines · On 05/12/2024 19:35, J. Neuschäfer via B4 Relay wrote:

From: "J. Neuschäfer" <[email protected]>

On 32-bit architectures, LAST_LINE (_LAST_LINE - 1UL) is 64 bits long, but size_t (from ARRAY_SIZE(...)) is 32 bits. This results in a warning because the max() macro expects the same type on both sides:

cmd/ufetch.c: In function ‘do_ufetch’:
include/linux/kernel.h:179:24: warning: comparison of distinct pointer types lacks a cast [-Wcompare-distinct-pointer-types]
179 | (void) (&_max1 == &_max2); \
| ^~
cmd/ufetch.c:92:25: note: in expansion of macro ‘max’
92 | int num_lines = max(LAST_LINE + 1, ARRAY_SIZE(logo_lines));
| ^~~

Quoted 3 lines

Fix this by casting LAST_LINE to size_t.

Quoted 18 lines
---
 cmd/ufetch.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
cmd/ufetch.c +1 -1
index 0b825d7e8c75f3b18933d3e3f77e5f40f2c7b658..5f3ef847b268dc384271fc6774720e5fd2337157 100644
@@ -89,7 +89,7 @@ enum output_lines {
 static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 		     char *const argv[])
 {
-	int num_lines = max(LAST_LINE + 1, ARRAY_SIZE(logo_lines));
+	int num_lines = max((size_t)LAST_LINE + 1, ARRAY_SIZE(logo_lines));
 	const char *model, *compatible;
 	char *ipaddr;
 	int n_cmds, n_cpus = 0, ret, compatlen;
Signature · 1 line
// Caleb (they/them)
#3 lore raw

PATCH 2/3 cmd: Allow building ufetch without CONFIG_BLK

In reply to J. Neuschäfer

The ufetch command is still quite useful on systems without block device support; remove the CONFIG_BLK dependency and make sure the code compiles/works with and without CONFIG_BLK.

---
 cmd/Kconfig  |  1 -
 cmd/ufetch.c | 10 +++++++---
 2 files changed, 7 insertions(+), 4 deletions(-)
cmd/Kconfig +0 -1
index 4936a70f3ef16ddb093ceafa12d011ca1b89e95c..547fd2a91f7883e2ae5982897ec93c4483d67852 100644
@@ -178,7 +178,6 @@ config CMD_CPU
 
 config CMD_UFETCH
 	bool "U-Boot fetch"
-	depends on BLK
 	help
 	  Fetch utility for U-Boot (akin to neofetch). Prints information
 	  about U-Boot and the board it is running on in a pleasing format.
cmd/ufetch.c +7 -3
index 5f3ef847b268dc384271fc6774720e5fd2337157..7aed0b447bda104b837d37c6adcbb21b80237aba 100644
@@ -92,11 +92,9 @@ static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 	int num_lines = max((size_t)LAST_LINE + 1, ARRAY_SIZE(logo_lines));
 	const char *model, *compatible;
 	char *ipaddr;
-	int n_cmds, n_cpus = 0, ret, compatlen;
+	int n_cmds, n_cpus = 0, compatlen;
 	size_t size;
 	ofnode np;
-	struct udevice *dev;
-	struct blk_desc *desc;
 	bool skip_ascii = false;
 
 	if (argc > 1 && strcmp(argv[1], "-n") == 0) {
@@ -200,6 +198,11 @@ static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 			break;
 		case STORAGE:
 		default:
+#ifdef CONFIG_BLK
+			struct udevice *dev;
+			struct blk_desc *desc;
+			int ret;
+
 			ret = uclass_find_device_by_seq(UCLASS_BLK, line - STORAGE, &dev);
 			if (!ret && dev) {
 				desc = dev_get_uclass_plat(dev);
@@ -213,6 +216,7 @@ static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 			} else if (ret == -ENODEV && (skip_ascii || line > ARRAY_SIZE(logo_lines))) {
 				break;
 			}
+#endif
 			printf("\n");
 		}
 	}
Signature · 1 line
2.45.2
Reviewed-by on v1 2/3
#4 lore raw

Re: [PATCH 2/3] cmd: Allow building ufetch without CONFIG_BLK

Quoted 7 lines · On 05/12/2024 19:35, J. Neuschäfer via B4 Relay wrote:

From: "J. Neuschäfer" <[email protected]>

The ufetch command is still quite useful on systems without block device support; remove the CONFIG_BLK dependency and make sure the code compiles/works with and without CONFIG_BLK.

Small nit below, but with that:

Quoted 35 lines
---
 cmd/Kconfig  |  1 -
 cmd/ufetch.c | 10 +++++++---
 2 files changed, 7 insertions(+), 4 deletions(-)
cmd/Kconfig +0 -1
index 4936a70f3ef16ddb093ceafa12d011ca1b89e95c..547fd2a91f7883e2ae5982897ec93c4483d67852 100644
@@ -178,7 +178,6 @@ config CMD_CPU
 
 config CMD_UFETCH
 	bool "U-Boot fetch"
-	depends on BLK
 	help
 	  Fetch utility for U-Boot (akin to neofetch). Prints information
 	  about U-Boot and the board it is running on in a pleasing format.
cmd/ufetch.c +1 -3
index 5f3ef847b268dc384271fc6774720e5fd2337157..7aed0b447bda104b837d37c6adcbb21b80237aba 100644
@@ -92,11 +92,9 @@ static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 	int num_lines = max((size_t)LAST_LINE + 1, ARRAY_SIZE(logo_lines));
 	const char *model, *compatible;
 	char *ipaddr;
-	int n_cmds, n_cpus = 0, ret, compatlen;
+	int n_cmds, n_cpus = 0, compatlen;
 	size_t size;
 	ofnode np;
-	struct udevice *dev;
-	struct blk_desc *desc;
 	bool skip_ascii = false;
 
 	if (argc > 1 && strcmp(argv[1], "-n") == 0) {

cmd/ufetch.c static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc, · line 202

Quoted 5 lines
cmd/ufetch.c +2 -0
@@ -200,6 +198,11 @@ static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 			break;
 		case STORAGE:
 		default:
+#ifdef CONFIG_BLK
+			struct udevice *dev;

At least without the #ifdef, I see a warning "A label followed by a declaration is a C23 extension". So to be on the safe side please put this section in a block:

		default: {

...

		}

Kind regards,

Quoted 15 lines
cmd/ufetch.c +4 -0
+			struct blk_desc *desc;
+			int ret;
+
 			ret = uclass_find_device_by_seq(UCLASS_BLK, line - STORAGE, &dev);
 			if (!ret && dev) {
 				desc = dev_get_uclass_plat(dev);
@@ -213,6 +216,7 @@ static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 			} else if (ret == -ENODEV && (skip_ascii || line > ARRAY_SIZE(logo_lines))) {
 				break;
 			}
+#endif
 			printf("\n");
 		}
 	}
Signature · 1 line
// Caleb (they/them)
#5 lore raw

Re: [PATCH 2/3] cmd: Allow building ufetch without CONFIG_BLK

Quoted 13 lines · Caleb Connolly
Quoted 7 lines · On 05/12/2024 19:35, J. Neuschäfer via B4 Relay wrote:

From: "J. Neuschäfer" <[email protected]>

The ufetch command is still quite useful on systems without block device support; remove the CONFIG_BLK dependency and make sure the code compiles/works with and without CONFIG_BLK.

Small nit below, but with that:

Quoted 1 line

---

[...]

Quoted 6 lines

cmd/ufetch.c static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc, · line 202

Quoted 3 lines
cmd/ufetch.c +2 -0
 		default:
+#ifdef CONFIG_BLK
+			struct udevice *dev;

At least without the #ifdef, I see a warning "A label followed by a declaration is a C23 extension".

Ah, I wondered about this. Apparently I have a compiler that uses C23 as the default, so I didn't see the warning. I'll fix it in the next revision.

Quoted 3 lines

So to be on the safe side please put this section in a block:

		default: {

An alternative that I prefer because it's a bit less intrusive with regards to indentation, is to use a semicolon:

		default:;
			/* code with the same indentation as usual */
			/* and no closing brace to remember */

What do you think?

-- jn

#6 lore raw

Re: [PATCH 2/3] cmd: Allow building ufetch without CONFIG_BLK

Quoted 27 lines · J. Neuschäfer
Quoted 13 lines · Caleb Connolly
Quoted 7 lines · On 05/12/2024 19:35, J. Neuschäfer via B4 Relay wrote:

From: "J. Neuschäfer" <[email protected]>

The ufetch command is still quite useful on systems without block device support; remove the CONFIG_BLK dependency and make sure the code compiles/works with and without CONFIG_BLK.

Small nit below, but with that:

Quoted 1 line

---

[...]

Quoted 6 lines

cmd/ufetch.c static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc, · line 202

Quoted 3 lines
cmd/ufetch.c +2 -0
 		default:
+#ifdef CONFIG_BLK
+			struct udevice *dev;

At least without the #ifdef, I see a warning "A label followed by a declaration is a C23 extension".

Ah, I wondered about this. Apparently I have a compiler that uses C23 as the default, so I didn't see the warning. I'll fix it in the next revision.

Quoted 2 lines

So to be on the safe side please put this section in a block:

Quoted 8 lines
Quoted 1 line
		default: {

An alternative that I prefer because it's a bit less intrusive with regards to indentation, is to use a semicolon:

		default:;
			/* code with the same indentation as usual */
			/* and no closing brace to remember */

Hmm, haven't seen that one before. fwiw I think you can leave the indentation as-is and just add the brackets.

Kind regards,

Quoted 5 lines

What do you think?

-- jn

Signature · 1 line
// Caleb (they/them)
#7 lore raw

PATCH 3/3 cmd: ufetch: Show CPU architecture under "CPU"

In reply to J. Neuschäfer

When looking at ufetch output it isn't immediately obvious which CPU architecture the presented board has. This patch therefore adds the CPU architecture string (for example "powerpc") to the "CPU:" line.

---
 cmd/ufetch.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
cmd/ufetch.c +1 -1
index 7aed0b447bda104b837d37c6adcbb21b80237aba..0ee50bfb606aa2223072c63cd46b3a70ef5449fd 100644
@@ -188,7 +188,7 @@ static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 				if (ofnode_name_eq(np, "cpu"))
 					n_cpus++;
 			}
-			printf("CPU:" RESET " %d (1 in use)\n", n_cpus);
+			printf("CPU: " RESET CONFIG_SYS_ARCH ", %d (1 in use)\n", n_cpus);
 			break;
 		case MEMORY:
 			for (int j = 0; j < CONFIG_NR_DRAM_BANKS && gd->bd->bi_dram[j].size; j++)
Signature · 1 line
2.45.2
#8 lore raw

Re: [PATCH 3/3] cmd: ufetch: Show CPU architecture under "CPU"

Thanks for the patch!

Quoted 5 lines · On 05/12/2024 19:35, J. Neuschäfer via B4 Relay wrote:

From: "J. Neuschäfer" <[email protected]>

When looking at ufetch output it isn't immediately obvious which CPU architecture the presented board has. This patch therefore adds the CPU architecture string (for example "powerpc") to the "CPU:" line.

It would be nice to have the proper fancy name, but this is an improvement nonetheless.

Quoted 11 lines
---
 cmd/ufetch.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
cmd/ufetch.c +0 -0
index 7aed0b447bda104b837d37c6adcbb21b80237aba..0ee50bfb606aa2223072c63cd46b3a70ef5449fd 100644

cmd/ufetch.c static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc, · line 191

Quoted 5 lines
cmd/ufetch.c +1 -1
@@ -188,7 +188,7 @@ static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc,
 				if (ofnode_name_eq(np, "cpu"))
 					n_cpus++;
 			}
-			printf("CPU:" RESET " %d (1 in use)\n", n_cpus);
+			printf("CPU: " RESET CONFIG_SYS_ARCH ", %d (1 in use)\n", n_cpus);

This will read like

CPU: arm 4 (1 in use)

which is a bit hard to parse at a glance. How about

CPU: arm [4 cores]

or some other separator between SYS_ARCH and the core count?

Kind regards,

Quoted 4 lines
cmd/ufetch.c +0 -0
 			break;
 		case MEMORY:
 			for (int j = 0; j < CONFIG_NR_DRAM_BANKS && gd->bd->bi_dram[j].size; j++)
Signature · 1 line
// Caleb (they/them)
#9 lore raw

Re: [PATCH 3/3] cmd: ufetch: Show CPU architecture under "CPU"

Quoted 11 lines · Caleb Connolly

Thanks for the patch!

Quoted 5 lines · On 05/12/2024 19:35, J. Neuschäfer via B4 Relay wrote:

From: "J. Neuschäfer" <[email protected]>

When looking at ufetch output it isn't immediately obvious which CPU architecture the presented board has. This patch therefore adds the CPU architecture string (for example "powerpc") to the "CPU:" line.

It would be nice to have the proper fancy name, but this is an improvement nonetheless.

True

Quoted 3 lines
Quoted 3 lines

---

[...]

Quoted 5 lines

cmd/ufetch.c static int do_ufetch(struct cmd_tbl *cmdtp, int flag, int argc, · line 191

Quoted 2 lines
cmd/ufetch.c +1 -1
-			printf("CPU:" RESET " %d (1 in use)\n", n_cpus);
+			printf("CPU: " RESET CONFIG_SYS_ARCH ", %d (1 in use)\n", n_cpus);

This will read like

Quoted 7 lines

CPU: arm 4 (1 in use)

which is a bit hard to parse at a glance. How about

CPU: arm [4 cores]

or some other separator between SYS_ARCH and the core count?

Ah yup, good idea. I also struggled with reading this line as it was; "4 cores" or similar makes it clear what the number means.

-- jn