Skip to contentSkip to conversation

drivers/scsi/arm/acornscsi.c: rename CONFIG_ACORNSCSI_CONSTANTS

PATCH In review Valentin Rothberg 3 participants 6 messages Apr 28, 2015 – May 6, 2015

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drivers/scsi/arm/acornscsi.c: rename CONFIG_ACORNSCSI_CONSTANTS

CONFIG_ACORNSCSI_CONSTANTS is a file local CPP identifier and thereby violates the naming convention of Kconfig options in Make and CPP syntax; only Kconfig options should carry the 'CONFIG_' prefix.

This patch removes the 'CONFIG_' prefix to apply to this convention and to make static analysis tools happy.

---
I detected this issue with ./scripts/checkkconfigsymbols.py

---
 drivers/scsi/arm/acornscsi.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/scsi/arm/acornscsi.c +3 -3
index deaaf84989cd..ce764c3ed99c 100644
@@ -96,7 +96,7 @@
  * Define this if you want to have verbose explanation of SCSI
  * status/messages.
  */
-#undef CONFIG_ACORNSCSI_CONSTANTS
+#undef ACORNSCSI_CONSTANTS
 /*
  * Define this if you want to use the on board DMAC [don't remove this option]
  * If not set, then use PIO mode (not currently supported).
@@ -399,7 +399,7 @@ void acornscsi_resetcard(AS_Host *host)
 /*=============================================================================================
  * Utility routines (eg. debug)
  */
-#ifdef CONFIG_ACORNSCSI_CONSTANTS
+#ifdef ACORNSCSI_CONSTANTS
 static char *acornscsi_interrupttype[] = {
   "rst",  "suc",  "p/a",  "3",
   "term", "5",	  "6",	  "7",
@@ -477,7 +477,7 @@ void print_scsi_status(unsigned int ssr)
 static
 void print_sbic_status(int asr, int ssr, int cmdphase)
 {
-#ifdef CONFIG_ACORNSCSI_CONSTANTS
+#ifdef ACORNSCSI_CONSTANTS
     printk("sbic: %c%c%c%c%c%c ",
 	    asr & ASR_INT ? 'I' : 'i',
 	    asr & ASR_LCI ? 'L' : 'l',
Signature · 1 line
2.1.4
#1 lore raw
Quoted 6 lines · Valentin Rothberg

CONFIG_ACORNSCSI_CONSTANTS is a file local CPP identifier and thereby violates the naming convention of Kconfig options in Make and CPP syntax; only Kconfig options should carry the 'CONFIG_' prefix.

This patch removes the 'CONFIG_' prefix to apply to this convention and to make static analysis tools happy.

Will the Erlangen bot still spot ACORNSCSI_CONSTANTS as a potential issue?

Paul Bolle

#2 lore raw

Hi Paul,

Quoted 10 lines · Paul Bolle
Quoted 6 lines · Valentin Rothberg

CONFIG_ACORNSCSI_CONSTANTS is a file local CPP identifier and thereby violates the naming convention of Kconfig options in Make and CPP syntax; only Kconfig options should carry the 'CONFIG_' prefix.

This patch removes the 'CONFIG_' prefix to apply to this convention and to make static analysis tools happy.

Will the Erlangen bot still spot ACORNSCSI_CONSTANTS as a potential issue?

No, undertaker-checkpatch won't complain about this. There are thousands of such cases (i.e., without CONFIG_ prefix) around in the code (mostly #ifdef DEBUG). But most of them are intentionally dead or related to debugging, so they are ignored to avoid having false positives.

Regards,
Valentin

Quoted 2 lines

Paul Bolle

#3 lore raw

Hi Valentin,

Quoted 9 lines · Valentin Rothberg
Quoted 2 lines · Paul Bolle

Will the Erlangen bot still spot ACORNSCSI_CONSTANTS as a potential issue?

No, undertaker-checkpatch won't complain about this. There are thousands of such cases (i.e., without CONFIG_ prefix) around in the code (mostly #ifdef DEBUG). But most of them are intentionally dead or related to debugging, so they are ignored to avoid having false positives.

Well, in a few years time, once undertaker-checkpatch has stomped out most of the faux Kconfig preprocessor checks, that might be an area to cover too. Or is that issue, ie pointless preprocessor checks, harder than one might naively think?

Thanks,

Paul Bolle

#4 lore raw
Quoted 13 lines · Paul Bolle

Hi Valentin,

Quoted 9 lines · Valentin Rothberg
Quoted 2 lines · Paul Bolle

Will the Erlangen bot still spot ACORNSCSI_CONSTANTS as a potential issue?

No, undertaker-checkpatch won't complain about this. There are thousands of such cases (i.e., without CONFIG_ prefix) around in the code (mostly #ifdef DEBUG). But most of them are intentionally dead or related to debugging, so they are ignored to avoid having false positives.

Quoted 4 lines

Well, in a few years time, once undertaker-checkpatch has stomped out most of the faux Kconfig preprocessor checks, that might be an area to cover too. Or is that issue, ie pointless preprocessor checks, harder than one might naively think?

To give a number from Linus' tree today: 4706 of such unprefixed dead and undead #ifdefs and 940 'real' ones. Most of them are intentional -- this doesn't mean that it's not a problem. Personally, I don't like to have code around that cannot at least be easily test compiled;
we manually need to (un)define the identifiers.

I like your proposal. For symbolic issues, we could even put that directly in checkkconfigsymbols.py, so a big part of the problem could be solved directly with Kernel tools. But I really hope that it won't take a few years until we get there : )

Valentin

Quoted 5 lines

Thanks,

Paul Bolle

Acked-by Russell King on v1 1/1
#5 lore raw
In reply to Valentin Rothberg
Quoted 8 lines · Valentin Rothberg

CONFIG_ACORNSCSI_CONSTANTS is a file local CPP identifier and thereby violates the naming convention of Kconfig options in Make and CPP syntax; only Kconfig options should carry the 'CONFIG_' prefix.

This patch removes the 'CONFIG_' prefix to apply to this convention and to make static analysis tools happy.

As you're merely renaming all instances of this:

Thanks.

Signature · 2 lines
FTTC broadband for 0.8mile line: currently at 10.5Mbps down 400kbps up
according to speedtest.net.