net: fddi: skfp: Include generic PCI definitions from pci_regs.h
Subsystems (2)
PATCH 1/1 net: fddi: skfp: Include generic PCI definitions from pci_regs.h
Include the generic PCI definitions from include/uapi/linux/pci_regs.h change PCI_REV_ID to PCI_REVISION_ID to make it compatible with the generic define.
This driver uses only one generic PCI define.
--- drivers/net/fddi/skfp/drvfbi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/fddi/skfp/drvfbi.c +2 -1
@@ -20,6 +20,7 @@#include "h/supern_2.h" #include "h/skfbiinc.h" #include <linux/bitrev.h> +#include <uapi/linux/pci_regs.h> #ifndef lintstatic const char ID_sccs[] ="@(#)drvfbi.c 1.63 99/02/11 (C) SK " ;
@@ -127,7 +128,7 @@ static void card_start(struct s_smc *smc)* at very first before any other initialization functions is * executed. */ - rev_id =inp (PCI_C (PCI_REV_ID )) ; + rev_id =inp (PCI_C (PCI_REVISION_ID )) ;if ((rev_id &0xf0 ) ==SK_ML_ID_1 || (rev_id &0xf0 ) ==SK_ML_ID_2 ) { smc->hw.hw_is_64bit = TRUE ; }else {
Signature · 1 line
2.21.0
Quoted 24 lines · Puranjay Mohan
Include the generic PCI definitions from
include/uapi/linux/pci_regs.hchangePCI_REV_IDtoPCI_REVISION_IDto make it compatible with the generic define.
This driver uses only one generic PCI define.Signed-off-by Puranjay Mohan <puranjay12 at gmail.com>--- drivers/net/fddi/skfp/drvfbi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/net/fddi/skfp/drvfbi.c +1
-0@@ -20,6 +20,7 @@#include "h/supern_2.h" #include "h/skfbiinc.h" #include <linux/bitrev.h> +#include <uapi/linux/pci_regs.h> #ifndef lintstatic const char ID_sccs[] ="@(#)drvfbi.c 1.63 99/02/11 (C) SK " ;
drivers/net/fddi/skfp/drvfbi.c static void card_start(struct s_smc *smc) · line 134
Quoted 8 lines
drivers/net/fddi/skfp/drvfbi.c +1
-1@@ -127,7 +128,7 @@ static void card_start(struct s_smc *smc)* at very first before any other initialization functions is * executed. */ - rev_id =inp (PCI_C (PCI_REV_ID )) ; + rev_id =inp (PCI_C (PCI_REVISION_ID )) ;if ((rev_id &0xf0 ) ==SK_ML_ID_1 || (rev_id &0xf0 ) ==SK_ML_ID_2 ) { smc->hw.hw_is_64bit = TRUE ; }else {
Why not delete the PCI_REV_ID define in:
drivers/net/fddi/skfp/h/skfbi.h
It looks like this header has duplicate PCI config space header defines, not just this one. Some of them are slightly different names:
e.g:
#define PCI_CACHE_LSZ 0x0c /* 8 bit Cache Line Size */
Looks like it defines the standard PCI config space instead of including and using the standard defines from uapi/linux/pci_regs.h
Something to look into.
thanks,
-- Shuah
A message that is not in the archive (sent off-list or not yet received): 20190619182122.GA4827@arch
PATCH 1/1 net: fddi: skfp: Include generic PCI definitions from pci_regs.h
Quoted 35 lines · Shuah Khan
Quoted 23 lines · Puranjay Mohan
Include the generic PCI definitions from
include/uapi/linux/pci_regs.hchangePCI_REV_IDtoPCI_REVISION_IDto make it compatible with the generic define.
This driver uses only one generic PCI define.Signed-off-by Puranjay Mohan <puranjay12 at gmail.com>--- drivers/net/fddi/skfp/drvfbi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/net/fddi/skfp/drvfbi.c +1
-0@@ -20,6 +20,7 @@#include "h/supern_2.h" #include "h/skfbiinc.h" #include <linux/bitrev.h> +#include <uapi/linux/pci_regs.h> #ifndef lintstatic const char ID_sccs[] ="@(#)drvfbi.c 1.63 99/02/11 (C) SK " ;drivers/net/fddi/skfp/drvfbi.c static void card_start(struct s_smc *smc) · line 134
Quoted 8 lines
drivers/net/fddi/skfp/drvfbi.c +1
-1@@ -127,7 +128,7 @@ static void card_start(struct s_smc *smc)* at very first before any other initialization functions is * executed. */ - rev_id =inp (PCI_C (PCI_REV_ID )) ; + rev_id =inp (PCI_C (PCI_REVISION_ID )) ;if ((rev_id &0xf0 ) ==SK_ML_ID_1 || (rev_id &0xf0 ) ==SK_ML_ID_2 ) { smc->hw.hw_is_64bit = TRUE ; }else {Why not delete the
PCI_REV_IDdefine in:
Quoted 2 lines
drivers/net/fddi/skfp/h/skfbi.h
I have removed all generic PCI definitions from skfbi.h in the next patch which I have sent, I wanted to keep it organised by sending two patches
Quoted 10 lines
It looks like this header has duplicate PCI config space header defines, not just this one. Some of them are slightly different names:
e.g:
#define
PCI_CACHE_LSZ0x0c/* 8 bit Cache Line Size */Looks like it defines the standard PCI config space instead of including and using the standard defines from
uapi/linux/pci_regs.h
It defines many duplicate definitions in skfbi.h, but only uses one of them, hence they are removed in the next patch as told by bjorn. It uses only one generic PCI define in driver code, i.e. PCI_REV_ID, it has been replaced by PCI_REVISION_ID to make it work with the define included with uapi/linux/pci_regs.h
Quoted 9 lines
Something to look into.
thanks,
-- Shuah
Signature · 5 lines
Thanks and Regards Yours Truly, Puranjay Mohan
Quoted 36 lines · Puranjay Mohan
Quoted 35 lines · Shuah Khan
Quoted 23 lines · Puranjay Mohan
Include the generic PCI definitions from
include/uapi/linux/pci_regs.hchangePCI_REV_IDtoPCI_REVISION_IDto make it compatible with the generic define.
This driver uses only one generic PCI define.Signed-off-by Puranjay Mohan <puranjay12 at gmail.com>--- drivers/net/fddi/skfp/drvfbi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/net/fddi/skfp/drvfbi.c +1
-0@@ -20,6 +20,7 @@#include "h/supern_2.h" #include "h/skfbiinc.h" #include <linux/bitrev.h> +#include <uapi/linux/pci_regs.h> #ifndef lintstatic const char ID_sccs[] ="@(#)drvfbi.c 1.63 99/02/11 (C) SK " ;drivers/net/fddi/skfp/drvfbi.c static void card_start(struct s_smc *smc) · line 134
Quoted 8 lines
drivers/net/fddi/skfp/drvfbi.c +1
-1@@ -127,7 +128,7 @@ static void card_start(struct s_smc *smc)* at very first before any other initialization functions is * executed. */ - rev_id =inp (PCI_C (PCI_REV_ID )) ; + rev_id =inp (PCI_C (PCI_REVISION_ID )) ;if ((rev_id &0xf0 ) ==SK_ML_ID_1 || (rev_id &0xf0 ) ==SK_ML_ID_2 ) { smc->hw.hw_is_64bit = TRUE ; }else {Why not delete the
PCI_REV_IDdefine in:
Quoted 6 lines
Quoted 2 lines
drivers/net/fddi/skfp/h/skfbi.hI have removed all generic PCI definitions from
skfbi.hin the next patch which I have sent, I wanted to keep it organised by sending two patches
Yeah. I saw your second patch come in after I sent my response. :)
Quoted 7 lines
Quoted 7 lines
It looks like this header has duplicate PCI config space header defines, not just this one. Some of them are slightly different names:
e.g:
#define
PCI_CACHE_LSZ0x0c/* 8 bit Cache Line Size */
Quoted 10 lines
Quoted 3 lines
Looks like it defines the standard PCI config space instead of including and using the standard defines from
uapi/linux/pci_regs.hIt defines many duplicate definitions in
skfbi.h, but only uses one of them, hence they are removed in the next patch as told by bjorn. It uses only one generic PCI define in driver code, i.e.PCI_REV_ID, it has been replaced byPCI_REVISION_IDto make it work with the define included withuapi/linux/pci_regs.hQuoted 2 lines
Something to look into.
Sounds like a plan.
thanks,
-- Shuah
Quoted 5 lines · Puranjay Mohan
Include the generic PCI definitions from
include/uapi/linux/pci_regs.hchangePCI_REV_IDtoPCI_REVISION_IDto make it compatible with the generic define.
This driver uses only one generic PCI define.
1) Start every sentence with a capital letter.
2) Use a period at the end of every sentence.
3) Use a blank line between paragraphs. A short line (like "generic define" above) *suggests* a new paragraph, but it's ambiguous, which makes it hard to read.
4) This patch must build correctly by itself. I didn't try it, but I'm a little suspicious that including pci_regs.h will cause redefinition of PCI_STATUS and other #defines that are the same between pci_regs.h and skfbi.h. You could either combine the two patches, or make the first patch simply rename PCI_REV_ID to PCI_REVISION_ID in skfbi.h and drvfbi.c Then the second patch could add the #include of pci_regs.h and remove the corresponding #defines from skfbi.h. Maybe a third patch would remove all the other unused PCI_* definitions. Arguably the second and third could be combined. But it's much easier for a maintainer to squash patches together than to split them apart, so err on the side of splitting them up.
Quoted 29 lines
Signed-off-by Puranjay Mohan <puranjay12 at gmail.com>--- drivers/net/fddi/skfp/drvfbi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/net/fddi/skfp/drvfbi.c +2
-2@@ -20,6 +20,7 @@#include "h/supern_2.h" #include "h/skfbiinc.h" #include <linux/bitrev.h> +#include <uapi/linux/pci_regs.h> #ifndef lintstatic const char ID_sccs[] ="@(#)drvfbi.c 1.63 99/02/11 (C) SK " ;@@ -127,7 +128,7 @@ static void card_start(struct s_smc *smc)* at very first before any other initialization functions is * executed. */ - rev_id =inp (PCI_C (PCI_REV_ID )) ; + rev_id =inp (PCI_C (PCI_REVISION_ID )) ;if ((rev_id &0xf0 ) ==SK_ML_ID_1 || (rev_id &0xf0 ) ==SK_ML_ID_2 ) { smc->hw.hw_is_64bit = TRUE ; }else { --2.21.0
David Miller
maintainerFrom: Puranjay Mohan <[email protected]>
Date: Wed, 19 Jun 2019 23:15:56 +0530
Quoted 15 lines
Include the generic PCI definitions from
include/uapi/linux/pci_regs.hchangePCI_REV_IDtoPCI_REVISION_IDto make it compatible with the generic define.
This driver uses only one generic PCI define.Signed-off-by Puranjay Mohan <puranjay12 at gmail.com>--- drivers/net/fddi/skfp/drvfbi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/net/fddi/skfp/drvfbi.c +0
-0
drivers/net/fddi/skfp/drvfbi.c line 23
Quoted 4 lines
drivers/net/fddi/skfp/drvfbi.c +1
-0@@ -20,6 +20,7 @@#include "h/supern_2.h" #include "h/skfbiinc.h" #include <linux/bitrev.h> +#include <uapi/linux/pci_regs.h>
You never need to use "uapi/" in header includes from the kernel source, please just use "linux/pci_regs.h"
Thank you.