Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, Mar 25, 2016 at 09:58:21AM +0100, Boris Brezillon wrote: > On Fri, 25 Mar 2016 08:57:59 +0100 > Richard Weinbergerwrote: > > > Am 25.03.2016 um 07:31 schrieb Yaowei Bai: > > > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > > >> ccing: Brian and Richard > > >> > > >> Hi Yao, > > >> Is that really necessary? I am not sure how much benefit we can > > >> achieve from this change. > > >> Could you explain more? > > > > > > Yes, according to these functions' name, a boolean return value is more > > > suitable and matchable. > > > > > > Also personally think this change maybe benfit function's return value > > > storage in the stack when called on certain architectures. > > > > On which archs? And what exactly is the benefit? > > I agree that bool might be a better choice for new functions > > but here you're touching existing and working(!) code. > > The only outcome is git history pollution that makes git blame > > less efficient. > > Indeed, you raised a good point. Having useless changes pollute git > blame output may be problematic. Not sure I want to apply those patches > anymore :-/. > > Anyway, Yao, I'm sure you can find other usefull things to contribute. OK, thanks for reviewing. > > -- > Boris Brezillon, Free Electrons > Embedded Linux and Kernel engineering > http://free-electrons.com
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, Mar 25, 2016 at 09:58:21AM +0100, Boris Brezillon wrote: > On Fri, 25 Mar 2016 08:57:59 +0100 > Richard Weinberger wrote: > > > Am 25.03.2016 um 07:31 schrieb Yaowei Bai: > > > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > > >> ccing: Brian and Richard > > >> > > >> Hi Yao, > > >> Is that really necessary? I am not sure how much benefit we can > > >> achieve from this change. > > >> Could you explain more? > > > > > > Yes, according to these functions' name, a boolean return value is more > > > suitable and matchable. > > > > > > Also personally think this change maybe benfit function's return value > > > storage in the stack when called on certain architectures. > > > > On which archs? And what exactly is the benefit? > > I agree that bool might be a better choice for new functions > > but here you're touching existing and working(!) code. > > The only outcome is git history pollution that makes git blame > > less efficient. > > Indeed, you raised a good point. Having useless changes pollute git > blame output may be problematic. Not sure I want to apply those patches > anymore :-/. > > Anyway, Yao, I'm sure you can find other usefull things to contribute. OK, thanks for reviewing. > > -- > Boris Brezillon, Free Electrons > Embedded Linux and Kernel engineering > http://free-electrons.com
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, 25 Mar 2016 08:57:59 +0100 Richard Weinbergerwrote: > Am 25.03.2016 um 07:31 schrieb Yaowei Bai: > > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > >> ccing: Brian and Richard > >> > >> Hi Yao, > >> Is that really necessary? I am not sure how much benefit we can > >> achieve from this change. > >> Could you explain more? > > > > Yes, according to these functions' name, a boolean return value is more > > suitable and matchable. > > > > Also personally think this change maybe benfit function's return value > > storage in the stack when called on certain architectures. > > On which archs? And what exactly is the benefit? > I agree that bool might be a better choice for new functions > but here you're touching existing and working(!) code. > The only outcome is git history pollution that makes git blame > less efficient. Indeed, you raised a good point. Having useless changes pollute git blame output may be problematic. Not sure I want to apply those patches anymore :-/. Anyway, Yao, I'm sure you can find other usefull things to contribute. -- Boris Brezillon, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, 25 Mar 2016 08:57:59 +0100 Richard Weinberger wrote: > Am 25.03.2016 um 07:31 schrieb Yaowei Bai: > > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > >> ccing: Brian and Richard > >> > >> Hi Yao, > >> Is that really necessary? I am not sure how much benefit we can > >> achieve from this change. > >> Could you explain more? > > > > Yes, according to these functions' name, a boolean return value is more > > suitable and matchable. > > > > Also personally think this change maybe benfit function's return value > > storage in the stack when called on certain architectures. > > On which archs? And what exactly is the benefit? > I agree that bool might be a better choice for new functions > but here you're touching existing and working(!) code. > The only outcome is git history pollution that makes git blame > less efficient. Indeed, you raised a good point. Having useless changes pollute git blame output may be problematic. Not sure I want to apply those patches anymore :-/. Anyway, Yao, I'm sure you can find other usefull things to contribute. -- Boris Brezillon, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, Mar 25, 2016 at 08:51:55AM +0100, Boris Brezillon wrote: > Hi Yao, > > On Fri, 25 Mar 2016 14:31:53 +0800 > Yaowei Baiwrote: > > > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > > > ccing: Brian and Richard > > > > > > Hi Yao, > > > Is that really necessary? I am not sure how much benefit we can > > > achieve from this change. > > > Could you explain more? > > > > Yes, according to these functions' name, a boolean return value is more > > suitable and matchable. > > > > Also personally think this change maybe benfit function's return value > > storage in the stack when called on certain architectures. > > At least be honest, and say this is for your patchcount statistics :P. > I'm fine taking the NAND related ones, but please, next time find > something more useful to submit. That's very nice of you, thanks. > > Best Regards, > > Boris > > > > > > > > > Yang > > > > > > On 03/25/2016 10:41 AM, Yaowei Bai wrote: > > > >This series only make several funcitons return bool to improve > > > >readability, no other funcitonal changes. > > > > > > > >Yaowei Bai (5): > > > > drivers/mtd: mtd_is_partition can be boolean > > > > drivers/mtd: cfi_interleave_supported can be boolean > > > > drivers/mtd: map_bankwidth_supported can be boolean > > > > drivers/mtd: mtd_nand_has_bch can be boolean > > > > drivers/mtd/nand: nand_opcode_8bits can be boolean > > > > > > > > drivers/mtd/mtdpart.c | 6 +++--- > > > > include/linux/mtd/cfi.h| 6 +++--- > > > > include/linux/mtd/map.h| 6 +++--- > > > > include/linux/mtd/nand.h | 6 +++--- > > > > include/linux/mtd/nand_bch.h | 4 ++-- > > > > include/linux/mtd/partitions.h | 2 +- > > > > 6 files changed, 15 insertions(+), 15 deletions(-) > > > > > > > > > > > > > > > -- > Boris Brezillon, Free Electrons > Embedded Linux and Kernel engineering > http://free-electrons.com
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, Mar 25, 2016 at 08:51:55AM +0100, Boris Brezillon wrote: > Hi Yao, > > On Fri, 25 Mar 2016 14:31:53 +0800 > Yaowei Bai wrote: > > > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > > > ccing: Brian and Richard > > > > > > Hi Yao, > > > Is that really necessary? I am not sure how much benefit we can > > > achieve from this change. > > > Could you explain more? > > > > Yes, according to these functions' name, a boolean return value is more > > suitable and matchable. > > > > Also personally think this change maybe benfit function's return value > > storage in the stack when called on certain architectures. > > At least be honest, and say this is for your patchcount statistics :P. > I'm fine taking the NAND related ones, but please, next time find > something more useful to submit. That's very nice of you, thanks. > > Best Regards, > > Boris > > > > > > > > > Yang > > > > > > On 03/25/2016 10:41 AM, Yaowei Bai wrote: > > > >This series only make several funcitons return bool to improve > > > >readability, no other funcitonal changes. > > > > > > > >Yaowei Bai (5): > > > > drivers/mtd: mtd_is_partition can be boolean > > > > drivers/mtd: cfi_interleave_supported can be boolean > > > > drivers/mtd: map_bankwidth_supported can be boolean > > > > drivers/mtd: mtd_nand_has_bch can be boolean > > > > drivers/mtd/nand: nand_opcode_8bits can be boolean > > > > > > > > drivers/mtd/mtdpart.c | 6 +++--- > > > > include/linux/mtd/cfi.h| 6 +++--- > > > > include/linux/mtd/map.h| 6 +++--- > > > > include/linux/mtd/nand.h | 6 +++--- > > > > include/linux/mtd/nand_bch.h | 4 ++-- > > > > include/linux/mtd/partitions.h | 2 +- > > > > 6 files changed, 15 insertions(+), 15 deletions(-) > > > > > > > > > > > > > > > -- > Boris Brezillon, Free Electrons > Embedded Linux and Kernel engineering > http://free-electrons.com
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, Mar 25, 2016 at 08:57:59AM +0100, Richard Weinberger wrote: > Am 25.03.2016 um 07:31 schrieb Yaowei Bai: > > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > >> ccing: Brian and Richard > >> > >> Hi Yao, > >> Is that really necessary? I am not sure how much benefit we can > >> achieve from this change. > >> Could you explain more? > > > > Yes, according to these functions' name, a boolean return value is more > > suitable and matchable. > > > > Also personally think this change maybe benfit function's return value > > storage in the stack when called on certain architectures. > > On which archs? And what exactly is the benefit? > I agree that bool might be a better choice for new functions > but here you're touching existing and working(!) code. > The only outcome is git history pollution that makes git blame > less efficient. Working code doesn't mean perfect code. :-) I still think this's a helpful change even though it's small, but you make the decision to merge or drop it. > > Thanks, > //richard >
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, Mar 25, 2016 at 08:57:59AM +0100, Richard Weinberger wrote: > Am 25.03.2016 um 07:31 schrieb Yaowei Bai: > > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > >> ccing: Brian and Richard > >> > >> Hi Yao, > >> Is that really necessary? I am not sure how much benefit we can > >> achieve from this change. > >> Could you explain more? > > > > Yes, according to these functions' name, a boolean return value is more > > suitable and matchable. > > > > Also personally think this change maybe benfit function's return value > > storage in the stack when called on certain architectures. > > On which archs? And what exactly is the benefit? > I agree that bool might be a better choice for new functions > but here you're touching existing and working(!) code. > The only outcome is git history pollution that makes git blame > less efficient. Working code doesn't mean perfect code. :-) I still think this's a helpful change even though it's small, but you make the decision to merge or drop it. > > Thanks, > //richard >
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
Am 25.03.2016 um 07:31 schrieb Yaowei Bai: > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: >> ccing: Brian and Richard >> >> Hi Yao, >> Is that really necessary? I am not sure how much benefit we can >> achieve from this change. >> Could you explain more? > > Yes, according to these functions' name, a boolean return value is more > suitable and matchable. > > Also personally think this change maybe benfit function's return value > storage in the stack when called on certain architectures. On which archs? And what exactly is the benefit? I agree that bool might be a better choice for new functions but here you're touching existing and working(!) code. The only outcome is git history pollution that makes git blame less efficient. Thanks, //richard
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
Am 25.03.2016 um 07:31 schrieb Yaowei Bai: > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: >> ccing: Brian and Richard >> >> Hi Yao, >> Is that really necessary? I am not sure how much benefit we can >> achieve from this change. >> Could you explain more? > > Yes, according to these functions' name, a boolean return value is more > suitable and matchable. > > Also personally think this change maybe benfit function's return value > storage in the stack when called on certain architectures. On which archs? And what exactly is the benefit? I agree that bool might be a better choice for new functions but here you're touching existing and working(!) code. The only outcome is git history pollution that makes git blame less efficient. Thanks, //richard
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
Hi Yao, On Fri, 25 Mar 2016 14:31:53 +0800 Yaowei Baiwrote: > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > > ccing: Brian and Richard > > > > Hi Yao, > > Is that really necessary? I am not sure how much benefit we can > > achieve from this change. > > Could you explain more? > > Yes, according to these functions' name, a boolean return value is more > suitable and matchable. > > Also personally think this change maybe benfit function's return value > storage in the stack when called on certain architectures. At least be honest, and say this is for your patchcount statistics :P. I'm fine taking the NAND related ones, but please, next time find something more useful to submit. Best Regards, Boris > > > > > Yang > > > > On 03/25/2016 10:41 AM, Yaowei Bai wrote: > > >This series only make several funcitons return bool to improve > > >readability, no other funcitonal changes. > > > > > >Yaowei Bai (5): > > > drivers/mtd: mtd_is_partition can be boolean > > > drivers/mtd: cfi_interleave_supported can be boolean > > > drivers/mtd: map_bankwidth_supported can be boolean > > > drivers/mtd: mtd_nand_has_bch can be boolean > > > drivers/mtd/nand: nand_opcode_8bits can be boolean > > > > > > drivers/mtd/mtdpart.c | 6 +++--- > > > include/linux/mtd/cfi.h| 6 +++--- > > > include/linux/mtd/map.h| 6 +++--- > > > include/linux/mtd/nand.h | 6 +++--- > > > include/linux/mtd/nand_bch.h | 4 ++-- > > > include/linux/mtd/partitions.h | 2 +- > > > 6 files changed, 15 insertions(+), 15 deletions(-) > > > > > > > -- Boris Brezillon, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
Hi Yao, On Fri, 25 Mar 2016 14:31:53 +0800 Yaowei Bai wrote: > On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > > ccing: Brian and Richard > > > > Hi Yao, > > Is that really necessary? I am not sure how much benefit we can > > achieve from this change. > > Could you explain more? > > Yes, according to these functions' name, a boolean return value is more > suitable and matchable. > > Also personally think this change maybe benfit function's return value > storage in the stack when called on certain architectures. At least be honest, and say this is for your patchcount statistics :P. I'm fine taking the NAND related ones, but please, next time find something more useful to submit. Best Regards, Boris > > > > > Yang > > > > On 03/25/2016 10:41 AM, Yaowei Bai wrote: > > >This series only make several funcitons return bool to improve > > >readability, no other funcitonal changes. > > > > > >Yaowei Bai (5): > > > drivers/mtd: mtd_is_partition can be boolean > > > drivers/mtd: cfi_interleave_supported can be boolean > > > drivers/mtd: map_bankwidth_supported can be boolean > > > drivers/mtd: mtd_nand_has_bch can be boolean > > > drivers/mtd/nand: nand_opcode_8bits can be boolean > > > > > > drivers/mtd/mtdpart.c | 6 +++--- > > > include/linux/mtd/cfi.h| 6 +++--- > > > include/linux/mtd/map.h| 6 +++--- > > > include/linux/mtd/nand.h | 6 +++--- > > > include/linux/mtd/nand_bch.h | 4 ++-- > > > include/linux/mtd/partitions.h | 2 +- > > > 6 files changed, 15 insertions(+), 15 deletions(-) > > > > > > > -- Boris Brezillon, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > ccing: Brian and Richard > > Hi Yao, > Is that really necessary? I am not sure how much benefit we can > achieve from this change. > Could you explain more? Yes, according to these functions' name, a boolean return value is more suitable and matchable. Also personally think this change maybe benfit function's return value storage in the stack when called on certain architectures. > > Yang > > On 03/25/2016 10:41 AM, Yaowei Bai wrote: > >This series only make several funcitons return bool to improve > >readability, no other funcitonal changes. > > > >Yaowei Bai (5): > > drivers/mtd: mtd_is_partition can be boolean > > drivers/mtd: cfi_interleave_supported can be boolean > > drivers/mtd: map_bankwidth_supported can be boolean > > drivers/mtd: mtd_nand_has_bch can be boolean > > drivers/mtd/nand: nand_opcode_8bits can be boolean > > > > drivers/mtd/mtdpart.c | 6 +++--- > > include/linux/mtd/cfi.h| 6 +++--- > > include/linux/mtd/map.h| 6 +++--- > > include/linux/mtd/nand.h | 6 +++--- > > include/linux/mtd/nand_bch.h | 4 ++-- > > include/linux/mtd/partitions.h | 2 +- > > 6 files changed, 15 insertions(+), 15 deletions(-) > > >
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
On Fri, Mar 25, 2016 at 10:54:51AM +0800, Dongsheng Yang wrote: > ccing: Brian and Richard > > Hi Yao, > Is that really necessary? I am not sure how much benefit we can > achieve from this change. > Could you explain more? Yes, according to these functions' name, a boolean return value is more suitable and matchable. Also personally think this change maybe benfit function's return value storage in the stack when called on certain architectures. > > Yang > > On 03/25/2016 10:41 AM, Yaowei Bai wrote: > >This series only make several funcitons return bool to improve > >readability, no other funcitonal changes. > > > >Yaowei Bai (5): > > drivers/mtd: mtd_is_partition can be boolean > > drivers/mtd: cfi_interleave_supported can be boolean > > drivers/mtd: map_bankwidth_supported can be boolean > > drivers/mtd: mtd_nand_has_bch can be boolean > > drivers/mtd/nand: nand_opcode_8bits can be boolean > > > > drivers/mtd/mtdpart.c | 6 +++--- > > include/linux/mtd/cfi.h| 6 +++--- > > include/linux/mtd/map.h| 6 +++--- > > include/linux/mtd/nand.h | 6 +++--- > > include/linux/mtd/nand_bch.h | 4 ++-- > > include/linux/mtd/partitions.h | 2 +- > > 6 files changed, 15 insertions(+), 15 deletions(-) > > >
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
ccing: Brian and Richard Hi Yao, Is that really necessary? I am not sure how much benefit we can achieve from this change. Could you explain more? Yang On 03/25/2016 10:41 AM, Yaowei Bai wrote: This series only make several funcitons return bool to improve readability, no other funcitonal changes. Yaowei Bai (5): drivers/mtd: mtd_is_partition can be boolean drivers/mtd: cfi_interleave_supported can be boolean drivers/mtd: map_bankwidth_supported can be boolean drivers/mtd: mtd_nand_has_bch can be boolean drivers/mtd/nand: nand_opcode_8bits can be boolean drivers/mtd/mtdpart.c | 6 +++--- include/linux/mtd/cfi.h| 6 +++--- include/linux/mtd/map.h| 6 +++--- include/linux/mtd/nand.h | 6 +++--- include/linux/mtd/nand_bch.h | 4 ++-- include/linux/mtd/partitions.h | 2 +- 6 files changed, 15 insertions(+), 15 deletions(-)
Re: [PATCH 0/5] drivers/mtd: make several functions return bool
ccing: Brian and Richard Hi Yao, Is that really necessary? I am not sure how much benefit we can achieve from this change. Could you explain more? Yang On 03/25/2016 10:41 AM, Yaowei Bai wrote: This series only make several funcitons return bool to improve readability, no other funcitonal changes. Yaowei Bai (5): drivers/mtd: mtd_is_partition can be boolean drivers/mtd: cfi_interleave_supported can be boolean drivers/mtd: map_bankwidth_supported can be boolean drivers/mtd: mtd_nand_has_bch can be boolean drivers/mtd/nand: nand_opcode_8bits can be boolean drivers/mtd/mtdpart.c | 6 +++--- include/linux/mtd/cfi.h| 6 +++--- include/linux/mtd/map.h| 6 +++--- include/linux/mtd/nand.h | 6 +++--- include/linux/mtd/nand_bch.h | 4 ++-- include/linux/mtd/partitions.h | 2 +- 6 files changed, 15 insertions(+), 15 deletions(-)
[PATCH 0/5] drivers/mtd: make several functions return bool
This series only make several funcitons return bool to improve readability, no other funcitonal changes. Yaowei Bai (5): drivers/mtd: mtd_is_partition can be boolean drivers/mtd: cfi_interleave_supported can be boolean drivers/mtd: map_bankwidth_supported can be boolean drivers/mtd: mtd_nand_has_bch can be boolean drivers/mtd/nand: nand_opcode_8bits can be boolean drivers/mtd/mtdpart.c | 6 +++--- include/linux/mtd/cfi.h| 6 +++--- include/linux/mtd/map.h| 6 +++--- include/linux/mtd/nand.h | 6 +++--- include/linux/mtd/nand_bch.h | 4 ++-- include/linux/mtd/partitions.h | 2 +- 6 files changed, 15 insertions(+), 15 deletions(-) -- 1.9.1
[PATCH 0/5] drivers/mtd: make several functions return bool
This series only make several funcitons return bool to improve readability, no other funcitonal changes. Yaowei Bai (5): drivers/mtd: mtd_is_partition can be boolean drivers/mtd: cfi_interleave_supported can be boolean drivers/mtd: map_bankwidth_supported can be boolean drivers/mtd: mtd_nand_has_bch can be boolean drivers/mtd/nand: nand_opcode_8bits can be boolean drivers/mtd/mtdpart.c | 6 +++--- include/linux/mtd/cfi.h| 6 +++--- include/linux/mtd/map.h| 6 +++--- include/linux/mtd/nand.h | 6 +++--- include/linux/mtd/nand_bch.h | 4 ++-- include/linux/mtd/partitions.h | 2 +- 6 files changed, 15 insertions(+), 15 deletions(-) -- 1.9.1