[PATCH 1/1] block: Use bits.h macros to improve readability

[PATCH 1/1] block: Use bits.h macros to improve readability

[Leonardo Bras]

Applies some bits.h macros in order to improve readability of
linux/blk_types.h.

Signed-off-by: Leonardo Bras <leonardo@linux.ibm.com>
---
 include/linux/blk_types.h | 55 ++++++++++++++++++++-------------------
 1 file changed, 28 insertions(+), 27 deletions(-)

diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index 95202f80676c..31c8c6d274f6 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -9,6 +9,7 @@
 #include <linux/types.h>
 #include <linux/bvec.h>
 #include <linux/ktime.h>
+#include <linux/bits.h>
 
 struct bio_set;
 struct bio;
@@ -101,13 +102,13 @@ static inline bool blk_path_error(blk_status_t error)
 #define BIO_ISSUE_SIZE_BITS     12
 #define BIO_ISSUE_RES_SHIFT     (64 - BIO_ISSUE_RES_BITS)
 #define BIO_ISSUE_SIZE_SHIFT    (BIO_ISSUE_RES_SHIFT - BIO_ISSUE_SIZE_BITS)
-#define BIO_ISSUE_TIME_MASK     ((1ULL << BIO_ISSUE_SIZE_SHIFT) - 1)
+#define BIO_ISSUE_TIME_MASK     GENMASK_ULL(BIO_ISSUE_SIZE_SHIFT - 1, 0)
 #define BIO_ISSUE_SIZE_MASK     \
-	(((1ULL << BIO_ISSUE_SIZE_BITS) - 1) << BIO_ISSUE_SIZE_SHIFT)
-#define BIO_ISSUE_RES_MASK      (~((1ULL << BIO_ISSUE_RES_SHIFT) - 1))
+	GENMASK_ULL(BIO_ISSUE_RES_SHIFT - 1, BIO_ISSUE_SIZE_SHIFT)
+#define BIO_ISSUE_RES_MASK      GENMASK_ULL(64, BIO_ISSUE_RES_SHIFT)
 
 /* Reserved bit for blk-throtl */
-#define BIO_ISSUE_THROTL_SKIP_LATENCY (1ULL << 63)
+#define BIO_ISSUE_THROTL_SKIP_LATENCY   BIT_ULL(63)
 
 struct bio_issue {
 	u64 value;
@@ -131,7 +132,7 @@ static inline sector_t bio_issue_size(struct bio_issue *issue)
 static inline void bio_issue_init(struct bio_issue *issue,
 				       sector_t size)
 {
-	size &= (1ULL << BIO_ISSUE_SIZE_BITS) - 1;
+	size &= GENMASK_ULL(BIO_ISSUE_SIZE_BITS - 1, 0);
 	issue->value = ((issue->value & BIO_ISSUE_RES_MASK) |
 			(ktime_get_ns() & BIO_ISSUE_TIME_MASK) |
 			((u64)size << BIO_ISSUE_SIZE_SHIFT));
@@ -270,7 +271,7 @@ typedef __u32 __bitwise blk_mq_req_flags_t;
  * meaning.
  */
 #define REQ_OP_BITS	8
-#define REQ_OP_MASK	((1 << REQ_OP_BITS) - 1)
+#define REQ_OP_MASK	GENMASK(REQ_OP_BITS - 1, 0)
 #define REQ_FLAG_BITS	24
 
 enum req_opf {
@@ -329,25 +330,25 @@ enum req_flag_bits {
 	__REQ_NR_BITS,		/* stops here */
 };
 
-#define REQ_FAILFAST_DEV	(1ULL << __REQ_FAILFAST_DEV)
-#define REQ_FAILFAST_TRANSPORT	(1ULL << __REQ_FAILFAST_TRANSPORT)
-#define REQ_FAILFAST_DRIVER	(1ULL << __REQ_FAILFAST_DRIVER)
-#define REQ_SYNC		(1ULL << __REQ_SYNC)
-#define REQ_META		(1ULL << __REQ_META)
-#define REQ_PRIO		(1ULL << __REQ_PRIO)
-#define REQ_NOMERGE		(1ULL << __REQ_NOMERGE)
-#define REQ_IDLE		(1ULL << __REQ_IDLE)
-#define REQ_INTEGRITY		(1ULL << __REQ_INTEGRITY)
-#define REQ_FUA			(1ULL << __REQ_FUA)
-#define REQ_PREFLUSH		(1ULL << __REQ_PREFLUSH)
-#define REQ_RAHEAD		(1ULL << __REQ_RAHEAD)
-#define REQ_BACKGROUND		(1ULL << __REQ_BACKGROUND)
-#define REQ_NOWAIT		(1ULL << __REQ_NOWAIT)
-#define REQ_NOUNMAP		(1ULL << __REQ_NOUNMAP)
-#define REQ_HIPRI		(1ULL << __REQ_HIPRI)
-
-#define REQ_DRV			(1ULL << __REQ_DRV)
-#define REQ_SWAP		(1ULL << __REQ_SWAP)
+#define REQ_FAILFAST_DEV	BIT_ULL(__REQ_FAILFAST_DEV)
+#define REQ_FAILFAST_TRANSPORT	BIT_ULL(__REQ_FAILFAST_TRANSPORT)
+#define REQ_FAILFAST_DRIVER	BIT_ULL(__REQ_FAILFAST_DRIVER)
+#define REQ_SYNC		BIT_ULL(__REQ_SYNC)
+#define REQ_META		BIT_ULL(__REQ_META)
+#define REQ_PRIO		BIT_ULL(__REQ_PRIO)
+#define REQ_NOMERGE		BIT_ULL(__REQ_NOMERGE)
+#define REQ_IDLE		BIT_ULL(__REQ_IDLE)
+#define REQ_INTEGRITY		BIT_ULL(__REQ_INTEGRITY)
+#define REQ_FUA			BIT_ULL(__REQ_FUA)
+#define REQ_PREFLUSH		BIT_ULL(__REQ_PREFLUSH)
+#define REQ_RAHEAD		BIT_ULL(__REQ_RAHEAD)
+#define REQ_BACKGROUND		BIT_ULL(__REQ_BACKGROUND)
+#define REQ_NOWAIT		BIT_ULL(__REQ_NOWAIT)
+#define REQ_NOUNMAP		BIT_ULL(__REQ_NOUNMAP)
+#define REQ_HIPRI		BIT_ULL(__REQ_HIPRI)
+
+#define REQ_DRV			BIT_ULL(__REQ_DRV)
+#define REQ_SWAP		BIT_ULL(__REQ_SWAP)
 
 #define REQ_FAILFAST_MASK \
 	(REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT | REQ_FAILFAST_DRIVER)
@@ -415,7 +416,7 @@ static inline int op_stat_group(unsigned int op)
 typedef unsigned int blk_qc_t;
 #define BLK_QC_T_NONE		-1U
 #define BLK_QC_T_SHIFT		16
-#define BLK_QC_T_INTERNAL	(1U << 31)
+#define BLK_QC_T_INTERNAL	BIT(31)
 
 static inline bool blk_qc_t_valid(blk_qc_t cookie)
 {
@@ -429,7 +430,7 @@ static inline unsigned int blk_qc_t_to_queue_num(blk_qc_t cookie)
 
 static inline unsigned int blk_qc_t_to_tag(blk_qc_t cookie)
 {
-	return cookie & ((1u << BLK_QC_T_SHIFT) - 1);
+	return cookie & GENMASK(BLK_QC_T_SHIFT - 1, 0);
 }
 
 static inline bool blk_qc_t_is_internal(blk_qc_t cookie)
-- 
2.20.1

Re: [PATCH 1/1] block: Use bits.h macros to improve readability

[Jens Axboe]

On 8/1/19 6:00 PM, Leonardo Bras wrote:
> Applies some bits.h macros in order to improve readability of
> linux/blk_types.h.
> 
> Signed-off-by: Leonardo Bras <leonardo@linux.ibm.com>
> ---
>   include/linux/blk_types.h | 55 ++++++++++++++++++++-------------------
>   1 file changed, 28 insertions(+), 27 deletions(-)
> 
> diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
> index 95202f80676c..31c8c6d274f6 100644
> --- a/include/linux/blk_types.h
> +++ b/include/linux/blk_types.h
> @@ -9,6 +9,7 @@
>   #include <linux/types.h>
>   #include <linux/bvec.h>
>   #include <linux/ktime.h>
> +#include <linux/bits.h>
>   
>   struct bio_set;
>   struct bio;
> @@ -101,13 +102,13 @@ static inline bool blk_path_error(blk_status_t error)
>   #define BIO_ISSUE_SIZE_BITS     12
>   #define BIO_ISSUE_RES_SHIFT     (64 - BIO_ISSUE_RES_BITS)
>   #define BIO_ISSUE_SIZE_SHIFT    (BIO_ISSUE_RES_SHIFT - BIO_ISSUE_SIZE_BITS)
> -#define BIO_ISSUE_TIME_MASK     ((1ULL << BIO_ISSUE_SIZE_SHIFT) - 1)
> +#define BIO_ISSUE_TIME_MASK     GENMASK_ULL(BIO_ISSUE_SIZE_SHIFT - 1, 0)

Not sure why we even have these helpers, I'd argue that patches like
this HURT readability, not improve it. When I see

((1ULL << SOME_SHIFT) - 1)

I know precisely what that does, whereas I have to think about the other
one, maybe even look it up to be sure. For instance, without looking
now, I have no idea what the second argument is. Looking at the git log,
I see numerous instances of:

"xxx: Fix misuses of GENMASK macro

 Arguments are supposed to be ordered high then low."

Hence it seems GENMASK_ULL is easy to misuse or get wrong, the very
opposite of what you'd want in a helper. How is it helping readability
if the helper is easy to misuse?

Ditto with

(1ULL << SOME_SHIFT)

vs BIT_ULL. But at least that one doesn't have a mysterious 2nd
argument.

Hence I'm not inclined to apply this patch.

-- 
Jens Axboe

Re: [PATCH 1/1] block: Use bits.h macros to improve readability

[Joe Perches]

On Fri, 2019-08-02 at 21:18 -0700, Jens Axboe wrote:
> On 8/1/19 6:00 PM, Leonardo Bras wrote:
> > Applies some bits.h macros in order to improve readability of
> > linux/blk_types.h.
[]
> I know precisely what that does, whereas I have to think about the other
> one, maybe even look it up to be sure. For instance, without looking
> now, I have no idea what the second argument is. Looking at the git log,
> I see numerous instances of:

While I'm not at all a proponent of GENMASK/GENMASK_ULL,
and so not a proponent of this patch, latent defects are
possible in both cases.

You'd likely have to look at SOME_SHIFT to see if it's 0
to verify the actual mask is what's really desired.

$ git grep -P '_SHIFT\s+\(?\s*0\s*\)?\b' | wc -l
11907

Re: [PATCH 1/1] block: Use bits.h macros to improve readability

[Jens Axboe]

On 8/4/19 1:17 PM, Joe Perches wrote:
> On Fri, 2019-08-02 at 21:18 -0700, Jens Axboe wrote:
>> On 8/1/19 6:00 PM, Leonardo Bras wrote:
>>> Applies some bits.h macros in order to improve readability of
>>> linux/blk_types.h.
> []
>> I know precisely what that does, whereas I have to think about the other
>> one, maybe even look it up to be sure. For instance, without looking
>> now, I have no idea what the second argument is. Looking at the git log,
>> I see numerous instances of:
> 
> While I'm not at all a proponent of GENMASK/GENMASK_ULL,
> and so not a proponent of this patch, latent defects are
> possible in both cases.
> 
> You'd likely have to look at SOME_SHIFT to see if it's 0
> to verify the actual mask is what's really desired.
> 
> $ git grep -P '_SHIFT\s+\(?\s*0\s*\)?\b' | wc -l
> 11907

Usually SOME_SHIFT is either a numeric value, and you already know,
or it's part of a series of enums/defines (like in the case that's
being done here) where you already have prior knowledge about the
values. For those cases, GENMASK/GENMASK_ULL adds nothing imho, it
only makes it less readable and more error prone.

-- 
Jens Axboe

Re: [PATCH 1/1] block: Use bits.h macros to improve readability

[Leonardo Bras]

[-- Attachment #1: Type: text/plain, Size: 2452 bytes --]

On Fri, 2019-08-02 at 21:18 -0700, Jens Axboe wrote:
> On 8/1/19 6:00 PM, Leonardo Bras wrote:
> > Applies some bits.h macros in order to improve readability of
> > linux/blk_types.h.
> > 
> > Signed-off-by: Leonardo Bras <leonardo@linux.ibm.com>
> > ---
> >   include/linux/blk_types.h | 55 ++++++++++++++++++++-------------------
> >   1 file changed, 28 insertions(+), 27 deletions(-)
> > 
> > diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
> > index 95202f80676c..31c8c6d274f6 100644
> > --- a/include/linux/blk_types.h
> > +++ b/include/linux/blk_types.h
> > @@ -9,6 +9,7 @@
> >   #include <linux/types.h>
> >   #include <linux/bvec.h>
> >   #include <linux/ktime.h>
> > +#include <linux/bits.h>
> >   
> >   struct bio_set;
> >   struct bio;
> > @@ -101,13 +102,13 @@ static inline bool blk_path_error(blk_status_t error)
> >   #define BIO_ISSUE_SIZE_BITS     12
> >   #define BIO_ISSUE_RES_SHIFT     (64 - BIO_ISSUE_RES_BITS)
> >   #define BIO_ISSUE_SIZE_SHIFT    (BIO_ISSUE_RES_SHIFT - BIO_ISSUE_SIZE_BITS)
> > -#define BIO_ISSUE_TIME_MASK     ((1ULL << BIO_ISSUE_SIZE_SHIFT) - 1)
> > +#define BIO_ISSUE_TIME_MASK     GENMASK_ULL(BIO_ISSUE_SIZE_SHIFT - 1, 0)
> 
> Not sure why we even have these helpers, I'd argue that patches like
> this HURT readability, not improve it. When I see
> 
> ((1ULL << SOME_SHIFT) - 1)
> 
> I know precisely what that does, whereas I have to think about the other
> one, maybe even look it up to be sure. For instance, without looking
> now, I have no idea what the second argument is. Looking at the git log,
> I see numerous instances of:
> 
> "xxx: Fix misuses of GENMASK macro
> 
>  Arguments are supposed to be ordered high then low."
> 
> Hence it seems GENMASK_ULL is easy to misuse or get wrong, the very
> opposite of what you'd want in a helper. How is it helping readability
> if the helper is easy to misuse?
> 
> Ditto with
> 
> (1ULL << SOME_SHIFT)
> 
> vs BIT_ULL. But at least that one doesn't have a mysterious 2nd
> argument.
> 
> Hence I'm not inclined to apply this patch.
> 

Well, as a newbie, macros like GENMASK(a,b) look way simpler. I to read
than (((1 << a) - 1) << b), or (~((1 << c) - 1)).
I mean, in GENMASK all bits >= a, <=b are set. 

But I agree that once you get used to reading masks created with <<, it
doesn't look that harder.

Well, thanks for this feedback.

Regards, 
Leonardo Brás

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 833 bytes --]