Skip to contentSkip to conversation

iwlwifi: mvm: Use div64_s64 instead of do_div in iwl_mvm_debug_range_resp

PATCH Superseded Nathan Chancellor 5 participants 10 messages Feb 19, 2019 – Feb 22, 2019

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 iwlwifi: mvm: Use div64_s64 instead of do_div in iwl_mvm_debug_range_resp

Clang warns:

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c:465:2: warning:
comparison of distinct pointer types ('typeof ((rtt_avg)) *' (aka 'long long *') and 'uint64_t *' (aka 'unsigned long long *')) [-Wcompare-distinct-pointer-types]

        do_div(rtt_avg, 6666);
        ^~~~~~~~~~~~~~~~~~~~~
include/asm-generic/div64.h:222:28: note: expanded from macro 'do_div'
        (void)(((typeof((n)) *)0) == ((uint64_t *)0));  \
               ~~~~~~~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~

1 warning generated.

do_div expects an unsigned dividend. Use div64_s64, which expects a signed dividend.

Fixes 937b10c0de68 ("iwlwifi: mvm: add debug prints for FTM")
---
 drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c +1 -1
index e9822a3ec373..92b22250eb7d 100644
@@ -462,7 +462,7 @@ static void iwl_mvm_debug_range_resp(struct iwl_mvm *mvm, u8 index,
 {
 	s64 rtt_avg = res->ftm.rtt_avg * 100;
 
-	do_div(rtt_avg, 6666);
+	div64_s64(rtt_avg, 6666);
 
 	IWL_DEBUG_INFO(mvm, "entry %d\n", index);
 	IWL_DEBUG_INFO(mvm, "\tstatus: %d\n", res->status);
Signature · 1 line
2.21.0.rc1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 12 lines · Nathan Chancellor

Clang warns:

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c:465:2: warning:
comparison of distinct pointer types ('typeof ((rtt_avg)) *' (aka 'long long *') and 'uint64_t *' (aka 'unsigned long long *')) [-Wcompare-distinct-pointer-types]

        do_div(rtt_avg, 6666);
        ^~~~~~~~~~~~~~~~~~~~~
include/asm-generic/div64.h:222:28: note: expanded from macro 'do_div'
        (void)(((typeof((n)) *)0) == ((uint64_t *)0));  \
               ~~~~~~~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~
Quoted 4 lines

1 warning generated.

do_div expects an unsigned dividend. Use div64_s64, which expects a signed dividend.

Eh, IIRC, signed vs unsigned division has implications for rounding towards zero or not, but I doubt that the round trip time average (RTT avg) should ever be negative. General rule of thumb for C is to keep arithmetic signed (even when working with non zero values), so rather than make the literal (6666) a unsigned long, I agree with your change to keep the division signed as well. Thanks for the fix.

Quoted 24 lines
Fixes 937b10c0de68 ("iwlwifi: mvm: add debug prints for FTM")
---
 drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c +1 -2
index e9822a3ec373..92b22250eb7d 100644
@@ -462,7 +462,7 @@ static void iwl_mvm_debug_range_resp(struct iwl_mvm *mvm, u8 index,
 {
        s64 rtt_avg = res->ftm.rtt_avg * 100;
 
-       do_div(rtt_avg, 6666);
+       div64_s64(rtt_avg, 6666);
 
        IWL_DEBUG_INFO(mvm, "entry %d\n", index);
        IWL_DEBUG_INFO(mvm, "\tstatus: %d\n", res->status);
--

2.21.0.rc1

Signature · 2 lines
Thanks,
~Nick Desaulniers
#2 lore raw
Quoted 29 lines · Nick Desaulniers
Quoted 14 lines · Nathan Chancellor

Clang warns:

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c:465:2:
warning:
comparison of distinct pointer types ('typeof ((rtt_avg)) *' (aka 'long
long *') and 'uint64_t *' (aka 'unsigned long long *')) [-Wcompare-distinct-pointer-types]

        do_div(rtt_avg, 6666);
        ^~~~~~~~~~~~~~~~~~~~~
include/asm-generic/div64.h:222:28: note: expanded from macro
'do_div'
        (void)(((typeof((n)) *)0) == ((uint64_t *)0));  \
               ~~~~~~~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~
Quoted 4 lines

1 warning generated.

do_div expects an unsigned dividend. Use div64_s64, which expects a signed dividend.

Eh, IIRC, signed vs unsigned division has implications for rounding towards zero or not, but I doubt that the round trip time average (RTT
avg) should ever be negative. General rule of thumb for C is to keep arithmetic signed (even when working with non zero values), so rather than make the literal (6666) a unsigned long, I agree with your change
to keep the division signed as well. Thanks for the fix.

Quoted 1 line

Thanks for the patch and for the review. I've applied this to our internal tree and it will be sent upstreaming following our normal upstreaming process.

--
Cheers,
Luca.

#3 lore raw
In reply to Nathan Chancellor

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c static void iwl_mvm_debug_range_resp(struct iwl_mvm *mvm, u8 index, · line 465

Quoted 12 lines · Nathan Chancellor
drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c +1 -1
index e9822a3ec373..92b22250eb7d 100644
@@ -462,7 +462,7 @@ static void iwl_mvm_debug_range_resp(struct iwl_mvm *mvm, u8 index,
 {
        s64 rtt_avg = res->ftm.rtt_avg * 100;
 
-       do_div(rtt_avg, 6666);
+       div64_s64(rtt_avg, 6666);

This is wrong: div64_s64 does not modify its argument like do_div(), but it returns the result instead. You also don't want to divide by a 64-bit value when the second argument is a small constant.

I think the correct way should be

       s64 rtt_avg = div_s64(res->ftm.rtt_avg * 100, 6666);

If you know that the value is positive, using unsigned types and div_u64() would be slightly faster.

Arnd

#4 lore raw
Quoted 19 lines · Arnd Bergmann

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c static void iwl_mvm_debug_range_resp(struct iwl_mvm *mvm, u8 index, · line 465

Quoted 12 lines · Nathan Chancellor
drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c +1 -1
index e9822a3ec373..92b22250eb7d 100644
@@ -462,7 +462,7 @@ static void iwl_mvm_debug_range_resp(struct iwl_mvm *mvm, u8 index,
 {
        s64 rtt_avg = res->ftm.rtt_avg * 100;
 
-       do_div(rtt_avg, 6666);
+       div64_s64(rtt_avg, 6666);

This is wrong: div64_s64 does not modify its argument like do_div(), but it returns the result instead. You also don't want to divide by a 64-bit value when the second argument is a small constant.

Quoted 8 lines

I think the correct way should be

       s64 rtt_avg = div_s64(res->ftm.rtt_avg * 100, 6666);

If you know that the value is positive, using unsigned types and div_u64() would be slightly faster.

Arnd

Thanks for the review and explanation, Arnd.

Luca, could you drop this version so I can resend it?

Nathan

#5 lore raw
Quoted 27 lines · Nathan Chancellor
Quoted 26 lines · Arnd Bergmann

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c static void iwl_mvm_debug_range_resp(struct · line 466

Quoted 13 lines · Nathan Chancellor
(quoted diff) +0 -0

initiator.c b/drivers/net/wireless/intel/iwlwifi/mvm/ftm- initiator.c
index e9822a3ec373..92b22250eb7d 100644

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c +1 -1
@@ -462,7 +462,7 @@ static void iwl_mvm_debug_range_resp(struct
 iwl_mvm *mvm, u8 index,
 {
        s64 rtt_avg = res->ftm.rtt_avg * 100;
 
-       do_div(rtt_avg, 6666);
+       div64_s64(rtt_avg, 6666);

This is wrong: div64_s64 does not modify its argument like do_div(), but
it returns the result instead. You also don't want to divide by a 64-bit
value when the second argument is a small constant.

I think the correct way should be

       s64 rtt_avg = div_s64(res->ftm.rtt_avg * 100, 6666);
Quoted 8 lines
Quoted 4 lines

If you know that the value is positive, using unsigned types and div_u64() would be slightly faster.

Arnd

Thanks for the review and explanation, Arnd.

Luca, could you drop this version so I can resend it?

Sure, please do! I already applied this internally, but I can just fix it with your new patch and that will be squashed before being sent upstream, so it will look like your second patch.

--
Cheers,
Luca.

#6 lore raw

PATCH v2 1/1 iwlwifi: mvm: Use div_s64 instead of do_div in iwl_mvm_debug_range_resp

In reply to Nathan Chancellor

Clang warns:

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c:465:2: warning:
comparison of distinct pointer types ('typeof ((rtt_avg)) *' (aka 'long long *') and 'uint64_t *' (aka 'unsigned long long *')) [-Wcompare-distinct-pointer-types]

        do_div(rtt_avg, 6666);
        ^~~~~~~~~~~~~~~~~~~~~
include/asm-generic/div64.h:222:28: note: expanded from macro 'do_div'
        (void)(((typeof((n)) *)0) == ((uint64_t *)0));  \
               ~~~~~~~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~

1 warning generated.

do_div expects an unsigned dividend. Use div_s64, which expects a signed dividend.

Fixes 937b10c0de68 ("iwlwifi: mvm: add debug prints for FTM")

---

v1 -> v2:

* Fix logic (as the return value of div{,64}_s64 must be used), thanks to Arnd for the review.

 drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c +1 -3
index e9822a3ec373..94132cfd1f56 100644
@@ -460,9 +460,7 @@ static int iwl_mvm_ftm_range_resp_valid(struct iwl_mvm *mvm, u8 request_id,
 static void iwl_mvm_debug_range_resp(struct iwl_mvm *mvm, u8 index,
 				     struct cfg80211_pmsr_result *res)
 {
-	s64 rtt_avg = res->ftm.rtt_avg * 100;
-
-	do_div(rtt_avg, 6666);
+	s64 rtt_avg = div_s64(res->ftm.rtt_avg * 100, 6666);
 
 	IWL_DEBUG_INFO(mvm, "entry %d\n", index);
 	IWL_DEBUG_INFO(mvm, "\tstatus: %d\n", res->status);
Signature · 1 line
2.21.0.rc1
Reviewed-by on v2 1/1
#7 lore raw

Re: [PATCH v2] iwlwifi: mvm: Use div_s64 instead of do_div in iwl_mvm_debug_range_resp

Quoted 21 lines · Nathan Chancellor

Clang warns:

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c:465:2: warning:
comparison of distinct pointer types ('typeof ((rtt_avg)) *' (aka 'long long *') and 'uint64_t *' (aka 'unsigned long long *')) [-Wcompare-distinct-pointer-types]

        do_div(rtt_avg, 6666);
        ^~~~~~~~~~~~~~~~~~~~~
include/asm-generic/div64.h:222:28: note: expanded from macro 'do_div'
        (void)(((typeof((n)) *)0) == ((uint64_t *)0));  \
               ~~~~~~~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~

1 warning generated.

do_div expects an unsigned dividend. Use div_s64, which expects a signed dividend.

Fixes 937b10c0de68 ("iwlwifi: mvm: add debug prints for FTM")
Quoted 6 lines

---

v1 -> v2:

* Fix logic (as the return value of div{,64}_s64 must be used), thanks to Arnd for the review.

oh boy, sorry I missed that in the initial code review, thanks Arnd for the sharp eye!

Side tangent: we see this kind of difference in APIs a lot (modifying the parameter vs returning a new value or making a copy then modifying that) in C++ when a call site isn't passing the explicit address of some variable or an identifier that's clearly a pointer. Ex.

int foo;
bar(foo);

Doesn't tell you whether bar mutates foo or not without looking at the definition of bar, as it could be:
void bar(int x);
or
void bar(int& x);

I miss the convention in Ruby of using ! suffixes on methods to differentiate between such cases. ex:

"hello".capitalize
vs
"hello".capitalize!

both return the same value, but the one with the ! mutates the existing object, while the one without creates a new object. And that's a very standard convention throughout the standard library. Whether or not people follow that convention is always another story.

One thing I'm curious about, is "why does do_div exist?" When should I use do_div vs div_u64 (not div_s64 as is used in this patch)?

Quoted 22 lines
 drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c +1 -4
index e9822a3ec373..94132cfd1f56 100644
@@ -460,9 +460,7 @@ static int iwl_mvm_ftm_range_resp_valid(struct iwl_mvm *mvm, u8 request_id,
 static void iwl_mvm_debug_range_resp(struct iwl_mvm *mvm, u8 index,
                                     struct cfg80211_pmsr_result *res)
 {
-       s64 rtt_avg = res->ftm.rtt_avg * 100;
-
-       do_div(rtt_avg, 6666);
+       s64 rtt_avg = div_s64(res->ftm.rtt_avg * 100, 6666);
 
        IWL_DEBUG_INFO(mvm, "entry %d\n", index);
        IWL_DEBUG_INFO(mvm, "\tstatus: %d\n", res->status);
--

2.21.0.rc1

Signature · 2 lines
Thanks,
~Nick Desaulniers
#8 lore raw

Re: [PATCH v2] iwlwifi: mvm: Use div_s64 instead of do_div in iwl_mvm_debug_range_resp

Quoted 30 lines · Nick Desaulniers
Quoted 28 lines · Nathan Chancellor

Clang warns:

drivers/net/wireless/intel/iwlwifi/mvm/ftm-initiator.c:465:2:
warning:
comparison of distinct pointer types ('typeof ((rtt_avg)) *' (aka 'long
long *') and 'uint64_t *' (aka 'unsigned long long *')) [-Wcompare-distinct-pointer-types]

        do_div(rtt_avg, 6666);
        ^~~~~~~~~~~~~~~~~~~~~
include/asm-generic/div64.h:222:28: note: expanded from macro
'do_div'
        (void)(((typeof((n)) *)0) == ((uint64_t *)0));  \
               ~~~~~~~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~

1 warning generated.

do_div expects an unsigned dividend. Use div_s64, which expects a signed
dividend.

Fixes 937b10c0de68 ("iwlwifi: mvm: add debug prints for FTM")

---

v1 -> v2:

Quoted 7 lines
Quoted 3 lines

* Fix logic (as the return value of div{,64}_s64 must be used), thanks
to Arnd for the review.

oh boy, sorry I missed that in the initial code review, thanks Arnd for the sharp eye!

Thanks, guys, I really didn't pay much attention when I applied the previous versions either.

I have applied this in our internal tree and will send it out instead of the previous one as part of our upstreaming process.

--
Cheers,
Luca.

#9 lore raw

Re: [PATCH v2] iwlwifi: mvm: Use div_s64 instead of do_div in iwl_mvm_debug_range_resp

In reply to Nick Desaulniers
Quoted 4 lines · Nick Desaulniers

On Thu, Feb 21, 2019 at 12:08 AM Nathan Chancellor <[email protected]> wrote:

One thing I'm curious about, is "why does do_div exist?" When should I use do_div vs div_u64 (not div_s64 as is used in this patch)?

I think do_div() is mostly historic, we've had it since the early days when C compilers were not as good with inline functions. The various other versions are regular functions, and I tend to prefer them for new code, but do_div() is widely known and documented, so I have little hope of it going away any time soon.

Arnd