[PATCH 0/3] common/sfc_efx/base: fix compiler warnings in X4 netport API

[PATCH 0/3] common/sfc_efx/base: fix compiler warnings in X4 netport API

[Ivan Malov]

The 25.07 release has got support for AMD Solarflare X45xx adapters [1].
This series is a follow-up to [1] to fix compiler warnings in the base
driver that are seen in Windows driver build, also based on this driver.

[1] https://patches.dpdk.org/project/dpdk/list/?series=35095

Ivan Malov (3):
  common/sfc_efx/base: fix compiler warnings in statistic code
  common/sfc_efx/base: fix compiler warnings in PHY link setup
  common/sfc_efx/base: fix compiler warning in PHY link getter

 drivers/common/sfc_efx/base/efx_np.c       | 7 ++-----
 drivers/common/sfc_efx/base/medford4_phy.c | 1 -
 2 files changed, 2 insertions(+), 6 deletions(-)

-- 
2.39.5

[PATCH 1/3] common/sfc_efx/base: fix compiler warnings in statistic code

[Ivan Malov]

Fixes: f2f77453cb9f ("common/sfc_efx/base: fill in software LUT for MAC statistics")

Suggested-by: Andy Moreton <andy.moreton@amd.com>
Signed-off-by: Ivan Malov <ivan.malov@arknetworks.am>
Reviewed-by: Andy Moreton <andy.moreton@amd.com>
---
 drivers/common/sfc_efx/base/efx_np.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/common/sfc_efx/base/efx_np.c b/drivers/common/sfc_efx/base/efx_np.c
index a19c986691..798fcce26c 100644
--- a/drivers/common/sfc_efx/base/efx_np.c
+++ b/drivers/common/sfc_efx/base/efx_np.c
@@ -750,7 +750,7 @@ efx_np_stat_describe(
 	return;
 
 found:
-	if (sw_id >= lut_nentries) {
+	if ((unsigned int)sw_id >= lut_nentries) {
 		/*
 		 * Static mapping size and the size of lookup
 		 * table are out-of-sync. Should never happen.
@@ -782,7 +782,6 @@ efx_np_stats_describe(
 	EFX_MCDI_DECLARE_BUF(payload,
 	    MC_CMD_MAC_STATISTICS_DESCRIPTOR_IN_LEN,
 	    MC_CMD_MAC_STATISTICS_DESCRIPTOR_OUT_LENMAX_MCDI2);
-	efx_port_t *epp = &(enp->en_port);
 	uint32_t nprocessed;
 	efx_mcdi_req_t req;
 	uint8_t *entries;
-- 
2.39.5

[PATCH 2/3] common/sfc_efx/base: fix compiler warnings in PHY link setup

[Ivan Malov]

Fixes: 8e79cd30230d ("common/sfc_efx/base: implement PHY link control for Medford4")

Suggested-by: Andy Moreton <andy.moreton@amd.com>
Signed-off-by: Ivan Malov <ivan.malov@arknetworks.am>
Reviewed-by: Andy Moreton <andy.moreton@amd.com>
---
 drivers/common/sfc_efx/base/efx_np.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/common/sfc_efx/base/efx_np.c b/drivers/common/sfc_efx/base/efx_np.c
index 798fcce26c..222d81c12e 100644
--- a/drivers/common/sfc_efx/base/efx_np.c
+++ b/drivers/common/sfc_efx/base/efx_np.c
@@ -1185,7 +1185,6 @@ efx_np_cap_sw_mask_to_hw_enum(
 	__out_opt			uint16_t *enum_hwp)
 {
 	unsigned int sw_nflags_req = 0;
-	unsigned int sw_nflags_sup = 0;
 	uint32_t sw_check_mask = 0;
 	unsigned int i;
 
@@ -1288,7 +1287,6 @@ efx_np_link_ctrl(
 	uint16_t cap_enum_hw;
 	boolean_t supported;
 	efx_mcdi_req_t req;
-	boolean_t phy_an;
 	efx_rc_t rc;
 	uint8_t fec;
 
@@ -1385,7 +1383,7 @@ efx_np_link_ctrl(
 	}
 
 	EFSYS_ASSERT(cap_enum_hw <= UINT8_MAX);
-	fec = cap_enum_hw;
+	fec = (uint8_t)cap_enum_hw;
 
 	MCDI_IN_SET_WORD(req, LINK_CTRL_IN_LINK_TECHNOLOGY, link_tech);
 	MCDI_IN_SET_DWORD(req, LINK_CTRL_IN_CONTROL_FLAGS, flags);
-- 
2.39.5

[PATCH 3/3] common/sfc_efx/base: fix compiler warning in PHY link getter

[Ivan Malov]

Fixes: 2a5cf77e6de8 ("common/sfc_efx/base: provide PHY link get method on Medford4")

Suggested-by: Andy Moreton <andy.moreton@amd.com>
Signed-off-by: Ivan Malov <ivan.malov@arknetworks.am>
Reviewed-by: Andy Moreton <andy.moreton@amd.com>
---
 drivers/common/sfc_efx/base/medford4_phy.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/common/sfc_efx/base/medford4_phy.c b/drivers/common/sfc_efx/base/medford4_phy.c
index 9ba6dfbc10..16e6dd2d84 100644
--- a/drivers/common/sfc_efx/base/medford4_phy.c
+++ b/drivers/common/sfc_efx/base/medford4_phy.c
@@ -34,7 +34,6 @@ medford4_phy_get_link(
 	efx_np_handle_t nph = enp->en_port.ep_np_handle;
 	efx_np_link_state_t ls;
 	efx_np_mac_state_t ms;
-	uint32_t fcntl;
 	efx_rc_t rc;
 
 	rc = efx_np_link_state(enp, nph, &ls);
-- 
2.39.5

Re: [PATCH 0/3] common/sfc_efx/base: fix compiler warnings in X4 netport API

[Stephen Hemminger]

On Fri, 11 Jul 2025 18:43:46 +0400
Ivan Malov <ivan.malov@arknetworks.am> wrote:

> The 25.07 release has got support for AMD Solarflare X45xx adapters [1].
> This series is a follow-up to [1] to fix compiler warnings in the base
> driver that are seen in Windows driver build, also based on this driver.
> 
> [1] https://patches.dpdk.org/project/dpdk/list/?series=35095
> 
> Ivan Malov (3):
>   common/sfc_efx/base: fix compiler warnings in statistic code
>   common/sfc_efx/base: fix compiler warnings in PHY link setup
>   common/sfc_efx/base: fix compiler warning in PHY link getter
> 
>  drivers/common/sfc_efx/base/efx_np.c       | 7 ++-----
>  drivers/common/sfc_efx/base/medford4_phy.c | 1 -
>  2 files changed, 2 insertions(+), 6 deletions(-)
> 

Have your tried enabling -Wshadow and -Wduplicated-branches and -Wduplicated-cond
Those flags are finding stuff in other places.

Re: [PATCH 0/3] common/sfc_efx/base: fix compiler warnings in X4 netport API

[Ivan Malov]

[-- Attachment #1: Type: text/plain, Size: 2326 bytes --]

On Fri, 11 Jul 2025, Stephen Hemminger wrote:

> On Fri, 11 Jul 2025 18:43:46 +0400
> Ivan Malov <ivan.malov@arknetworks.am> wrote:
>
>> The 25.07 release has got support for AMD Solarflare X45xx adapters [1].
>> This series is a follow-up to [1] to fix compiler warnings in the base
>> driver that are seen in Windows driver build, also based on this driver.
>>
>> [1] https://patches.dpdk.org/project/dpdk/list/?series=35095
>>
>> Ivan Malov (3):
>>   common/sfc_efx/base: fix compiler warnings in statistic code
>>   common/sfc_efx/base: fix compiler warnings in PHY link setup
>>   common/sfc_efx/base: fix compiler warning in PHY link getter
>>
>>  drivers/common/sfc_efx/base/efx_np.c       | 7 ++-----
>>  drivers/common/sfc_efx/base/medford4_phy.c | 1 -
>>  2 files changed, 2 insertions(+), 6 deletions(-)
>>
>
> Have your tried enabling -Wshadow and -Wduplicated-branches and -Wduplicated-cond
> Those flags are finding stuff in other places.
>
Thanks Stephen. That is very useful. I tried adding these flags via meson.build
for common/sfc_efx, common/sfc_efx/base and for net/sfc trees and the only thing
I was able to find is as follows:

In file included from ../lib/eal/x86/include/rte_byteorder.h:9,
                  from ../drivers/net/sfc/sfc_ef10_essb_rx.c:14:
../drivers/net/sfc/sfc_ef10_essb_rx.c: In function ‘sfc_ef10_essb_rx_qsize_up_rings’:
../drivers/net/sfc/sfc_dp.h:27:27: warning: declaration of ‘_a’ shadows a previous local [-Wshadow]
    27 |                 typeof(a) _a = (a);     \
       |                           ^~
../lib/eal/include/rte_common.h:817:34: note: in definition of macro ‘RTE_MAX’
   817 |                 typeof (a) _a = (a); \
       |                                  ^
../drivers/net/sfc/sfc_ef10_essb_rx.c:535:33: note: in expansion of macro ‘SFC_DIV_ROUND_UP’
   535 |         nb_hw_rx_desc = RTE_MAX(SFC_DIV_ROUND_UP(nb_rx_desc,
       |                                 ^~~~~~~~~~~~~~~~
../lib/eal/include/rte_common.h:817:28: note: shadowed declaration is here
   817 |                 typeof (a) _a = (a); \
       |                            ^~
../drivers/net/sfc/sfc_ef10_essb_rx.c:535:25: note: in expansion of macro ‘RTE_MAX’
   535 |         nb_hw_rx_desc = RTE_MAX(SFC_DIV_ROUND_UP(nb_rx_desc,
       |

But no extra warnings in the base driver.

Thank you.

Re: [PATCH 0/3] common/sfc_efx/base: fix compiler warnings in X4 netport API

[Stephen Hemminger]

On Fri, 11 Jul 2025 19:13:01 +0400 (+04)
Ivan Malov <ivan.malov@arknetworks.am> wrote:

> On Fri, 11 Jul 2025, Stephen Hemminger wrote:
> 
> > On Fri, 11 Jul 2025 18:43:46 +0400
> > Ivan Malov <ivan.malov@arknetworks.am> wrote:
> >  
> >> The 25.07 release has got support for AMD Solarflare X45xx adapters [1].
> >> This series is a follow-up to [1] to fix compiler warnings in the base
> >> driver that are seen in Windows driver build, also based on this driver.
> >>
> >> [1] https://patches.dpdk.org/project/dpdk/list/?series=35095
> >>
> >> Ivan Malov (3):
> >>   common/sfc_efx/base: fix compiler warnings in statistic code
> >>   common/sfc_efx/base: fix compiler warnings in PHY link setup
> >>   common/sfc_efx/base: fix compiler warning in PHY link getter
> >>
> >>  drivers/common/sfc_efx/base/efx_np.c       | 7 ++-----
> >>  drivers/common/sfc_efx/base/medford4_phy.c | 1 -
> >>  2 files changed, 2 insertions(+), 6 deletions(-)
> >>  
> >
> > Have your tried enabling -Wshadow and -Wduplicated-branches and -Wduplicated-cond
> > Those flags are finding stuff in other places.
> >  
> Thanks Stephen. That is very useful. I tried adding these flags via meson.build
> for common/sfc_efx, common/sfc_efx/base and for net/sfc trees and the only thing
> I was able to find is as follows:
> 
> In file included from ../lib/eal/x86/include/rte_byteorder.h:9,
>                   from ../drivers/net/sfc/sfc_ef10_essb_rx.c:14:
> ../drivers/net/sfc/sfc_ef10_essb_rx.c: In function ‘sfc_ef10_essb_rx_qsize_up_rings’:
> ../drivers/net/sfc/sfc_dp.h:27:27: warning: declaration of ‘_a’ shadows a previous local [-Wshadow]
>     27 |                 typeof(a) _a = (a);     \
>        |                           ^~
> ../lib/eal/include/rte_common.h:817:34: note: in definition of macro ‘RTE_MAX’
>    817 |                 typeof (a) _a = (a); \
>        |                                  ^
> ../drivers/net/sfc/sfc_ef10_essb_rx.c:535:33: note: in expansion of macro ‘SFC_DIV_ROUND_UP’
>    535 |         nb_hw_rx_desc = RTE_MAX(SFC_DIV_ROUND_UP(nb_rx_desc,
>        |                                 ^~~~~~~~~~~~~~~~
> ../lib/eal/include/rte_common.h:817:28: note: shadowed declaration is here
>    817 |                 typeof (a) _a = (a); \
>        |                            ^~
> ../drivers/net/sfc/sfc_ef10_essb_rx.c:535:25: note: in expansion of macro ‘RTE_MAX’
>    535 |         nb_hw_rx_desc = RTE_MAX(SFC_DIV_ROUND_UP(nb_rx_desc,
>        |
> 
> But no extra warnings in the base driver.
> 
> Thank you.

Good thanks, RTE_MIN/RTE_MAX has issues when nested. Not sure if there is an easy fix.

Re: [PATCH 0/3] common/sfc_efx/base: fix compiler warnings in X4 netport API

[Stephen Hemminger]

On Fri, 11 Jul 2025 18:43:46 +0400
Ivan Malov <ivan.malov@arknetworks.am> wrote:

> The 25.07 release has got support for AMD Solarflare X45xx adapters [1].
> This series is a follow-up to [1] to fix compiler warnings in the base
> driver that are seen in Windows driver build, also based on this driver.
> 
> [1] https://patches.dpdk.org/project/dpdk/list/?series=35095
> 
> Ivan Malov (3):
>   common/sfc_efx/base: fix compiler warnings in statistic code
>   common/sfc_efx/base: fix compiler warnings in PHY link setup
>   common/sfc_efx/base: fix compiler warning in PHY link getter
> 
>  drivers/common/sfc_efx/base/efx_np.c       | 7 ++-----
>  drivers/common/sfc_efx/base/medford4_phy.c | 1 -
>  2 files changed, 2 insertions(+), 6 deletions(-)
> 
> -- 
Queued to next-net for 25.11

Re: [PATCH 1/3] common/sfc_efx/base: fix compiler warnings in statistic code

[Andrew Rybchenko]

On 7/11/25 17:43, Ivan Malov wrote:
> Fixes: f2f77453cb9f ("common/sfc_efx/base: fill in software LUT for MAC statistics")
> 
> Suggested-by: Andy Moreton <andy.moreton@amd.com>
> Signed-off-by: Ivan Malov <ivan.malov@arknetworks.am>
> Reviewed-by: Andy Moreton <andy.moreton@amd.com>
> ---
>   drivers/common/sfc_efx/base/efx_np.c | 3 +--
>   1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/common/sfc_efx/base/efx_np.c b/drivers/common/sfc_efx/base/efx_np.c
> index a19c986691..798fcce26c 100644
> --- a/drivers/common/sfc_efx/base/efx_np.c
> +++ b/drivers/common/sfc_efx/base/efx_np.c
> @@ -750,7 +750,7 @@ efx_np_stat_describe(
>   	return;
>   
>   found:
> -	if (sw_id >= lut_nentries) {
> +	if ((unsigned int)sw_id >= lut_nentries) {

Unfortunately it is not trivial to understand what's happening here
without looking at code. I bit more verbose description would be
helpful.

>   		/*
>   		 * Static mapping size and the size of lookup
>   		 * table are out-of-sync. Should never happen.
> @@ -782,7 +782,6 @@ efx_np_stats_describe(
>   	EFX_MCDI_DECLARE_BUF(payload,
>   	    MC_CMD_MAC_STATISTICS_DESCRIPTOR_IN_LEN,
>   	    MC_CMD_MAC_STATISTICS_DESCRIPTOR_OUT_LENMAX_MCDI2);
> -	efx_port_t *epp = &(enp->en_port);

I guess the variable is simply unused.

>   	uint32_t nprocessed;
>   	efx_mcdi_req_t req;
>   	uint8_t *entries;

IMHO it would be better to group similar fixes as unused variables into
single patch which clearly explains what's happening.
The goal is to "fix unused variable warnings".
The way is to "Remove unused variables."