The patch b1c17215d718: "stmmac: add ipq806x glue layer", leads to the
following static checker warning:
.../stmmac/dwmac-ipq806x.c:314 ipq806x_gmac_probe()
warn: double left shift '1 << (1 << gmac->id)'
The NSS_COMMON_CLK_SRC_CTRL_OFFSET macro is used once as an offset, and
once as a mask, which is a bug indeed. We'll fix it by defining the
offset as the real offset value and computing the mask from it when
required.
Tested on IPQ806x ref designs AP148 & DB149.
Reported-by: Dan Carpenter <[email protected]>
Signed-off-by: Mathieu Olivari <[email protected]>
---
drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c
index 333489f..9d89bdb 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c
@@ -42,7 +42,7 @@
#define NSS_COMMON_CLK_DIV_MASK 0x7f
#define NSS_COMMON_CLK_SRC_CTRL 0x14
-#define NSS_COMMON_CLK_SRC_CTRL_OFFSET(x) (1 << x)
+#define NSS_COMMON_CLK_SRC_CTRL_OFFSET(x) (x)
/* Mode is coded on 1 bit but is different depending on the MAC ID:
* MAC0: QSGMII=0 RGMII=1
* MAC1: QSGMII=0 SGMII=0 RGMII=1
@@ -308,7 +308,7 @@ static int ipq806x_gmac_probe(struct platform_device *pdev)
/* Configure the clock src according to the mode */
regmap_read(gmac->nss_common, NSS_COMMON_CLK_SRC_CTRL, &val);
- val &= ~NSS_COMMON_CLK_SRC_CTRL_OFFSET(gmac->id);
+ val &= ~(1 << NSS_COMMON_CLK_SRC_CTRL_OFFSET(gmac->id));
switch (gmac->phy_mode) {
case PHY_INTERFACE_MODE_RGMII:
val |= NSS_COMMON_CLK_SRC_CTRL_RGMII(gmac->id) <<
--
2.1.4
From: Mathieu Olivari <[email protected]>
Date: Tue, 4 Aug 2015 17:25:02 -0700
> The patch b1c17215d718: "stmmac: add ipq806x glue layer", leads to the
> following static checker warning:
>
> .../stmmac/dwmac-ipq806x.c:314 ipq806x_gmac_probe()
> warn: double left shift '1 << (1 << gmac->id)'
>
> The NSS_COMMON_CLK_SRC_CTRL_OFFSET macro is used once as an offset, and
> once as a mask, which is a bug indeed. We'll fix it by defining the
> offset as the real offset value and computing the mask from it when
> required.
>
> Tested on IPQ806x ref designs AP148 & DB149.
>
> Reported-by: Dan Carpenter <[email protected]>
> Signed-off-by: Mathieu Olivari <[email protected]>
Applied.