On 22-10-21, 07:26, Peter Geis wrote:

On Fri, Oct 22, 2021 at 6:51 AM Vinod Koul [off-list ref] wrote:

quoted

On 13-10-21, 18:19, Yifeng Zhao wrote:

quoted

quoted

+#define RK3568_T22_PHYREG6           (0x6 << 2)
+#define T22_PHYREG6_TX_RTERM_MASK    GENMASK(7, 4)
+#define T22_PHYREG6_TX_RTERM_SHIFT   4
+#define T22_PHYREG6_TX_RTERM_50OHM   0x8
+#define T22_PHYREG6_RX_RTERM_MASK    GENMASK(3, 0)
+#define T22_PHYREG6_RX_RTERM_SHIFT   0
+#define T22_PHYREG6_RX_RTERM_44OHM   0xF
+
+#define RK3568_T22_PHYREG7           (0x7 << 2)

Pls use GENMASK for these?

quoted

+#define T22_PHYREG7_SSC_EN           BIT(4)
+
+#define RK3568_T22_PHYREG10          (0xA << 2)
+#define T22_PHYREG10_SU_TRIM_0_7     0xF0
+
+#define RK3568_T22_PHYREG11          (0xB << 2)
+#define T22_PHYREG11_PLL_LPF_ADJ     0x4
+
+#define RK3568_T22_PHYREG12          (0xC << 2)
+#define T22_PHYREG12_RESISTER_MASK   GENMASK(5, 4)
+#define T22_PHYREG12_RESISTER_SHIFT  0x4

bitfield.h has nice helpers which can extract/program values and avoid
one to define these shifts

They aren't values, they are registers.

Yes!

This is a remnant from the downstream driver's attempt at obfuscating
the register it's touching.
Please define these correctly.

The point of bitfield.h is one defines register bit fields using
BIT/GENMASK, no need to define SHIFT etc and use the helpers to
extract/program values and avoid defining these shifts etc!

-- 
~Vinod