Re: [PATCH v2 18/22] phy: apple: Add Apple Type-C PHY

[Alyssa Anne Rosenzweig <alyssa@rosenzweig.io>]

Looks a lot nicer than the initial version I looked at a few months ago,
nice work Sven :-)

If we're going to define all those mask32/clear32/etc convenience
helpers, there are a couple more we should probably add and use too:

1.
    void cond_set32(void __iomem *reg, bool cond, u32 mask) {
        if (cond) {
            set32(reg, mask);
        } else {
            clear32(reg, mask);
        }
    }

Not sure on the name but this shows up a bunch of places and turns
messy sequences into straight-line code at least.

2.
   #define bit_to_OV(reg, bit) \
        clear32(reg, bit); \
        set32(reg, bit ## _OV); \

Also not sure on the name, but this would make a bunch of sequences
more compact. For example

> clear32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_POST);
> set32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_POST_OV);
> clear32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_POST_2R);
> set32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_POST_2R_OV);
> clear32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_POST_4R);
> set32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_POST_4R_OV);
> clear32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_PRE);
> set32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_PRE_OV);
> clear32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_PRE_2R);
> set32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_PRE_2R_OV);
> clear32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_PRE_4R);
> set32(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_PRE_4R_OV);

turns into

> bit_to_OV(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_POST);
> bit_to_OV(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_POST_2R);
> bit_to_OV(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_POST_4R);
> bit_to_OV(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_PRE);
> bit_to_OV(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_PRE_2R);
> bit_to_OV(tx_shm + LN_AUSPMA_TX_SHM_TXA_IMP_REG3, LN_TXA_MARGIN_PRE_4R);

3. static inline const struct atcphy_mode_configuration
*get_mode_cfg(struct apple_atcphy *atcphy, enum atcphy_mode) {
	if (atcphy->swap_lanes)
		return &atcphy_modes[mode].swapped;
	else
		return &atcphy_modes[mode].normal;
   }

This only shows up two places but both would be improved by its use.

---

With those cleanups (or an explanation why they're silly), ttbomk this
is r-b me, thank you!