On Fri, Mar 31, 2023 at 04:45:46PM -0500, Andrew Halaney wrote: > Some platforms have dwmac4 implementations that have a different > address space layout than the default, resulting in the need to define > their own DMA/MTL offsets. > > Extend the functions to allow a platform driver to indicate what its > addresses are, overriding the defaults. > > Signed-off-by: Andrew Halaney <ahalaney@xxxxxxxxxx> > --- > > This patch (and the prior patch) are replacements for > https://lore.kernel.org/netdev/20230320204153.21736840@xxxxxxxxxx/ > as was requested. Hopefully I was understanding the intent correctly :) > > I'm pretty sure further refinement will be requested for this one, but > it is the best I could come up with myself! Specifically some of the > naming, dealing with spacing in some older spots of dwmac4, > where the addresses should live in the structure hierarchy, etc are > things I would not be surprised to have to rework if this is still > preferred over the wrapper approach. > > Changes since v2: > * New, replacing old wrapper approach > > drivers/net/ethernet/stmicro/stmmac/dwmac4.h | 91 ++++++++-- > .../net/ethernet/stmicro/stmmac/dwmac4_core.c | 36 ++-- > .../net/ethernet/stmicro/stmmac/dwmac4_dma.c | 157 ++++++++++-------- > .../net/ethernet/stmicro/stmmac/dwmac4_dma.h | 51 +++--- > .../net/ethernet/stmicro/stmmac/dwmac4_lib.c | 67 +++++--- > include/linux/stmmac.h | 19 +++ > 6 files changed, 279 insertions(+), 142 deletions(-) > > diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4.h b/drivers/net/ethernet/stmicro/stmmac/dwmac4.h > index ccd49346d3b3..a0c0ee1dc13f 100644 > --- a/drivers/net/ethernet/stmicro/stmmac/dwmac4.h > +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4.h > @@ -336,14 +336,23 @@ enum power_event { > > #define MTL_CHAN_BASE_ADDR 0x00000d00 > #define MTL_CHAN_BASE_OFFSET 0x40 > -#define MTL_CHANX_BASE_ADDR(x) (MTL_CHAN_BASE_ADDR + \ > - (x * MTL_CHAN_BASE_OFFSET)) > - > -#define MTL_CHAN_TX_OP_MODE(x) MTL_CHANX_BASE_ADDR(x) > -#define MTL_CHAN_TX_DEBUG(x) (MTL_CHANX_BASE_ADDR(x) + 0x8) > -#define MTL_CHAN_INT_CTRL(x) (MTL_CHANX_BASE_ADDR(x) + 0x2c) > -#define MTL_CHAN_RX_OP_MODE(x) (MTL_CHANX_BASE_ADDR(x) + 0x30) > -#define MTL_CHAN_RX_DEBUG(x) (MTL_CHANX_BASE_ADDR(x) + 0x38) > +#define MTL_CHANX_BASE_ADDR(addrs, x) \ > +({ \ > + const struct dwmac4_addrs *__addrs = addrs; \ > + const u32 __x = x; \ > + u32 __addr; \ > + if (__addrs) \ > + __addr = __addrs->mtl_chan + (__x * __addrs->mtl_chan_offset); \ > + else \ > + __addr = MTL_CHAN_BASE_ADDR + (__x * MTL_CHAN_BASE_OFFSET); \ > + __addr; \ > +}) Could this and similar macros added by this patch be functions? >From my pov a benefit would be slightly more type safety. And as a bonus there wouldn't be any need to handle aliasing of input.
