On 2/14/19 9:26 AM, Alexandre Torgue wrote:
> GMAC IP is little-endian and used on several kind of CPU (big or little
> endian). Main callbacks functions of the stmmac drivers take care about
> it. It was not the case for dwmac4_get_timestamp function.
This is clearly a bugfix, so you should be targeting the 'net' tree and
provide a Fixes: tag so this can be backported to relevant stable kernels.
>
> Signed-off-by: Alexandre Torgue <alexandre.tor...@st.com>
>
> diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c
> b/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c
> index 20299f6..736e296 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c
> +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c
> @@ -241,15 +241,18 @@ static inline void dwmac4_get_timestamp(void *desc, u32
> ats, u64 *ts)
> static int dwmac4_rx_check_timestamp(void *desc)
> {
> struct dma_desc *p = (struct dma_desc *)desc;
> + unsigned int rdes0 = le32_to_cpu(p->des0);
> + unsigned int rdes1 = le32_to_cpu(p->des1);
> + unsigned int rdes3 = le32_to_cpu(p->des3);
> u32 own, ctxt;
> int ret = 1;
>
> - own = p->des3 & RDES3_OWN;
> - ctxt = ((p->des3 & RDES3_CONTEXT_DESCRIPTOR)
> + own = rdes3 & RDES3_OWN;
> + ctxt = ((rdes3 & RDES3_CONTEXT_DESCRIPTOR)
> >> RDES3_CONTEXT_DESCRIPTOR_SHIFT);
>
> if (likely(!own && ctxt)) {
> - if ((p->des0 == 0xffffffff) && (p->des1 == 0xffffffff))
> + if ((rdes0 == 0xffffffff) && (rdes1 == 0xffffffff))
> /* Corrupted value */
> ret = -EINVAL;
> else
>
--
Florian
