| Message ID | 6435706.peaM34LhnV@wuerfel |
|---|---|
| State | Accepted, archived |
| Delegated to: | David Miller |
| Headers | show |
From: Arnd Bergmann <arnd@arndb.de> Date: Tue, 08 Dec 2015 16:32:59 +0100 > The handling of epib and psdata remains a bit unclear in the driver, > as we access the same fields both as CPU-endian and through DMA > from the device. > > Sparse warns about this: > ti/netcp_core.c:1147:21: warning: incorrect type in assignment (different base types) > ti/netcp_core.c:1147:21: expected unsigned int [usertype] *[assigned] epib > ti/netcp_core.c:1147:21: got restricted __le32 *<noident> > > This uses __le32 types in a few places and uses __force where the code > looks fishy. The previous patch should really have produced the correct > behavior, but this second patch is needed to shut up the warnings about > it. Ideally it would be slightly rewritten to not need those casts, > but I don't dare do that without access to the hardware for proper > testing. > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> Applied to net-next. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/net/ethernet/ti/netcp.h b/drivers/net/ethernet/ti/netcp.h index bb1bb72121c0..17a26a429b71 100644 --- a/drivers/net/ethernet/ti/netcp.h +++ b/drivers/net/ethernet/ti/netcp.h @@ -113,7 +113,7 @@ struct netcp_intf { #define NETCP_PSDATA_LEN KNAV_DMA_NUM_PS_WORDS struct netcp_packet { struct sk_buff *skb; - u32 *epib; + __le32 *epib; u32 *psdata; unsigned int psdata_len; struct netcp_intf *netcp; diff --git a/drivers/net/ethernet/ti/netcp_core.c b/drivers/net/ethernet/ti/netcp_core.c index eb2585e777e1..92d08eb262c2 100644 --- a/drivers/net/ethernet/ti/netcp_core.c +++ b/drivers/net/ethernet/ti/netcp_core.c @@ -1145,8 +1145,8 @@ static int netcp_tx_submit_skb(struct netcp_intf *netcp, p_info.ts_context = NULL; p_info.txtstamp_complete = NULL; p_info.epib = desc->epib; - p_info.psdata = desc->psdata; - memset(p_info.epib, 0, KNAV_DMA_NUM_EPIB_WORDS * sizeof(u32)); + p_info.psdata = (u32 __force *)desc->psdata; + memset(p_info.epib, 0, KNAV_DMA_NUM_EPIB_WORDS * sizeof(__le32)); /* Find out where to inject the packet for transmission */ list_for_each_entry(tx_hook, &netcp->txhook_list_head, list) { @@ -1170,11 +1170,12 @@ static int netcp_tx_submit_skb(struct netcp_intf *netcp, /* update descriptor */ if (p_info.psdata_len) { - u32 *psdata = p_info.psdata; + /* psdata points to both native-endian and device-endian data */ + __le32 *psdata = (void __force *)p_info.psdata; memmove(p_info.psdata, p_info.psdata + p_info.psdata_len, p_info.psdata_len); - set_words(psdata, p_info.psdata_len, psdata); + set_words(p_info.psdata, p_info.psdata_len, psdata); tmp |= (p_info.psdata_len & KNAV_DMA_DESC_PSLEN_MASK) << KNAV_DMA_DESC_PSLEN_SHIFT; }
The handling of epib and psdata remains a bit unclear in the driver, as we access the same fields both as CPU-endian and through DMA from the device. Sparse warns about this: ti/netcp_core.c:1147:21: warning: incorrect type in assignment (different base types) ti/netcp_core.c:1147:21: expected unsigned int [usertype] *[assigned] epib ti/netcp_core.c:1147:21: got restricted __le32 *<noident> This uses __le32 types in a few places and uses __force where the code looks fishy. The previous patch should really have produced the correct behavior, but this second patch is needed to shut up the warnings about it. Ideally it would be slightly rewritten to not need those casts, but I don't dare do that without access to the hardware for proper testing. Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- drivers/net/ethernet/ti/netcp.h | 2 +- drivers/net/ethernet/ti/netcp_core.c | 9 +++++---- 2 files changed, 6 insertions(+), 5 deletions(-)