Messages in this thread | | | Date | Sun, 27 Sep 2015 15:19:09 +0800 | From | huangdaode <> | Subject | Re: [PATCH net-next v3] net: Fix Hisilicon Network Subsystem Support Compilation |
| |
On 2015/9/26 12:42, David Miller wrote: > From: huangdaode <huangdaode@hisilicon.com> > Date: Fri, 25 Sep 2015 16:29:50 +0800 > >> @@ -191,9 +191,12 @@ static void hns_rcb_ring_init(struct ring_pair_cb *ring_pair, int ring_type) >> if (ring_type == RX_RING) { >> dsaf_write_dev(q, RCB_RING_RX_RING_BASEADDR_L_REG, >> (u32)dma); >> +#ifdef CONFIG_64BIT >> dsaf_write_dev(q, RCB_RING_RX_RING_BASEADDR_H_REG, >> (u32)(dma >> 32)); >> - >> +#else >> + dsaf_write_dev(q, RCB_RING_RX_RING_BASEADDR_H_REG, 0); >> +#endif >> dsaf_write_dev(q, RCB_RING_RX_RING_BD_LEN_REG, >> bd_size_type); >> dsaf_write_dev(q, RCB_RING_RX_RING_BD_NUM_REG, > CONFIG_64BIT doesn't tell you if _DMA_ addresses are 64-bit or not. > > Furthermore there is a portable way to shift a 32-bit value down > 32-bits whilst avoiding warnings. > > (x >> 31) >> 1 > > So you should definitely use that instead of the CONFIG_64BIT test. > > .
agreed.
| |