| Message ID | 20150319171301.26446.71081.stgit@tlendack-t1.amdoffice.net |
|---|---|
| State | Changes Requested, archived |
| Delegated to: | David Miller |
| Headers | show |
On 03/19/2015 12:32 PM, David Laight wrote: > From: Tom Lendacky >> The hardware supplies a value that indicates the DMA range that it >> is capable of using. Use this value rather than hard-coding it in >> the driver. > ... >> + /* Translate the address width setting into actual number */ >> + switch (hw_feat->dma_width) { >> + case 0: >> + hw_feat->dma_width = 32; >> + break; >> + case 1: >> + hw_feat->dma_width = 40; >> + break; >> + case 2: >> + hw_feat->dma_width = 48; >> + break; >> + } >> + > > I see a glaring problem should the hardware ever return 3. Except this version of the hardware can't return a 3 and any future versions will have a different device tree id or ACPI id and won't match this driver (or the driver will be updated at that point to support the newer versions). I could have it default to 32 if you feel it's really needed, though. Thanks, Tom > > David > -- 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
From: Tom Lendacky <thomas.lendacky@amd.com> Date: Thu, 19 Mar 2015 12:40:05 -0500 > Except this version of the hardware can't return a 3 Yep, I've never _ever_ seen HW return garbage in a register. Nope. No reason to every program defensively for that. -- 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
On 03/19/2015 01:43 PM, David Miller wrote: > From: Tom Lendacky <thomas.lendacky@amd.com> > Date: Thu, 19 Mar 2015 12:40:05 -0500 > >> Except this version of the hardware can't return a 3 > > Yep, I've never _ever_ seen HW return garbage in a register. > > Nope. > > No reason to every program defensively for that. Point taken :-). I'll add a default to 32 bit and submit v2. Thanks, Tom > -- 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/amd/xgbe/xgbe-common.h b/drivers/net/ethernet/amd/xgbe/xgbe-common.h index 29a0927..34c28aa 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-common.h +++ b/drivers/net/ethernet/amd/xgbe/xgbe-common.h @@ -365,6 +365,8 @@ #define MAC_HWF0R_TXCOESEL_WIDTH 1 #define MAC_HWF0R_VLHASH_INDEX 4 #define MAC_HWF0R_VLHASH_WIDTH 1 +#define MAC_HWF1R_ADDR64_INDEX 14 +#define MAC_HWF1R_ADDR64_WIDTH 2 #define MAC_HWF1R_ADVTHWORD_INDEX 13 #define MAC_HWF1R_ADVTHWORD_WIDTH 1 #define MAC_HWF1R_DBGMEMA_INDEX 19 diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-drv.c b/drivers/net/ethernet/amd/xgbe/xgbe-drv.c index 8635c94..abaf092 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-drv.c +++ b/drivers/net/ethernet/amd/xgbe/xgbe-drv.c @@ -519,6 +519,7 @@ void xgbe_get_all_hw_features(struct xgbe_prv_data *pdata) RXFIFOSIZE); hw_feat->tx_fifo_size = XGMAC_GET_BITS(mac_hfr1, MAC_HWF1R, TXFIFOSIZE); + hw_feat->dma_width = XGMAC_GET_BITS(mac_hfr1, MAC_HWF1R, ADDR64); hw_feat->dcb = XGMAC_GET_BITS(mac_hfr1, MAC_HWF1R, DCBEN); hw_feat->sph = XGMAC_GET_BITS(mac_hfr1, MAC_HWF1R, SPHEN); hw_feat->tso = XGMAC_GET_BITS(mac_hfr1, MAC_HWF1R, TSOEN); @@ -553,6 +554,19 @@ void xgbe_get_all_hw_features(struct xgbe_prv_data *pdata) break; } + /* Translate the address width setting into actual number */ + switch (hw_feat->dma_width) { + case 0: + hw_feat->dma_width = 32; + break; + case 1: + hw_feat->dma_width = 40; + break; + case 2: + hw_feat->dma_width = 48; + break; + } + /* The Queue, Channel and TC counts are zero based so increment them * to get the actual number */ diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-main.c b/drivers/net/ethernet/amd/xgbe/xgbe-main.c index 32dd651..2e4c22d 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-main.c +++ b/drivers/net/ethernet/amd/xgbe/xgbe-main.c @@ -374,15 +374,6 @@ static int xgbe_probe(struct platform_device *pdev) pdata->awcache = XGBE_DMA_SYS_AWCACHE; } - /* Set the DMA mask */ - if (!dev->dma_mask) - dev->dma_mask = &dev->coherent_dma_mask; - ret = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(40)); - if (ret) { - dev_err(dev, "dma_set_mask_and_coherent failed\n"); - goto err_io; - } - /* Get the device interrupt */ ret = platform_get_irq(pdev, 0); if (ret < 0) { @@ -409,6 +400,16 @@ static int xgbe_probe(struct platform_device *pdev) /* Set default configuration data */ xgbe_default_config(pdata); + /* Set the DMA mask */ + if (!dev->dma_mask) + dev->dma_mask = &dev->coherent_dma_mask; + ret = dma_set_mask_and_coherent(dev, + DMA_BIT_MASK(pdata->hw_feat.dma_width)); + if (ret) { + dev_err(dev, "dma_set_mask_and_coherent failed\n"); + goto err_io; + } + /* Calculate the number of Tx and Rx rings to be created * -Tx (DMA) Channels map 1-to-1 to Tx Queues so set * the number of Tx queues to the number of Tx channels diff --git a/drivers/net/ethernet/amd/xgbe/xgbe.h b/drivers/net/ethernet/amd/xgbe/xgbe.h index 1eea3e5..bfe11fb 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe.h +++ b/drivers/net/ethernet/amd/xgbe/xgbe.h @@ -632,6 +632,7 @@ struct xgbe_hw_features { unsigned int rx_fifo_size; /* MTL Receive FIFO Size */ unsigned int tx_fifo_size; /* MTL Transmit FIFO Size */ unsigned int adv_ts_hi; /* Advance Timestamping High Word */ + unsigned int dma_width; /* DMA width */ unsigned int dcb; /* DCB Feature */ unsigned int sph; /* Split Header Feature */ unsigned int tso; /* TCP Segmentation Offload */
The hardware supplies a value that indicates the DMA range that it is capable of using. Use this value rather than hard-coding it in the driver. Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com> --- drivers/net/ethernet/amd/xgbe/xgbe-common.h | 2 ++ drivers/net/ethernet/amd/xgbe/xgbe-drv.c | 14 ++++++++++++++ drivers/net/ethernet/amd/xgbe/xgbe-main.c | 19 ++++++++++--------- drivers/net/ethernet/amd/xgbe/xgbe.h | 1 + 4 files changed, 27 insertions(+), 9 deletions(-) -- 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