Message-ID: <6b7de79d-b149-9888-2d0f-2acee5c2905e@gmx.de>
Date:   Mon, 28 Nov 2016 22:41:46 +0100
From:   Lino Sanfilippo <LinoSanfilippo@....de>
To:     Florian Fainelli <f.fainelli@...il.com>, davem@...emloft.net,
        charrer@...critech.com, liodot@...il.com,
        gregkh@...uxfoundation.org, andrew@...n.ch
Cc:     devel@...verdev.osuosl.org, linux-kernel@...r.kernel.org,
        netdev@...r.kernel.org
Subject: Re: [PATCH v3 net-next 1/2] net: ethernet: slicoss: add slicoss
 gigabit ethernet driver

Hi Florian,

On 28.11.2016 05:56, Florian Fainelli wrote:
> On 11/26/2016 04:20 AM, Lino Sanfilippo wrote:
>> Add driver for Alacritech gigabit ethernet cards with SLIC (session-layer
>> interface control) technology. The driver provides basic support without
>> SLIC for the following devices:
>> 
>> - Mojave cards (single port PCI Gigabit) both copper and fiber
>> - Oasis cards (single and dual port PCI-x Gigabit) copper and fiber
>> - Kalahari cards (dual and quad port PCI-e Gigabit) copper and fiber
> 
> This looks great, a few nits below:
> 
> 
>> +#define SLIC_MAX_TX_COMPLETIONS		100
> 
> You usually don't want to limit the number of TX completion, if the
> entire TX ring needs to be cleaned, you would want to allow that.
> 

The problem is that the HW does not provide a tx completion index. Instead we have to 
iterate the status descriptors until we get an invalid idx which indicates that there 
are no further tx descriptors done for now. I am afraid that if we do not limit the 
number of descriptors processed in the tx completion handler, a continuous transmission 
of frames could keep the loop in xmit_complete() run endlessly. I dont know if this 
can actually happen but I wanted to make sure that this is avoided.

> [snip]
> 
>> +	while (slic_get_free_rx_descs(rxq) > SLIC_MAX_REQ_RX_DESCS) {
>> +		skb = alloc_skb(maplen + ALIGN_MASK, gfp);
>> +		if (!skb)
>> +			break;
>> +
>> +		paddr = dma_map_single(&sdev->pdev->dev, skb->data, maplen,
>> +				       DMA_FROM_DEVICE);
>> +		if (dma_mapping_error(&sdev->pdev->dev, paddr)) {
>> +			netdev_err(dev, "mapping rx packet failed\n");
>> +			/* drop skb */
>> +			dev_kfree_skb_any(skb);
>> +			break;
>> +		}
>> +		/* ensure head buffer descriptors are 256 byte aligned */
>> +		offset = 0;
>> +		misalign = paddr & ALIGN_MASK;
>> +		if (misalign) {
>> +			offset = SLIC_RX_BUFF_ALIGN - misalign;
>> +			skb_reserve(skb, offset);
>> +		}
>> +		/* the HW expects dma chunks for descriptor + frame data */
>> +		desc = (struct slic_rx_desc *)skb->data;
>> +		memset(desc, 0, sizeof(*desc));
> 
> Do you really need to zero-out the prepending RX descriptor? Are not you
> missing a write barrier here?

Indeed, it should be sufficient to make sure that the bit SLIC_IRHDDR_SVALID is not set.
I will adjust it. 
Concerning the write barrier: You mean a wmb() before slic_write() to ensure that the zeroing
 of the status desc is done before the descriptor is passed to the HW, right?


> [snip]
> 
>> +
>> +		dma_sync_single_for_cpu(&sdev->pdev->dev,
>> +					dma_unmap_addr(buff, map_addr),
>> +					buff->addr_offset + sizeof(*desc),
>> +					DMA_FROM_DEVICE);
>> +
>> +		status = le32_to_cpu(desc->status);
>> +		if (!(status & SLIC_IRHDDR_SVALID))
>> +			break;
>> +
>> +		buff->skb = NULL;
>> +
>> +		dma_unmap_single(&sdev->pdev->dev,
>> +				 dma_unmap_addr(buff, map_addr),
>> +				 dma_unmap_len(buff, map_len),
>> +				 DMA_FROM_DEVICE);
> 
> This is potentially inefficient, you already did a cache invalidation
> for the RX descriptor here, you could be more efficient with just
> invalidating the packet length, minus the descriptor length.
> 

I am not sure I understand: We have to unmap the complete dma area, no matter if we synced
part of it before, dont we? AFAIK a dma sync is different from unmapping dma, or do I miss
something?


>> +
>> +		/* skip rx descriptor that is placed before the frame data */
>> +		skb_reserve(skb, SLIC_RX_BUFF_HDR_SIZE);
>> +
>> +		if (unlikely(status & SLIC_IRHDDR_ERR)) {
>> +			slic_handle_frame_error(sdev, skb);
>> +			dev_kfree_skb_any(skb);
>> +		} else {
>> +			struct ethhdr *eh = (struct ethhdr *)skb->data;
>> +
>> +			if (is_multicast_ether_addr(eh->h_dest))
>> +				SLIC_INC_STATS_COUNTER(&sdev->stats, rx_mcasts);
>> +
>> +			len = le32_to_cpu(desc->length) & SLIC_IRHDDR_FLEN_MSK;
>> +			skb_put(skb, len);
>> +			skb->protocol = eth_type_trans(skb, dev);
>> +			skb->ip_summed = CHECKSUM_UNNECESSARY;
>> +			skb->dev = dev;
> 
> eth_type_trans() already assigns skb->dev = dev;
> 

Right, this is unnecessary, I will fix it.

>> +static int slic_poll(struct napi_struct *napi, int todo)
>> +{
>> +	struct slic_device *sdev = container_of(napi, struct slic_device, napi);
>> +	struct slic_shmem *sm = &sdev->shmem;
>> +	struct slic_shmem_data *sm_data = sm->shmem_data;
>> +	u32 isr = le32_to_cpu(sm_data->isr);
>> +	unsigned int done = 0;
>> +
>> +	slic_handle_irq(sdev, isr, todo, &done);
>> +
>> +	if (done < todo) {
>> +		napi_complete(napi);
> 
> napi_complete_done() since you know how many packets you completed.
> 

Ok, will change it.

>> +		/* reenable irqs */
>> +		sm_data->isr = 0;
>> +		/* make sure sm_data->isr is cleard before irqs are reenabled */
>> +		wmb();
>> +		slic_write(sdev, SLIC_REG_ISR, 0);
>> +		slic_flush_write(sdev);
>> +	}
>> +
>> +	return done;
>> +}
>> +
>> +static irqreturn_t slic_irq(int irq, void *dev_id)
>> +{
>> +	struct slic_device *sdev = dev_id;
>> +	struct slic_shmem *sm = &sdev->shmem;
>> +	struct slic_shmem_data *sm_data = sm->shmem_data;
>> +
>> +	slic_write(sdev, SLIC_REG_ICR, SLIC_ICR_INT_MASK);
>> +	slic_flush_write(sdev);
>> +	/* make sure sm_data->isr is read after ICR_INT_MASK is set */
>> +	wmb();
>> +
>> +	if (!sm_data->isr) {
>> +		dma_rmb();
>> +		/* spurious interrupt */
>> +		slic_write(sdev, SLIC_REG_ISR, 0);
>> +		slic_flush_write(sdev);
>> +		return IRQ_NONE;
>> +	}
>> +
>> +	napi_schedule(&sdev->napi);
> 
> Likewise napi_schedule_irqoff() can be used here.
> 

Ok, will change it.

Thanks a lot Florian!

Regards,
Lino

Powered by blists - more mailing lists