[Patch net-next 4/4] net: fec: Workaround for imx6sx enet tx hang when enable three queues

[fugang.duan at freescale.com]

From: Shawn Guo <shawn.guo at linaro.org> Sent: Tuesday, September 16, 2014 8:52 AM
>To: Li Frank-B20596
>Cc: Duan Fugang-B38611; davem at davemloft.net; netdev at vger.kernel.org;
>lznuaa at gmail.com; linux-arm-kernel at lists.infradead.org
>Subject: Re: [Patch net-next 4/4] net: fec: Workaround for imx6sx enet tx
>hang when enable three queues
>
>On Tue, Sep 16, 2014 at 01:12:57AM +0800, Frank.Li at freescale.com wrote:
>> From: Fugang Duan <B38611 at freescale.com>
>>
>> When enable three queues on imx6sx enet, and then do tx performance
>> test with iperf tool, after some time running, tx hang.
>>
>> Found that:
>> 	If uDMA is running, software set TDAR may cause tx hang.
>> 	If uDMA is in idle, software set TDAR don't cause tx hang.
>>
>> There is a TDAR race condition for mutliQ when the software sets TDAR
>> and the UDMA clears TDAR simultaneously or in a small window (2-4
>cycles).
>> This will cause the udma_tx and udma_tx_arbiter state machines to hang.
>> The issue exist at i.MX6SX enet IP.
>>
>> So, the Workaround is checking TDAR status four time, if TDAR cleared
>> by hardware and then write TDAR, otherwise don't set TDAR.
>>
>> The patch is only one Workaround for the issue TKT210582.
>>
>> Signed-off-by: Fugang Duan <B38611 at freescale.com>
>> Signed-off-by: Frank Li <Frank.Li at freescale.com>
>> ---
>>  drivers/net/ethernet/freescale/fec_main.c | 18 ++++++++++++++++--
>>  1 file changed, 16 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/freescale/fec_main.c
>> b/drivers/net/ethernet/freescale/fec_main.c
>> index 524ddbe..452e576 100644
>> --- a/drivers/net/ethernet/freescale/fec_main.c
>> +++ b/drivers/net/ethernet/freescale/fec_main.c
>> @@ -111,6 +111,13 @@ static void fec_enet_itr_coal_init(struct
>net_device *ndev);
>>   *   independent rings
>>   */
>>  #define FEC_QUIRK_HAS_AVB		(1 << 8)
>> +/*
>> + * There is a TDAR race condition for mutliQ when the software sets
>> +TDAR
>> + * and the UDMA clears TDAR simultaneously or in a small window (2-4
>cycles).
>> + * This will cause the udma_tx and udma_tx_arbiter state machines to
>hang.
>> + * The issue exist at i.MX6SX enet IP.
>> + */
>> +#define FEC_QUIRK_TKT210582		(1 << 9)
>
>I think TKTxxx number is used by Freescale design team to track issues
>internally, and there should be a corresponding errata number like ERRxxx
>which should be accessible by external people in errata document?
>
>Shawn
The issue ticket is: TDAR race condition for mutliQ - Errata: ERR007885

>
>>
>>  static struct platform_device_id fec_devtype[] = {
>>  	{
>> @@ -139,7 +146,7 @@ static struct platform_device_id fec_devtype[] = {
>>  		.driver_data = FEC_QUIRK_ENET_MAC | FEC_QUIRK_HAS_GBIT |
>>  				FEC_QUIRK_HAS_BUFDESC_EX | FEC_QUIRK_HAS_CSUM |
>>  				FEC_QUIRK_HAS_VLAN | FEC_QUIRK_ERR006358 |
>> -				FEC_QUIRK_HAS_AVB,
>> +				FEC_QUIRK_HAS_AVB | FEC_QUIRK_TKT210582,
>>  	}, {
>>  		/* sentinel */
>>  	}
>> @@ -709,6 +716,8 @@ static int fec_enet_txq_submit_tso(struct
>fec_enet_priv_tx_q *txq,
>>  	struct tso_t tso;
>>  	unsigned int index = 0;
>>  	int ret;
>> +	const struct platform_device_id *id_entry =
>> +				platform_get_device_id(fep->pdev);
>>
>>  	if (tso_count_descs(skb) >= fec_enet_get_free_txdesc_num(fep, txq))
>{
>>  		dev_kfree_skb_any(skb);
>> @@ -770,7 +779,12 @@ static int fec_enet_txq_submit_tso(struct
>fec_enet_priv_tx_q *txq,
>>  	txq->cur_tx = bdp;
>>
>>  	/* Trigger transmission start */
>> -	writel(0, fep->hwp + FEC_X_DES_ACTIVE(queue));
>> +	if (!(id_entry->driver_data & FEC_QUIRK_TKT210582) ||
>> +		!readl(fep->hwp + FEC_X_DES_ACTIVE(queue)) ||
>> +		!readl(fep->hwp + FEC_X_DES_ACTIVE(queue)) ||
>> +		!readl(fep->hwp + FEC_X_DES_ACTIVE(queue)) ||
>> +		!readl(fep->hwp + FEC_X_DES_ACTIVE(queue)))
>> +		writel(0, fep->hwp + FEC_X_DES_ACTIVE(queue));
>>
>>  	return 0;
>>
>> --
>> 1.9.1
>>