Re: [dpdk-dev] [PATCH v4 3/4] net/mlx5: remove device register remap

[Yongseok Koh <yskoh@mellanox.com>]

> On Apr 10, 2019, at 10:50 AM, Ferruh Yigit <ferruh.yigit@intel.com> wrote:
> 
> On 4/10/2019 12:13 AM, Yongseok Koh wrote:
>> UAR (User Access Region) register does not need to be remapped for primary
>> process but it should be remapped only for secondary process. UAR register
>> table is in the process private structure in rte_eth_devices[],
>> 	(struct mlx5_proc_priv *)rte_eth_devices[port_id].process_private
>> 
>> The actual UAR table follows the data structure and the table is used for
>> both Tx and Rx.
>> 
>> For Tx, BlueFlame in UAR is used to ring the doorbell. MLX5_TX_BFREG(txq)
>> is defined to get a register for the txq. Processes access its own private
>> data to acquire the register from the UAR table.
>> 
>> For Rx, the doorbell in UAR is required in arming CQ event. However, it is
>> a known issue that the register isn't remapped for secondary process.
>> 
>> Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
> 
> <...>
> 
>> @@ -229,13 +229,99 @@ mlx5_tx_queue_release(void *dpdk_txq)
>> 		}
>> }
>> 
>> +/**
>> + * Initialize Tx UAR registers for primary process.
>> + *
>> + * @param txq_ctrl
>> + *   Pointer to Tx queue control structure.
>> + */
>> +static void
>> +txq_uar_init(struct mlx5_txq_ctrl *txq_ctrl)
>> +{
>> +	struct mlx5_priv *priv = txq_ctrl->priv;
>> +	struct mlx5_proc_priv *ppriv = MLX5_PROC_PRIV(PORT_ID(priv));
>> +
>> +	assert(rte_eal_process_type() == RTE_PROC_PRIMARY);
>> +	assert(ppriv);
>> +	ppriv->uar_table[txq_ctrl->txq.idx] = txq_ctrl->bf_reg;
>> +#ifndef RTE_ARCH_64
>> +	struct mlx5_priv *priv = txq_ctrl->priv;
>> +	struct mlx5_txq_data *txq = &txq_ctrl->txq;
>> +	unsigned int lock_idx;
>> +	/* Assign an UAR lock according to UAR page number */
>> +	lock_idx = (txq_ctrl->uar_mmap_offset / page_size) &
>> +		   MLX5_UAR_PAGE_NUM_MASK;
>> +	txq->uar_lock = &priv->uar_lock[lock_idx];
>> +#endif
>> +}
> 
> This won't compile for arch is not 64bits, since 'page_size' in that block is
> not defined.

It is embarrassing that I have committed so many mistakes on this last patchset.
So many contexts in my head... Or, this patches must be haunted. :-(

I always test 32-bit but it looks like a mistake when rebasing it, not sure...
My apologies. I've sent out v5.

For your convenience, here's the diff.

$ git diff yskoh/upstr-remove-uar-remap
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index 904c4f5c03..6224b3be1a 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -56,16 +56,6 @@
 /** Enable extending memsegs when creating a MR. */
 #define MLX4_MR_EXT_MEMSEG_EN_KVARG "mr_ext_memseg_en"

-/* Reserved address space for UAR mapping. */
-#define MLX4_UAR_SIZE (1ULL << (sizeof(uintptr_t) * 4))
-
-/* Offset of reserved UAR address space to hugepage memory. Offset is used here
- * to minimize possibility of address next to hugepage being used by other code
- * in either primary or secondary process, failing to map TX UAR would make TX
- * packets invisible to HW.
- */
-#define MLX4_UAR_OFFSET (2ULL << (sizeof(uintptr_t) * 4))
-
 enum {
        PCI_VENDOR_ID_MELLANOX = 0x15b3,
 };
diff --git a/drivers/net/mlx5/mlx5_defs.h b/drivers/net/mlx5/mlx5_defs.h
index bfe6655800..69b6960e94 100644
--- a/drivers/net/mlx5/mlx5_defs.h
+++ b/drivers/net/mlx5/mlx5_defs.h
@@ -91,16 +91,6 @@
 /* Timeout in seconds to get a valid link status. */
 #define MLX5_LINK_STATUS_TIMEOUT 10

-/* Reserved address space for UAR mapping. */
-#define MLX5_UAR_SIZE (1ULL << (sizeof(uintptr_t) * 4))
-
-/* Offset of reserved UAR address space to hugepage memory. Offset is used here
- * to minimize possibility of address next to hugepage being used by other code
- * in either primary or secondary process, failing to map TX UAR would make TX
- * packets invisible to HW.
- */
-#define MLX5_UAR_OFFSET (1ULL << (sizeof(uintptr_t) * 4))
-
 /* Maximum number of UAR pages used by a port,
  * These are the size and mask for an array of mutexes used to synchronize
  * the access to port's UARs on platforms that do not support 64 bit writes.
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index 5fb1761955..9965b2b771 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -240,18 +240,19 @@ txq_uar_init(struct mlx5_txq_ctrl *txq_ctrl)
 {
        struct mlx5_priv *priv = txq_ctrl->priv;
        struct mlx5_proc_priv *ppriv = MLX5_PROC_PRIV(PORT_ID(priv));
+#ifndef RTE_ARCH_64
+       unsigned int lock_idx;
+       const size_t page_size = sysconf(_SC_PAGESIZE);
+#endif

        assert(rte_eal_process_type() == RTE_PROC_PRIMARY);
        assert(ppriv);
        ppriv->uar_table[txq_ctrl->txq.idx] = txq_ctrl->bf_reg;
 #ifndef RTE_ARCH_64
-       struct mlx5_priv *priv = txq_ctrl->priv;
-       struct mlx5_txq_data *txq = &txq_ctrl->txq;
-       unsigned int lock_idx;
        /* Assign an UAR lock according to UAR page number */
        lock_idx = (txq_ctrl->uar_mmap_offset / page_size) &
                   MLX5_UAR_PAGE_NUM_MASK;
-       txq->uar_lock = &priv->uar_lock[lock_idx];
+       txq_ctrl->txq.uar_lock = &priv->uar_lock[lock_idx];
 #endif
 }