Re: [dpdk-dev] [PATCH v8 06/12] eal: Add pci_uio_alloc_resource()

[Tetsuya Mukawa <mukawa@igel.co.jp>]

On 2015/07/07 17:04, David Marchand wrote:
> On Mon, Jul 6, 2015 at 8:24 AM, Tetsuya Mukawa <mukawa@igel.co.jp
> <mailto:mukawa@igel.co.jp>> wrote:
>
>     From: "Tetsuya.Mukawa" <mukawa@igel.co.jp <mailto:mukawa@igel.co.jp>>
>
>     This patch adds a new function called pci_uio_alloc_resource().
>     The function hides how to prepare uio resource in linuxapp and bsdapp.
>     With the function, pci_uio_map_resource() will be more abstracted.
>
>     Signed-off-by: Tetsuya Mukawa <mukawa@igel.co.jp
>     <mailto:mukawa@igel.co.jp>>
>     ---
>      lib/librte_eal/bsdapp/eal/eal_pci.c       | 84
>     +++++++++++++++++---------
>      lib/librte_eal/linuxapp/eal/eal_pci_uio.c | 98
>     +++++++++++++++++++++----------
>      2 files changed, 123 insertions(+), 59 deletions(-)
>
>     diff --git a/lib/librte_eal/bsdapp/eal/eal_pci.c
>     b/lib/librte_eal/bsdapp/eal/eal_pci.c
>     index 92d9886..ce0ca07 100644
>     --- a/lib/librte_eal/bsdapp/eal/eal_pci.c
>     +++ b/lib/librte_eal/bsdapp/eal/eal_pci.c
>     @@ -189,28 +189,17 @@ pci_uio_map_secondary(struct rte_pci_device
>     *dev)
>             return 1;
>      }
>
>     -/* map the PCI resource of a PCI device in virtual memory */
>      static int
>     -pci_uio_map_resource(struct rte_pci_device *dev)
>     +pci_uio_alloc_resource(struct rte_pci_device *dev,
>     +               struct mapped_pci_resource **uio_res)
>      {
>     -       int i, map_idx = 0;
>             char devname[PATH_MAX]; /* contains the /dev/uioX */
>     -       void *mapaddr;
>     -       uint64_t phaddr;
>     -       uint64_t offset;
>     -       uint64_t pagesz;
>     -       struct rte_pci_addr *loc = &dev->addr;
>     -       struct mapped_pci_resource *uio_res = NULL;
>     -       struct mapped_pci_res_list *uio_res_list =
>     -                       RTE_TAILQ_CAST(rte_uio_tailq.head,
>     mapped_pci_res_list);
>     -       struct pci_map *maps;
>     +       struct rte_pci_addr *loc;
>
>     -       dev->intr_handle.fd = -1;
>     -       dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
>     +       if ((dev == NULL) || (uio_res == NULL))
>     +               return -1;
>
>
> This is an "internal" function, dev cannot be null, neither uio_res.

I will fix it.

>  
>
>     -       /* secondary processes - use already recorded details */
>     -       if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>     -               return pci_uio_map_secondary(dev);
>     +       loc = &dev->addr;
>
>             snprintf(devname, sizeof(devname), "/dev/uio@pci:%u:%u:%u",
>                             dev->addr.bus, dev->addr.devid,
>     dev->addr.function);
>     @@ -231,18 +220,57 @@ pci_uio_map_resource(struct rte_pci_device *dev)
>             dev->intr_handle.type = RTE_INTR_HANDLE_UIO;
>
>             /* allocate the mapping details for secondary processes*/
>     -       if ((uio_res = rte_zmalloc("UIO_RES", sizeof (*uio_res),
>     0)) == NULL) {
>     +       *uio_res = rte_zmalloc("UIO_RES", sizeof(**uio_res), 0);
>     +       if (*uio_res == NULL) {
>                     RTE_LOG(ERR, EAL,
>                             "%s(): cannot store uio mmap details\n",
>     __func__);
>                     goto error;
>             }
>
>     -       snprintf(uio_res->path, sizeof(uio_res->path), "%s", devname);
>     -       memcpy(&uio_res->pci_addr, &dev->addr,
>     sizeof(uio_res->pci_addr));
>     +       snprintf((*uio_res)->path, sizeof((*uio_res)->path), "%s",
>     devname);
>     +       memcpy(&(*uio_res)->pci_addr, &dev->addr,
>     sizeof((*uio_res)->pci_addr));
>
>     +       return 0;
>     +
>     +error:
>     +       if (dev->intr_handle.fd) {
>     +               close(dev->intr_handle.fd);
>     +               dev->intr_handle.fd = -1;
>     +               dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
>     +       }
>     +       return -1;
>     +}
>     +
>     +/* map the PCI resource of a PCI device in virtual memory */
>     +static int
>     +pci_uio_map_resource(struct rte_pci_device *dev)
>     +{
>     +       int i, map_idx = 0, ret;
>     +       char *devname;
>     +       void *mapaddr;
>     +       uint64_t phaddr;
>     +       uint64_t offset;
>     +       uint64_t pagesz;
>     +       struct mapped_pci_resource *uio_res = NULL;
>     +       struct mapped_pci_res_list *uio_res_list =
>     +               RTE_TAILQ_CAST(rte_uio_tailq.head,
>     mapped_pci_res_list);
>     +       struct pci_map *maps;
>     +
>     +       dev->intr_handle.fd = -1;
>     +       dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
>     +
>     +       /* secondary processes - use already recorded details */
>     +       if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>     +               return pci_uio_map_secondary(dev);
>     +
>     +       /* allocate uio resource */
>     +       ret = pci_uio_alloc_resource(dev, &uio_res);
>     +       if ((ret != 0) || (uio_res == NULL))
>     +               return ret;
>
>
> If ret != 0, uio_res cannot be NULL.

Also I will fix it.

>
>  
>
>             /* Map all BARs */
>             pagesz = sysconf(_SC_PAGESIZE);
>     +       devname = uio_res->path;
>
>             maps = uio_res->maps;
>             for (i = 0; i != PCI_MAX_RESOURCE; i++) {
>     @@ -298,13 +326,15 @@ pci_uio_map_resource(struct rte_pci_device *dev)
>      error:
>             for (i = 0; i < map_idx; i++)
>                     rte_free(maps[i].path);
>     -       if (uio_res)
>     -               rte_free(uio_res);
>     -       if (dev->intr_handle.fd >= 0) {
>     -               close(dev->intr_handle.fd);
>     -               dev->intr_handle.fd = -1;
>     -               dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
>     -       }
>     +
>     +       /* 'uio_res' has valid value here */
>
>
> Not sure what this comment means, do you mean != NULL ? It does not
> matter.

I thought rte_free() fails when NULL is passed.
Because of above, I wanted to mention here we don't need to check NULL
because uio_res will have non-NULL value.
But, as you mentioned in other patch, we don't need to check it.
Anyway, I will remove the comment.

>  
>
>     +       rte_free(uio_res);
>     +
>     +       /* 'fd' has valid value here */
>     +       close(dev->intr_handle.fd);
>     +       dev->intr_handle.fd = -1;
>     +       dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
>     +
>             return -1;
>      }
>
>
> In the end, if you add this "alloc" function, why not introduce a
> "free" function that does this cleanup ? 
>
>
> Same comments apply to linux modifications.
>

Sure, I will introduce it.

Tetsuya