[dpdk-dev] [PATCH v2] eal_pci: Fix max_vfs missing for none igb_uio driver

[Qiu, Michael]

On 2/12/2015 1:50 PM, David Marchand wrote:
> On Wed, Feb 11, 2015 at 2:25 PM, Michael Qiu <michael.qiu at intel.com
> <mailto:michael.qiu at intel.com>> wrote:
>
>     max_vfs will only be created by igb_uio driver, for other
>     drivers like vfio or pci_uio_generic, max_vfs will miss.
>
>     But sriov_numvfs is not driver related, just get the vf numbers
>     from that field.
>
>     Signed-off-by: Michael Qiu <michael.qiu at intel.com
>     <mailto:michael.qiu at intel.com>>
>     ---
>     v2 --> v1:
>             backport for kernel version less than 3.8
>
>      lib/librte_eal/linuxapp/eal/eal_pci.c | 10 +++++++++-
>      1 file changed, 9 insertions(+), 1 deletion(-)
>
>     diff --git a/lib/librte_eal/linuxapp/eal/eal_pci.c
>     b/lib/librte_eal/linuxapp/eal/eal_pci.c
>     index b5f5410..fcf5511 100644
>     --- a/lib/librte_eal/linuxapp/eal/eal_pci.c
>     +++ b/lib/librte_eal/linuxapp/eal/eal_pci.c
>     @@ -272,8 +272,16 @@ pci_scan_one(const char *dirname, uint16_t
>     domain, uint8_t bus,
>             dev->max_vfs = 0;
>             snprintf(filename, sizeof(filename), "%s/max_vfs", dirname);
>             if (!access(filename, F_OK) &&
>     -           eal_parse_sysfs_value(filename, &tmp) == 0) {
>     +           eal_parse_sysfs_value(filename, &tmp) == 0)
>                     dev->max_vfs = (uint16_t)tmp;
>     +       else {
>     +               /* for none igb_uio driver, need kernel
>     +                * version greater than 3.8 */
>     +               snprintf(filename, sizeof(filename),
>     +                        "%s/sriov_numvfs", dirname);
>     +               if (!access(filename, F_OK) &&
>     +                   eal_parse_sysfs_value(filename, &tmp) == 0)
>     +                       dev->max_vfs = (uint16_t)tmp;
>             }
>
>             /* get numa node */
>
>
> I can see a different solution : maybe faking a sriov_numvfs parameter
> from igb_uio for kernels that don't have it and doing nothing for
> kernels that already provide sriov_numvfs.

Yes, your solution is much more better, but little complex. My solution
can be a quick fix.

> But your solution is more straightforward (and we don't need to update
> documentation and/or existing scripts afaik).
>
> Acked-by: David Marchand <david.marchand at 6wind.com
> <mailto:david.marchand at 6wind.com>>
>

Thanks David for reviewing this patch and your good comments

Thanks,
Michale
> -- 
> David Marchand