[dpdk-dev] [PATCH v2 00/61] net/qede/base: qede PMD enhancements

Ferruh Yigit ferruh.yigit at intel.com
Mon Mar 20 17:59:08 CET 2017


On 3/18/2017 7:05 AM, Rasesh Mody wrote:
> Hi,
> 
> This patch set adds support for new firmware 8.18.9.0, new features and
> bug fixes.
> 
> Please apply to dpdk-net-next for 17.05 release. Note that this patch set
> depends on http://dpdk.org/dev/patchwork/patch/21896.
> 
> v1..v2
>  - address all the review comments received so far
> 
> Thanks!
> Rasesh
> 
> Harish Patil (3):
>   net/qede/base: add support for arfs mode
>   net/qede: add ntuple and flow director filter support
>   net/qede: add LRO/TSO offloads support
> 
> Rasesh Mody (58):
>   net/qede/base: return an initialized return value
>   net/qede/base: send FW version driver state to MFW
>   net/qede/base: mask Rx buffer attention bits
>   net/qede/base: print various indication on Tx-timeouts
>   net/qede/base: utilize FW 8.18.9.0
>   net/qede: upgrade the FW to 8.18.9.0
>   net/qede/base: decrease maximum HW func per device
>   net/qede/base: move mask constants defining NIC type
>   net/qede/base: remove attribute from update current config
>   net/qede/base: add nvram options
>   net/qede/base: add comment
>   net/qede/base: use default MTU from shared memory
>   net/qede/base: change queue/sb-id from 8 bit to 16 bit
>   net/qede/base: update MFW when default MTU is changed
>   net/qede/base: prevent device init failure
>   net/qede/base: read card personality via MFW commands
>   net/qede/base: allow probe to succeed with minor HW-issues
>   net/qede/base: remove unneeded step in HW init
>   net/qede/base: allow only trusted VFs to be promisc
>   net/qede/base: qm initialization revamp
>   net/qede/base: print firmware MFW and MBI versions
>   net/qede/base: check active VF queues before stopping
>   net/qede/base: set driver type before sending load request
>   net/qede/base: prevent driver laod with invalid resources
>   net/qede/base: add interfaces for MFW TLV request processing
>   net/qede/base: code refactoring of SP queues
>   net/qede/base: make L2 queues handle based
>   net/qede/base: add support for handling TLV request from MFW
>   net/qede/base: optimize cache-line access
>   net/qede/base: infrastructure changes for VF tunnelling
>   net/qede/base: revise tunnel APIs/structs
>   net/qede/base: add tunnelling support for VFs
>   net/qede/base: formatting changes
>   net/qede/base: prevent transmitter stuck condition
>   net/qede/base: add mask/shift defines for resource command
>   net/qede/base: add API for using MFW resource lock
>   net/qede/base: remove clock slowdown option
>   net/qede/base: add new image types
>   net/qede/base: use L2-handles for RSS configuration
>   net/qede/base: change valloc to vzalloc
>   net/qede/base: add support for previous driver unload
>   net/qede/base: add non-L2 dcbx tlv application support
>   net/qede/base: update bulletin board during VF init
>   net/qede/base: add coalescing support for VFs
>   net/qede/base: add macro got resource value message
>   net/qede/base: add mailbox for resource allocation
>   net/qede/base: add macro for unsupported command
>   net/qede/base: set max values for soft resoruces
>   net/qede/base: add return code check
>   net/qede/base: zero out MFW mailbox data
>   net/qede/base: move code bits
>   net/qede/base: add PF parameter
>   net/qede/base: allow PMD to control vport and RSS engine ids
>   net/qede/base: add udp ports in bulletin board message
>   net/qede/base: prevent DMAE transactions during recovery
>   net/qede/base: multi-Txq support on same queue-zone for VFs
>   net/qede/base: prevent race condition during unload
>   net/qede/base: semantic changes
> 

Hi Rasesh,

Getting following build errors, one with clang [1] and other with 32bit
[2], I have not investigated which patch cause the error, just
copy-pasting the build errors.

These looks like same build errors with previous version of the patchset.


[1]
.../drivers/net/qede/qede_rxtx.c:1202:21: error: variable 'pad' is
uninitialized when used here [-Werror,-Wuninitialized]
                rx_mb->data_off = pad + RTE_PKTMBUF_HEADROOM;
                                  ^~~
.../drivers/net/qede/qede_rxtx.c:997:14: note: initialize the variable
'pad' to silence this warning
        uint16_t pad;
                    ^
                     = 0


[2]
.../drivers/net/qede/qede_fdir.c: In function ‘qede_config_cmn_fdir_filter’:
.../drivers/net/qede/qede_fdir.c:126:44: error: format ‘%lx’ expects
argument of type ‘long unsigned int’, but argument 4 has type ‘uint64_t
{aka long long unsigned int}’ [-Werror=format=]
  snprintf(mz_name, sizeof(mz_name) - 1, "%lx", rte_get_timer_cycles());
                                            ^



More information about the dev mailing list