[dpdk-dev] [Bug 228] rte_mp_request_sync memleak with multiple recipients

[dpdk-dev] [Bug 228] rte_mp_request_sync memleak with multiple recipients

[bugzilla]

https://bugs.dpdk.org/show_bug.cgi?id=228

            Bug ID: 228
           Summary: rte_mp_request_sync memleak with multiple recipients
           Product: DPDK
           Version: unspecified
          Hardware: All
                OS: All
            Status: CONFIRMED
          Severity: minor
          Priority: Normal
         Component: core
          Assignee: dev@dpdk.org
          Reporter: dariusz.stojaczyk@intel.com
  Target Milestone: ---

When sending multiple requests, rte_mp_request_sync
can succeed sending a few of those requests, but then
fail on a later one and in the end return with rc=-1.
The upper layers - e.g. device hotplug - currently
handles this case as if no messages were sent and no
memory for response buffers was allocated, which is
not true.

Fixing the above is not so obvious, as rte_mp_request_sync
can fail before the response buffer pointer is even
set (it is unitialized by default). So the caller
cannot safely access it (or free it) just if the
function returns with rc=-1. One way to fix this is
to always initialize the response buffer pointer and
always require the caller to free the response buffer.
This, however, asks for a redesign and should be
addressed with slightly more effort put into it.
Maybe the response buffer pointer should be initialized
by the caller from the very beginning?

-- 
You are receiving this mail because:
You are the assignee for the bug.

[dpdk-dev] [Bug 228] rte_mp_request_sync memleak with multiple recipients

[bugzilla]

https://bugs.dpdk.org/show_bug.cgi?id=228

            Bug ID: 228
           Summary: rte_mp_request_sync memleak with multiple recipients
           Product: DPDK
           Version: unspecified
          Hardware: All
                OS: All
            Status: CONFIRMED
          Severity: minor
          Priority: Normal
         Component: core
          Assignee: dev@dpdk.org
          Reporter: dariusz.stojaczyk@intel.com
  Target Milestone: ---

When sending multiple requests, rte_mp_request_sync
can succeed sending a few of those requests, but then
fail on a later one and in the end return with rc=-1.
The upper layers - e.g. device hotplug - currently
handles this case as if no messages were sent and no
memory for response buffers was allocated, which is
not true.

Fixing the above is not so obvious, as rte_mp_request_sync
can fail before the response buffer pointer is even
set (it is unitialized by default). So the caller
cannot safely access it (or free it) just if the
function returns with rc=-1. One way to fix this is
to always initialize the response buffer pointer and
always require the caller to free the response buffer.
This, however, asks for a redesign and should be
addressed with slightly more effort put into it.
Maybe the response buffer pointer should be initialized
by the caller from the very beginning?

-- 
You are receiving this mail because:
You are the assignee for the bug.