summaryrefslogtreecommitdiffstats
path: root/tools/perf/scripts/python/check-perf-trace.py
diff options
context:
space:
mode:
authorAaron Durbin <adurbin@chromium.org>2014-02-07 16:25:51 -0800
committerJohn W. Linville <linville@tuxdriver.com>2014-02-12 15:36:15 -0500
commit189b3299fe46c3d3f7555e1c80e8e8691e71faf1 (patch)
tree7266f6b4f4c21a3c7d2464982b82f3b2ecbbe611 /tools/perf/scripts/python/check-perf-trace.py
parentdbccc92b5d543d1ead727f1416af9e113a3ccc4a (diff)
mwifiex: don't leak DMA command skbuffs
The current mwifiex pcie driver assumed that it would get its cmdrsp_complete() callback called before another command was sent to unmap the command's skbuff. However, that is not true. The mwifiex_check_ps_cond() will send a sleep command to the card without having adapter->curr_cmd set. Within the workqueue's state machine the adapter's state would be set to allow commands (curr_cmd = NULL && cmd_sent = false) after having receieved the response from the sleep command. The card->cmd_buf would then be overridden with the new command but the first command's skbuff was not unmapped. This leaks mapped skbuffs when a bounce buffer is employed. To rectify this unmap the card->cmd_buf when the response is received from the card instead of waiting for the cmdrsp_complete() callback. Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Paul Stewart <pstew@chromium.org> Reviewed-by: Avinash Patil <patila@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Diffstat (limited to 'tools/perf/scripts/python/check-perf-trace.py')
0 files changed, 0 insertions, 0 deletions