summaryrefslogtreecommitdiffstats
path: root/net/ipv4/tcp_ulp.c
diff options
context:
space:
mode:
authorMichal Kubecek <mkubecek@suse.cz>2020-07-09 12:11:50 +0200
committerDavid S. Miller <davem@davemloft.net>2020-07-09 21:35:33 +0200
commit365f9ae4ee36037e2a9268fe7296065356840b4c (patch)
tree6c88d2bceeb96f8bd8a7ab14d7016f6b688981a2 /net/ipv4/tcp_ulp.c
parentnet_sched: fix a memory leak in atm_tc_init() (diff)
downloadlinux-365f9ae4ee36037e2a9268fe7296065356840b4c.tar.xz
linux-365f9ae4ee36037e2a9268fe7296065356840b4c.zip
ethtool: fix genlmsg_put() failure handling in ethnl_default_dumpit()
If the genlmsg_put() call in ethnl_default_dumpit() fails, we bail out without checking if we already have some messages in current skb like we do with ethnl_default_dump_one() failure later. Therefore if existing messages almost fill up the buffer so that there is not enough space even for netlink and genetlink header, we lose all prepared messages and return and error. Rather than duplicating the skb->len check, move the genlmsg_put(), genlmsg_cancel() and genlmsg_end() calls into ethnl_default_dump_one(). This is also more logical as all message composition will be in ethnl_default_dump_one() and only iteration logic will be left in ethnl_default_dumpit(). Fixes: 728480f12442 ("ethtool: default handlers for GET requests") Reported-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Michal Kubecek <mkubecek@suse.cz> Reviewed-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/ipv4/tcp_ulp.c')
0 files changed, 0 insertions, 0 deletions