I do not approve this patch. The type field is always set to either NES_TIMER_TYPE_SEND or NES_TIMER_TYPE_CLOSE when schedule_nes_timer() is called. Even if we wanted to make sure the value of type is one or the other, proper patch can be just a if stmt to check the type before new_send allocation.
Thanks Faisal > -----Original Message----- > From: [email protected] [mailto:linux-rdma- > [email protected]] On Behalf Of Jesper Juhl > Sent: Monday, February 20, 2012 2:56 PM > To: [email protected] > Cc: [email protected]; Hal Rosenstock; Hefty, Sean; Roland > Dreier; Latif, Faisal > Subject: [PATCH] IB: Fix potential mem leak in schedule_nes_timer() > > In drivers/infiniband/hw/nes/nes_cm.c::schedule_nes_timer(), if we do > not take the true branch of either the > (type == NES_TIMER_TYPE_SEND) > or > (type == NES_TIMER_TYPE_SEND) > 'if' statements (which it looks like we potentially could since that > enum can have other values), then we'll leak the memory allocated to > 'new_send'. > > This (admittedly, rather clumsy) patch addresses that issue by turning > the two 'if's into 'if' 'else if' and then adding a 'else' branch > where we note that we need to free the mem and then later, when we are > done with it, does a conditional kfree() based on the bool we set in > the 'else' branch. > > I also removed a redundant variable ('was_timer_set'). > > Signed-off-by: Jesper Juhl <[email protected]> > --- > drivers/infiniband/hw/nes/nes_cm.c | 15 ++++++++------- > 1 files changed, 8 insertions(+), 7 deletions(-) > > Compile tested only. > > diff --git a/drivers/infiniband/hw/nes/nes_cm.c > b/drivers/infiniband/hw/nes/nes_cm.c > index a4972ab..434a25b 100644 > --- a/drivers/infiniband/hw/nes/nes_cm.c > +++ b/drivers/infiniband/hw/nes/nes_cm.c > @@ -663,8 +663,8 @@ int schedule_nes_timer(struct nes_cm_node *cm_node, > struct sk_buff *skb, > unsigned long flags; > struct nes_cm_core *cm_core = cm_node->cm_core; > struct nes_timer_entry *new_send; > + bool free_new_send = false; > int ret = 0; > - u32 was_timer_set; > > new_send = kzalloc(sizeof(*new_send), GFP_ATOMIC); > if (!new_send) > @@ -688,9 +688,7 @@ int schedule_nes_timer(struct nes_cm_node *cm_node, > struct sk_buff *skb, > return -EINVAL; > } > cm_node->recv_entry = new_send; > - } > - > - if (type == NES_TIMER_TYPE_SEND) { > + } else if (type == NES_TIMER_TYPE_SEND) { > new_send->seq_num = ntohl(tcp_hdr(skb)->seq); > atomic_inc(&new_send->skb->users); > spin_lock_irqsave(&cm_node->retrans_list_lock, flags); > @@ -714,15 +712,18 @@ int schedule_nes_timer(struct nes_cm_node > *cm_node, struct sk_buff *skb, > return ret; > } > } > + } else { > + free_new_send = true; > } > > - was_timer_set = timer_pending(&cm_core->tcp_timer); > - > - if (!was_timer_set) { > + if (!timer_pending(&cm_core->tcp_timer)) { > cm_core->tcp_timer.expires = new_send->timetosend; > add_timer(&cm_core->tcp_timer); > } > > + if (free_new_send) > + kfree(new_send); > + > return ret; > } > > -- > 1.7.9.1 > > > -- > Jesper Juhl <[email protected]> http://www.chaosbits.net/ > Don't top-post http://www.catb.org/jargon/html/T/top-post.html > Plain text mails only, please. > > -- > To unsubscribe from this list: send the line "unsubscribe linux-rdma" > in > the body of a message to [email protected] > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
