| Message ID | 1359854463-2538-10-git-send-email-tj@kernel.org |
|---|---|
| State | Not Applicable, archived |
| Delegated to: | David Miller |
| Headers | show |
I don't quite understand your comment. idr_pre_get() returns 0 in the case of failure. On Sat, 2 Feb 2013 17:20:10 -0800 Tejun Heo <tj@kernel.org> wrote: > Convert to the much saner new idr interface. The existing code looks > buggy to me - ID 0 is treated as no-ID but allocation specifies 0 as > lower limit and there's no error handling after parial success. This > conversion keeps the bugs unchanged. > > Only compile tested. > > Signed-off-by: Tejun Heo <tj@kernel.org> > Cc: Chas Williams <chas@cmf.nrl.navy.mil> > Cc: netdev@vger.kernel.org > --- > This patch depends on an earlier idr changes and I think it would be > best to route these together through -mm. Please holler if there's > any objection. Thanks. > > drivers/atm/nicstar.c | 27 ++++++++++++--------------- > 1 file changed, 12 insertions(+), 15 deletions(-) > > diff --git a/drivers/atm/nicstar.c b/drivers/atm/nicstar.c > index 628787e..7fd6834 100644 > --- a/drivers/atm/nicstar.c > +++ b/drivers/atm/nicstar.c > @@ -1026,24 +1026,21 @@ static void push_rxbufs(ns_dev * card, struct sk_buff *skb) > card->lbfqc += 2; > } > > - do { > - if (!idr_pre_get(&card->idr, GFP_ATOMIC)) { > - printk(KERN_ERR > - "nicstar%d: no free memory for idr\n", > - card->index); > + if (!id1) { > + id1 = idr_alloc(&card->idr, handle1, 0, 0, GFP_ATOMIC); > + if (id1 < 0) { > + err = id1; > goto out; > } > + } > > - if (!id1) > - err = idr_get_new_above(&card->idr, handle1, 0, &id1); > - > - if (!id2 && err == 0) > - err = idr_get_new_above(&card->idr, handle2, 0, &id2); > - > - } while (err == -EAGAIN); > - > - if (err) > - goto out; > + if (!id2) { > + id2 = idr_alloc(&card->idr, handle2, 0, 0, GFP_ATOMIC); > + if (id2 < 0) { > + err = id2; > + goto out; > + } > + } > > spin_lock_irqsave(&card->res_lock, flags); > while (CMD_BUSY(card)) ; -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Hello, Chas. On Mon, Feb 04, 2013 at 09:04:10AM -0500, chas williams - CONTRACTOR wrote: > I don't quite understand your comment. idr_pre_get() returns 0 in the > case of failure. So, if you do the following, int id1 = 0, id2 = 0; if (!id1) err = idr_get_new_above(&card->idr, handle1, 0, &id1); if (!id2 && err == 0) err = idr_get_new_above(&card->idr, handle2, 0, &id2); if (err) goto out; You have no way of telling whether id1/2 are allocated or not. 0 is the special "not allocated" value but it also is a valid ID. The error path should be freeing id1/2 if either of them has been allocated but it doesn't and can't with 0 as the non-allocated value. Thanks.
On Mon, 4 Feb 2013 09:06:24 -0800 Tejun Heo <tj@kernel.org> wrote: > Hello, Chas. > > On Mon, Feb 04, 2013 at 09:04:10AM -0500, chas williams - CONTRACTOR wrote: > > I don't quite understand your comment. idr_pre_get() returns 0 in the > > case of failure. > > So, if you do the following, > > int id1 = 0, id2 = 0; > > if (!id1) > err = idr_get_new_above(&card->idr, handle1, 0, &id1); > if (!id2 && err == 0) > err = idr_get_new_above(&card->idr, handle2, 0, &id2); > if (err) > goto out; > > You have no way of telling whether id1/2 are allocated or not. 0 is > the special "not allocated" value but it also is a valid ID. The > error path should be freeing id1/2 if either of them has been > allocated but it doesn't and can't with 0 as the non-allocated value. yeah, i see now. it didn't understand what idr_get_new_above() was doing with the start id. i assumed that it would return something greater than the start id. i guess it never did return the starting id (atleast after some initial failure). additionally, yes, cleaning up if the second allocation failed was never done. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/atm/nicstar.c b/drivers/atm/nicstar.c index 628787e..7fd6834 100644 --- a/drivers/atm/nicstar.c +++ b/drivers/atm/nicstar.c @@ -1026,24 +1026,21 @@ static void push_rxbufs(ns_dev * card, struct sk_buff *skb) card->lbfqc += 2; } - do { - if (!idr_pre_get(&card->idr, GFP_ATOMIC)) { - printk(KERN_ERR - "nicstar%d: no free memory for idr\n", - card->index); + if (!id1) { + id1 = idr_alloc(&card->idr, handle1, 0, 0, GFP_ATOMIC); + if (id1 < 0) { + err = id1; goto out; } + } - if (!id1) - err = idr_get_new_above(&card->idr, handle1, 0, &id1); - - if (!id2 && err == 0) - err = idr_get_new_above(&card->idr, handle2, 0, &id2); - - } while (err == -EAGAIN); - - if (err) - goto out; + if (!id2) { + id2 = idr_alloc(&card->idr, handle2, 0, 0, GFP_ATOMIC); + if (id2 < 0) { + err = id2; + goto out; + } + } spin_lock_irqsave(&card->res_lock, flags); while (CMD_BUSY(card)) ;
Convert to the much saner new idr interface. The existing code looks buggy to me - ID 0 is treated as no-ID but allocation specifies 0 as lower limit and there's no error handling after parial success. This conversion keeps the bugs unchanged. Only compile tested. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Chas Williams <chas@cmf.nrl.navy.mil> Cc: netdev@vger.kernel.org --- This patch depends on an earlier idr changes and I think it would be best to route these together through -mm. Please holler if there's any objection. Thanks. drivers/atm/nicstar.c | 27 ++++++++++++--------------- 1 file changed, 12 insertions(+), 15 deletions(-)