| Message ID | 1690892953-25201-1-git-send-email-schakrabarti@linux.microsoft.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [V7,net] net: mana: Fix MANA VF unload when hardware is | expand |
On Tue, Aug 01, 2023 at 05:29:13AM -0700, Souradeep Chakrabarti wrote: ... Hi Souradeep, > + for (i = 0; i < apc->num_queues; i++) { > + txq = &apc->tx_qp[i].txq; > + while (skb = skb_dequeue(&txq->pending_skbs)) { W=1 builds with both clang-16 and gcc-12 complain that they would like an extra set of parentheses around an assignment used as a truth value. > + mana_unmap_skb(skb, apc); > + dev_consume_skb_any(skb); > + } > + atomic_set(&txq->pending_sends, 0); > + } > /* We're 100% sure the queues can no longer be woken up, because > * we're sure now mana_poll_tx_cq() can't be running. > */ > -- > 2.34.1 > >
On 8/1/2023 5:29 AM, Souradeep Chakrabarti wrote: > When unloading the MANA driver, mana_dealloc_queues() waits for the MANA > hardware to complete any inflight packets and set the pending send count > to zero. But if the hardware has failed, mana_dealloc_queues() > could wait forever. > > Fix this by adding a timeout to the wait. Set the timeout to 120 seconds, tx timeout in stack defaults to 5 seconds, do you not have that on? What happens when you start getting resets while unloading? > which is a somewhat arbitrary value that is more than long enough for > functional hardware to complete any sends. I'd say 2 or 5 seconds is probably plenty of time to hang up a driver unload. > > Cc: stable@vger.kernel.org > Fixes: ca9c54d2d6a5 ("net: mana: Add a driver for Microsoft Azure Network Adapter (MANA)") > > Signed-off-by: Souradeep Chakrabarti <schakrabarti@linux.microsoft.com> keep s-o-b and other trailers together please, no spaces, it messes up git and doesn't conform to kernel standards. > --- > V6 -> V7: > * Optimized the while loop for freeing skb. > > V5 -> V6: > * Added pcie_flr to reset the pci after timeout. > * Fixed the position of changelog. > * Removed unused variable like cq. > > V4 -> V5: > * Added fixes tag > * Changed the usleep_range from static to incremental value. > * Initialized timeout in the begining. > > V3 -> V4: > * Removed the unnecessary braces from mana_dealloc_queues(). > > V2 -> V3: > * Removed the unnecessary braces from mana_dealloc_queues(). > > V1 -> V2: > * Added net branch > * Removed the typecasting to (struct mana_context*) of void pointer > * Repositioned timeout variable in mana_dealloc_queues() > * Repositioned vf_unload_timeout in mana_context struct, to utilise the > 6 bytes hole > --- > drivers/net/ethernet/microsoft/mana/mana_en.c | 37 +++++++++++++++++-- > 1 file changed, 33 insertions(+), 4 deletions(-) > > diff --git a/drivers/net/ethernet/microsoft/mana/mana_en.c b/drivers/net/ethernet/microsoft/mana/mana_en.c > index a499e460594b..3c5552a176d0 100644 > --- a/drivers/net/ethernet/microsoft/mana/mana_en.c > +++ b/drivers/net/ethernet/microsoft/mana/mana_en.c > @@ -8,6 +8,7 @@ > #include <linux/ethtool.h> > #include <linux/filter.h> > #include <linux/mm.h> > +#include <linux/pci.h> > > #include <net/checksum.h> > #include <net/ip6_checksum.h> > @@ -2345,9 +2346,12 @@ int mana_attach(struct net_device *ndev) > static int mana_dealloc_queues(struct net_device *ndev) > { > struct mana_port_context *apc = netdev_priv(ndev); > + unsigned long timeout = jiffies + 120 * HZ; > struct gdma_dev *gd = apc->ac->gdma_dev; > struct mana_txq *txq; > + struct sk_buff *skb; > int i, err; > + u32 tsleep; > > if (apc->port_is_up) > return -EINVAL; > @@ -2363,15 +2367,40 @@ static int mana_dealloc_queues(struct net_device *ndev) > * to false, but it doesn't matter since mana_start_xmit() drops any > * new packets due to apc->port_is_up being false. > * > - * Drain all the in-flight TX packets > + * Drain all the in-flight TX packets. > + * A timeout of 120 seconds for all the queues is used. > + * This will break the while loop when h/w is not responding. > + * This value of 120 has been decided here considering max > + * number of queues. > */ > + > for (i = 0; i < apc->num_queues; i++) { > txq = &apc->tx_qp[i].txq; > - > - while (atomic_read(&txq->pending_sends) > 0) > - usleep_range(1000, 2000); > + tsleep = 1000; > + while (atomic_read(&txq->pending_sends) > 0 && > + time_before(jiffies, timeout)) { > + usleep_range(tsleep, tsleep + 1000); > + tsleep <<= 1; > + } > + if (atomic_read(&txq->pending_sends)) { > + err = pcie_flr(to_pci_dev(gd->gdma_context->dev)); > + if (err) { > + netdev_err(ndev, "flr failed %d with %d pkts pending in txq %u\n", > + err, atomic_read(&txq->pending_sends), > + txq->gdma_txq_id); > + } > + break; > + } > } > > + for (i = 0; i < apc->num_queues; i++) { > + txq = &apc->tx_qp[i].txq; > + while (skb = skb_dequeue(&txq->pending_skbs)) { > + mana_unmap_skb(skb, apc); > + dev_consume_skb_any(skb); dev_kfree_skb_any() would be more correct here since this is an error path and the transmit is presumed dropped, isn't it? > + } > + atomic_set(&txq->pending_sends, 0); > + } > /* We're 100% sure the queues can no longer be woken up, because > * we're sure now mana_poll_tx_cq() can't be running. > */
>-----Original Message----- >From: Jesse Brandeburg <jesse.brandeburg@intel.com> >Sent: Tuesday, August 1, 2023 11:34 PM >To: Souradeep Chakrabarti <schakrabarti@linux.microsoft.com>; KY Srinivasan ><kys@microsoft.com>; Haiyang Zhang <haiyangz@microsoft.com>; >wei.liu@kernel.org; Dexuan Cui <decui@microsoft.com>; >davem@davemloft.net; edumazet@google.com; kuba@kernel.org; >pabeni@redhat.com; Long Li <longli@microsoft.com>; Ajay Sharma ><sharmaajay@microsoft.com>; leon@kernel.org; cai.huoqing@linux.dev; >ssengar@linux.microsoft.com; vkuznets <vkuznets@redhat.com>; >tglx@linutronix.de; linux-hyperv@vger.kernel.org; netdev@vger.kernel.org; >linux-kernel@vger.kernel.org; linux-rdma@vger.kernel.org >Cc: Souradeep Chakrabarti <schakrabarti@microsoft.com>; >stable@vger.kernel.org >Subject: [EXTERNAL] Re: [PATCH V7 net] net: mana: Fix MANA VF unload when >hardware is > >On 8/1/2023 5:29 AM, Souradeep Chakrabarti wrote: >> When unloading the MANA driver, mana_dealloc_queues() waits for the >> MANA hardware to complete any inflight packets and set the pending >> send count to zero. But if the hardware has failed, >> mana_dealloc_queues() could wait forever. >> >> Fix this by adding a timeout to the wait. Set the timeout to 120 >> seconds, > >tx timeout in stack defaults to 5 seconds, do you not have that on? What >happens when you start getting resets while unloading? > >> which is a somewhat arbitrary value that is more than long enough for >> functional hardware to complete any sends. > >I'd say 2 or 5 seconds is probably plenty of time to hang up a driver unload. > Thank you for the comment. This was already discussed in V4. I am just sharing the summary here: This waiting time is usually much shorter than 120 sec. The long wait only happens in rare and unexpected NIC HW non-responding cases. At that point, we don't actually care if the pending packets are sent or not. But if we free the queues too soon, and the HW is slow for unexpected reasons, a delayed completion notice will DMA into the freed memory and cause corruption. That's why we have a longer waiting time. >> >> Cc: stable@vger.kernel.org >> Fixes: ca9c54d2d6a5 ("net: mana: Add a driver for Microsoft Azure >> Network Adapter (MANA)") >> >> Signed-off-by: Souradeep Chakrabarti >> <schakrabarti@linux.microsoft.com> > >keep s-o-b and other trailers together please, no spaces, it messes up git and >doesn't conform to kernel standards. > > >> --- >> V6 -> V7: >> * Optimized the while loop for freeing skb. >> >> V5 -> V6: >> * Added pcie_flr to reset the pci after timeout. >> * Fixed the position of changelog. >> * Removed unused variable like cq. >> >> V4 -> V5: >> * Added fixes tag >> * Changed the usleep_range from static to incremental value. >> * Initialized timeout in the begining. >> >> V3 -> V4: >> * Removed the unnecessary braces from mana_dealloc_queues(). >> >> V2 -> V3: >> * Removed the unnecessary braces from mana_dealloc_queues(). >> >> V1 -> V2: >> * Added net branch >> * Removed the typecasting to (struct mana_context*) of void pointer >> * Repositioned timeout variable in mana_dealloc_queues() >> * Repositioned vf_unload_timeout in mana_context struct, to utilise >> the >> 6 bytes hole >> --- >> drivers/net/ethernet/microsoft/mana/mana_en.c | 37 >> +++++++++++++++++-- >> 1 file changed, 33 insertions(+), 4 deletions(-) >> >> diff --git a/drivers/net/ethernet/microsoft/mana/mana_en.c >> b/drivers/net/ethernet/microsoft/mana/mana_en.c >> index a499e460594b..3c5552a176d0 100644 >> --- a/drivers/net/ethernet/microsoft/mana/mana_en.c >> +++ b/drivers/net/ethernet/microsoft/mana/mana_en.c >> @@ -8,6 +8,7 @@ >> #include <linux/ethtool.h> >> #include <linux/filter.h> >> #include <linux/mm.h> >> +#include <linux/pci.h> >> >> #include <net/checksum.h> >> #include <net/ip6_checksum.h> >> @@ -2345,9 +2346,12 @@ int mana_attach(struct net_device *ndev) >> static int mana_dealloc_queues(struct net_device *ndev) { >> struct mana_port_context *apc = netdev_priv(ndev); >> + unsigned long timeout = jiffies + 120 * HZ; >> struct gdma_dev *gd = apc->ac->gdma_dev; >> struct mana_txq *txq; >> + struct sk_buff *skb; >> int i, err; >> + u32 tsleep; >> >> if (apc->port_is_up) >> return -EINVAL; >> @@ -2363,15 +2367,40 @@ static int mana_dealloc_queues(struct >net_device *ndev) >> * to false, but it doesn't matter since mana_start_xmit() drops any >> * new packets due to apc->port_is_up being false. >> * >> - * Drain all the in-flight TX packets >> + * Drain all the in-flight TX packets. >> + * A timeout of 120 seconds for all the queues is used. >> + * This will break the while loop when h/w is not responding. >> + * This value of 120 has been decided here considering max >> + * number of queues. >> */ >> + >> for (i = 0; i < apc->num_queues; i++) { >> txq = &apc->tx_qp[i].txq; >> - >> - while (atomic_read(&txq->pending_sends) > 0) >> - usleep_range(1000, 2000); >> + tsleep = 1000; >> + while (atomic_read(&txq->pending_sends) > 0 && >> + time_before(jiffies, timeout)) { >> + usleep_range(tsleep, tsleep + 1000); >> + tsleep <<= 1; >> + } >> + if (atomic_read(&txq->pending_sends)) { >> + err = pcie_flr(to_pci_dev(gd->gdma_context->dev)); >> + if (err) { >> + netdev_err(ndev, "flr failed %d with %d pkts >pending in txq %u\n", >> + err, atomic_read(&txq- >>pending_sends), >> + txq->gdma_txq_id); >> + } >> + break; >> + } >> } >> >> + for (i = 0; i < apc->num_queues; i++) { >> + txq = &apc->tx_qp[i].txq; >> + while (skb = skb_dequeue(&txq->pending_skbs)) { >> + mana_unmap_skb(skb, apc); >> + dev_consume_skb_any(skb); > >dev_kfree_skb_any() would be more correct here since this is an error path and >the transmit is presumed dropped, isn't it? Yes, dev_kfree_skb_any() will be a better approach in this scenario. Will change it in next version. > > >> + } >> + atomic_set(&txq->pending_sends, 0); >> + } >> /* We're 100% sure the queues can no longer be woken up, because >> * we're sure now mana_poll_tx_cq() can't be running. >> */
>-----Original Message----- >From: Simon Horman <horms@kernel.org> >Sent: Tuesday, August 1, 2023 9:01 PM >To: Souradeep Chakrabarti <schakrabarti@linux.microsoft.com> >Cc: KY Srinivasan <kys@microsoft.com>; Haiyang Zhang ><haiyangz@microsoft.com>; wei.liu@kernel.org; Dexuan Cui ><decui@microsoft.com>; davem@davemloft.net; edumazet@google.com; >kuba@kernel.org; pabeni@redhat.com; Long Li <longli@microsoft.com>; Ajay >Sharma <sharmaajay@microsoft.com>; leon@kernel.org; >cai.huoqing@linux.dev; ssengar@linux.microsoft.com; vkuznets ><vkuznets@redhat.com>; tglx@linutronix.de; linux-hyperv@vger.kernel.org; >netdev@vger.kernel.org; linux-kernel@vger.kernel.org; linux- >rdma@vger.kernel.org; Souradeep Chakrabarti ><schakrabarti@microsoft.com>; stable@vger.kernel.org >Subject: [EXTERNAL] Re: [PATCH V7 net] net: mana: Fix MANA VF unload when >hardware is > >On Tue, Aug 01, 2023 at 05:29:13AM -0700, Souradeep Chakrabarti wrote: > >... > >Hi Souradeep, > > >> + for (i = 0; i < apc->num_queues; i++) { >> + txq = &apc->tx_qp[i].txq; >> + while (skb = skb_dequeue(&txq->pending_skbs)) { > >W=1 builds with both clang-16 and gcc-12 complain that they would like an >extra set of parentheses around an assignment used as a truth value. Thanks for letting me know. I will fix it in next version. > >> + mana_unmap_skb(skb, apc); >> + dev_consume_skb_any(skb); >> + } >> + atomic_set(&txq->pending_sends, 0); >> + } >> /* We're 100% sure the queues can no longer be woken up, because >> * we're sure now mana_poll_tx_cq() can't be running. >> */ >> -- >> 2.34.1 >> >>
Hi Souradeep, It looks like the subject line is not complete. I could see "net: mana: Fix MANA VF unload when hardware is". Is that correct? Regards, Kalesh On Wed, Aug 2, 2023 at 12:29 AM Souradeep Chakrabarti < schakrabarti@microsoft.com> wrote: > > > >-----Original Message----- > >From: Simon Horman <horms@kernel.org> > >Sent: Tuesday, August 1, 2023 9:01 PM > >To: Souradeep Chakrabarti <schakrabarti@linux.microsoft.com> > >Cc: KY Srinivasan <kys@microsoft.com>; Haiyang Zhang > ><haiyangz@microsoft.com>; wei.liu@kernel.org; Dexuan Cui > ><decui@microsoft.com>; davem@davemloft.net; edumazet@google.com; > >kuba@kernel.org; pabeni@redhat.com; Long Li <longli@microsoft.com>; Ajay > >Sharma <sharmaajay@microsoft.com>; leon@kernel.org; > >cai.huoqing@linux.dev; ssengar@linux.microsoft.com; vkuznets > ><vkuznets@redhat.com>; tglx@linutronix.de; linux-hyperv@vger.kernel.org; > >netdev@vger.kernel.org; linux-kernel@vger.kernel.org; linux- > >rdma@vger.kernel.org; Souradeep Chakrabarti > ><schakrabarti@microsoft.com>; stable@vger.kernel.org > >Subject: [EXTERNAL] Re: [PATCH V7 net] net: mana: Fix MANA VF unload when > >hardware is > > > >On Tue, Aug 01, 2023 at 05:29:13AM -0700, Souradeep Chakrabarti wrote: > > > >... > > > >Hi Souradeep, > > > > > >> + for (i = 0; i < apc->num_queues; i++) { > >> + txq = &apc->tx_qp[i].txq; > >> + while (skb = skb_dequeue(&txq->pending_skbs)) { > > > >W=1 builds with both clang-16 and gcc-12 complain that they would like an > >extra set of parentheses around an assignment used as a truth value. > Thanks for letting me know. I will fix it in next version. > > > >> + mana_unmap_skb(skb, apc); > >> + dev_consume_skb_any(skb); > >> + } > >> + atomic_set(&txq->pending_sends, 0); > >> + } > >> /* We're 100% sure the queues can no longer be woken up, because > >> * we're sure now mana_poll_tx_cq() can't be running. > >> */ > >> -- > >> 2.34.1 > >> > >> > >
On Wed, Aug 02, 2023 at 10:57:52AM +0530, Kalesh Anakkur Purayil wrote: > Hi Souradeep, > > It looks like the subject line is not complete. I could see "net: mana: Fix > MANA VF unload when hardware is". > > Is that correct? > > Regards, > Kalesh > Yes, it got truncated. Will fix it in next version. > On Wed, Aug 2, 2023 at 12:29 AM Souradeep Chakrabarti < > schakrabarti@microsoft.com> wrote: > > > > > > > >-----Original Message----- > > >From: Simon Horman <horms@kernel.org> > > >Sent: Tuesday, August 1, 2023 9:01 PM > > >To: Souradeep Chakrabarti <schakrabarti@linux.microsoft.com> > > >Cc: KY Srinivasan <kys@microsoft.com>; Haiyang Zhang > > ><haiyangz@microsoft.com>; wei.liu@kernel.org; Dexuan Cui > > ><decui@microsoft.com>; davem@davemloft.net; edumazet@google.com; > > >kuba@kernel.org; pabeni@redhat.com; Long Li <longli@microsoft.com>; Ajay > > >Sharma <sharmaajay@microsoft.com>; leon@kernel.org; > > >cai.huoqing@linux.dev; ssengar@linux.microsoft.com; vkuznets > > ><vkuznets@redhat.com>; tglx@linutronix.de; linux-hyperv@vger.kernel.org; > > >netdev@vger.kernel.org; linux-kernel@vger.kernel.org; linux- > > >rdma@vger.kernel.org; Souradeep Chakrabarti > > ><schakrabarti@microsoft.com>; stable@vger.kernel.org > > >Subject: [EXTERNAL] Re: [PATCH V7 net] net: mana: Fix MANA VF unload when > > >hardware is > > > > > >On Tue, Aug 01, 2023 at 05:29:13AM -0700, Souradeep Chakrabarti wrote: > > > > > >... > > > > > >Hi Souradeep, > > > > > > > > >> + for (i = 0; i < apc->num_queues; i++) { > > >> + txq = &apc->tx_qp[i].txq; > > >> + while (skb = skb_dequeue(&txq->pending_skbs)) { > > > > > >W=1 builds with both clang-16 and gcc-12 complain that they would like an > > >extra set of parentheses around an assignment used as a truth value. > > Thanks for letting me know. I will fix it in next version. > > > > > >> + mana_unmap_skb(skb, apc); > > >> + dev_consume_skb_any(skb); > > >> + } > > >> + atomic_set(&txq->pending_sends, 0); > > >> + } > > >> /* We're 100% sure the queues can no longer be woken up, because > > >> * we're sure now mana_poll_tx_cq() can't be running. > > >> */ > > >> -- > > >> 2.34.1 > > >> > > >> > > > > > > -- > Regards, > Kalesh A P
Hi Souradeep, kernel test robot noticed the following build warnings: [auto build test WARNING on net/main] url: https://github.com/intel-lab-lkp/linux/commits/Souradeep-Chakrabarti/net-mana-Fix-MANA-VF-unload-when-hardware-is/20230801-203141 base: net/main patch link: https://lore.kernel.org/r/1690892953-25201-1-git-send-email-schakrabarti%40linux.microsoft.com patch subject: [PATCH V7 net] net: mana: Fix MANA VF unload when hardware is config: x86_64-allyesconfig (https://download.01.org/0day-ci/archive/20230802/202308021532.8iYkExDh-lkp@intel.com/config) compiler: gcc-12 (Debian 12.2.0-14) 12.2.0 reproduce: (https://download.01.org/0day-ci/archive/20230802/202308021532.8iYkExDh-lkp@intel.com/reproduce) If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot <lkp@intel.com> | Closes: https://lore.kernel.org/oe-kbuild-all/202308021532.8iYkExDh-lkp@intel.com/ All warnings (new ones prefixed by >>): drivers/net/ethernet/microsoft/mana/mana_en.c: In function 'mana_dealloc_queues': >> drivers/net/ethernet/microsoft/mana/mana_en.c:2398:24: warning: suggest parentheses around assignment used as truth value [-Wparentheses] 2398 | while (skb = skb_dequeue(&txq->pending_skbs)) { | ^~~ vim +2398 drivers/net/ethernet/microsoft/mana/mana_en.c 2345 2346 static int mana_dealloc_queues(struct net_device *ndev) 2347 { 2348 struct mana_port_context *apc = netdev_priv(ndev); 2349 unsigned long timeout = jiffies + 120 * HZ; 2350 struct gdma_dev *gd = apc->ac->gdma_dev; 2351 struct mana_txq *txq; 2352 struct sk_buff *skb; 2353 int i, err; 2354 u32 tsleep; 2355 2356 if (apc->port_is_up) 2357 return -EINVAL; 2358 2359 mana_chn_setxdp(apc, NULL); 2360 2361 if (gd->gdma_context->is_pf) 2362 mana_pf_deregister_filter(apc); 2363 2364 /* No packet can be transmitted now since apc->port_is_up is false. 2365 * There is still a tiny chance that mana_poll_tx_cq() can re-enable 2366 * a txq because it may not timely see apc->port_is_up being cleared 2367 * to false, but it doesn't matter since mana_start_xmit() drops any 2368 * new packets due to apc->port_is_up being false. 2369 * 2370 * Drain all the in-flight TX packets. 2371 * A timeout of 120 seconds for all the queues is used. 2372 * This will break the while loop when h/w is not responding. 2373 * This value of 120 has been decided here considering max 2374 * number of queues. 2375 */ 2376 2377 for (i = 0; i < apc->num_queues; i++) { 2378 txq = &apc->tx_qp[i].txq; 2379 tsleep = 1000; 2380 while (atomic_read(&txq->pending_sends) > 0 && 2381 time_before(jiffies, timeout)) { 2382 usleep_range(tsleep, tsleep + 1000); 2383 tsleep <<= 1; 2384 } 2385 if (atomic_read(&txq->pending_sends)) { 2386 err = pcie_flr(to_pci_dev(gd->gdma_context->dev)); 2387 if (err) { 2388 netdev_err(ndev, "flr failed %d with %d pkts pending in txq %u\n", 2389 err, atomic_read(&txq->pending_sends), 2390 txq->gdma_txq_id); 2391 } 2392 break; 2393 } 2394 } 2395 2396 for (i = 0; i < apc->num_queues; i++) { 2397 txq = &apc->tx_qp[i].txq; > 2398 while (skb = skb_dequeue(&txq->pending_skbs)) { 2399 mana_unmap_skb(skb, apc); 2400 dev_consume_skb_any(skb); 2401 } 2402 atomic_set(&txq->pending_sends, 0); 2403 } 2404 /* We're 100% sure the queues can no longer be woken up, because 2405 * we're sure now mana_poll_tx_cq() can't be running. 2406 */ 2407 2408 apc->rss_state = TRI_STATE_FALSE; 2409 err = mana_config_rss(apc, TRI_STATE_FALSE, false, false); 2410 if (err) { 2411 netdev_err(ndev, "Failed to disable vPort: %d\n", err); 2412 return err; 2413 } 2414 2415 mana_destroy_vport(apc); 2416 2417 return 0; 2418 } 2419
diff --git a/drivers/net/ethernet/microsoft/mana/mana_en.c b/drivers/net/ethernet/microsoft/mana/mana_en.c index a499e460594b..3c5552a176d0 100644 --- a/drivers/net/ethernet/microsoft/mana/mana_en.c +++ b/drivers/net/ethernet/microsoft/mana/mana_en.c @@ -8,6 +8,7 @@ #include <linux/ethtool.h> #include <linux/filter.h> #include <linux/mm.h> +#include <linux/pci.h> #include <net/checksum.h> #include <net/ip6_checksum.h> @@ -2345,9 +2346,12 @@ int mana_attach(struct net_device *ndev) static int mana_dealloc_queues(struct net_device *ndev) { struct mana_port_context *apc = netdev_priv(ndev); + unsigned long timeout = jiffies + 120 * HZ; struct gdma_dev *gd = apc->ac->gdma_dev; struct mana_txq *txq; + struct sk_buff *skb; int i, err; + u32 tsleep; if (apc->port_is_up) return -EINVAL; @@ -2363,15 +2367,40 @@ static int mana_dealloc_queues(struct net_device *ndev) * to false, but it doesn't matter since mana_start_xmit() drops any * new packets due to apc->port_is_up being false. * - * Drain all the in-flight TX packets + * Drain all the in-flight TX packets. + * A timeout of 120 seconds for all the queues is used. + * This will break the while loop when h/w is not responding. + * This value of 120 has been decided here considering max + * number of queues. */ + for (i = 0; i < apc->num_queues; i++) { txq = &apc->tx_qp[i].txq; - - while (atomic_read(&txq->pending_sends) > 0) - usleep_range(1000, 2000); + tsleep = 1000; + while (atomic_read(&txq->pending_sends) > 0 && + time_before(jiffies, timeout)) { + usleep_range(tsleep, tsleep + 1000); + tsleep <<= 1; + } + if (atomic_read(&txq->pending_sends)) { + err = pcie_flr(to_pci_dev(gd->gdma_context->dev)); + if (err) { + netdev_err(ndev, "flr failed %d with %d pkts pending in txq %u\n", + err, atomic_read(&txq->pending_sends), + txq->gdma_txq_id); + } + break; + } } + for (i = 0; i < apc->num_queues; i++) { + txq = &apc->tx_qp[i].txq; + while (skb = skb_dequeue(&txq->pending_skbs)) { + mana_unmap_skb(skb, apc); + dev_consume_skb_any(skb); + } + atomic_set(&txq->pending_sends, 0); + } /* We're 100% sure the queues can no longer be woken up, because * we're sure now mana_poll_tx_cq() can't be running. */
When unloading the MANA driver, mana_dealloc_queues() waits for the MANA hardware to complete any inflight packets and set the pending send count to zero. But if the hardware has failed, mana_dealloc_queues() could wait forever. Fix this by adding a timeout to the wait. Set the timeout to 120 seconds, which is a somewhat arbitrary value that is more than long enough for functional hardware to complete any sends. Cc: stable@vger.kernel.org Fixes: ca9c54d2d6a5 ("net: mana: Add a driver for Microsoft Azure Network Adapter (MANA)") Signed-off-by: Souradeep Chakrabarti <schakrabarti@linux.microsoft.com> --- V6 -> V7: * Optimized the while loop for freeing skb. V5 -> V6: * Added pcie_flr to reset the pci after timeout. * Fixed the position of changelog. * Removed unused variable like cq. V4 -> V5: * Added fixes tag * Changed the usleep_range from static to incremental value. * Initialized timeout in the begining. V3 -> V4: * Removed the unnecessary braces from mana_dealloc_queues(). V2 -> V3: * Removed the unnecessary braces from mana_dealloc_queues(). V1 -> V2: * Added net branch * Removed the typecasting to (struct mana_context*) of void pointer * Repositioned timeout variable in mana_dealloc_queues() * Repositioned vf_unload_timeout in mana_context struct, to utilise the 6 bytes hole --- drivers/net/ethernet/microsoft/mana/mana_en.c | 37 +++++++++++++++++-- 1 file changed, 33 insertions(+), 4 deletions(-)