* [PATCH net-next-2.6] sfq: fix sfq stats handling
From: Eric Dumazet @ 2010-12-22 6:14 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Jarek Poplawski
David, since this one fixes a bug, could you apply it before the pending
SFQ patch (sch_sfq: allow big packets and be fair), I'll respin this one
if necessary.
Thanks
[PATCH net-next-2.6] sfq: fix sfq class stats handling
sfq_walk() runs without qdisc lock. By the time it selects a non empty
hash slot and sfq_dump_class_stats() is run (with lock held), slot might
have been freed : We then access q->slots[SFQ_EMPTY_SLOT], out of
bounds, and crash in slot_queue_walk()
On previous kernels, bug is here but out of bounds qs[SFQ_DEPTH] and
allot[SFQ_DEPTH] are located in struct sfq_sched_data, so no illegal
memory access happens, only possibly wrong data reported to user.
Also, slot_dequeue_tail() should make sure slot skb chain is correctly
terminated, or sfq_dump_class_stats() can access freed skbs.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Jarek Poplawski <jarkao2@gmail.com>
---
net/sched/sch_sfq.c | 16 +++++++++++-----
1 files changed, 11 insertions(+), 5 deletions(-)
diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c
index 13322e8..6a2f88f 100644
--- a/net/sched/sch_sfq.c
+++ b/net/sched/sch_sfq.c
@@ -281,6 +281,7 @@ static inline struct sk_buff *slot_dequeue_tail(struct sfq_slot *slot)
struct sk_buff *skb = slot->skblist_prev;
slot->skblist_prev = skb->prev;
+ skb->prev->next = (struct sk_buff *)slot;
skb->next = skb->prev = NULL;
return skb;
}
@@ -608,14 +609,19 @@ static int sfq_dump_class_stats(struct Qdisc *sch, unsigned long cl,
struct gnet_dump *d)
{
struct sfq_sched_data *q = qdisc_priv(sch);
- const struct sfq_slot *slot = &q->slots[q->ht[cl - 1]];
- struct gnet_stats_queue qs = { .qlen = slot->qlen };
- struct tc_sfq_xstats xstats = { .allot = slot->allot };
+ sfq_index idx = q->ht[cl - 1];
+ struct gnet_stats_queue qs = { 0 };
+ struct tc_sfq_xstats xstats = { 0 };
struct sk_buff *skb;
- slot_queue_walk(slot, skb)
- qs.backlog += qdisc_pkt_len(skb);
+ if (idx != SFQ_EMPTY_SLOT) {
+ const struct sfq_slot *slot = &q->slots[idx];
+ xstats.allot = slot->allot;
+ qs.qlen = slot->qlen;
+ slot_queue_walk(slot, skb)
+ qs.backlog += qdisc_pkt_len(skb);
+ }
if (gnet_stats_copy_queue(d, &qs) < 0)
return -1;
return gnet_stats_copy_app(d, &xstats, sizeof(xstats));
^ permalink raw reply related
* Re: [PATCH net-next-2.6 v2 1/1] can: c_can: Added support for Bosch C_CAN controller
From: Marc Kleine-Budde @ 2010-12-22 6:52 UTC (permalink / raw)
To: Bhupesh SHARMA
Cc: Socketcan-core-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org,
netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
Wolfgang Grandegger
In-Reply-To: <D5ECB3C7A6F99444980976A8C6D896384DEAE48416-8vAmw3ZAcdzhJTuQ9jeba9BPR1lH4CV8@public.gmane.org>
[-- Attachment #1.1: Type: text/plain, Size: 2835 bytes --]
On 12/22/2010 04:36 AM, Bhupesh SHARMA wrote:
> Hi Wolfgang,
>
>> Hi Bhupesh,
>>
>> On 12/21/2010 05:48 AM, Bhupesh SHARMA wrote:
>>> Hi Wolfgang,
>> ...
>>>> In the meantime I compared the CAN chapter of the PCH manual with
>> the
>>>> C_CAN manual. The paragraphs I checked are *identical*. This makes
>>>> clear, that the "pch_can" is a clone of the C_CAN CAN controller,
>> with
>>>> a few extensions, though. Therefore it would make sense, to
>> implement a
>>>> bus sensitive interface like for the SJA1000 allowing to handle both
>>>> CAN
>>>> controllers with one driver sooner than later. Therefore, could you
>>>> please implement:
>>>>
>>>> drivers/net/can/c_can/c_can.c
>>>> /c_can_platform.c
>>>>
>>>> Then an interface to the PCI based PCH CAN controller could be added
>>>> easily, e.g. as "pch_pci.c". You already had something similar in
>> your
>>>> RFC version of the patch, IIRC.
>>>
>>> This was the approach I initially proposed in my RFC V1 patch :)
>>> But unfortunately we could not agree to it.
>>
>> I know. But at that time I was not aware of any other bus used for the
>> C_CAN controller.
>>
>>> So, please let me reiterate what I understood and what was present
>>> in RFC version of the patch. Please add your comments/views:
>>>
>>> - drivers/net/can/c_can/c_can.c (similar on lines of
>> sja1000.c)
>>> i.e. a)no *probe* / *remove* functions here,
>>> b)register read/write implemented here.
>>>
>>> - drivers/net/can/c_can/c_can_platform.c (similar on lines of
>> sja1000_platform.c)
>>> i.e. *probe* / *remove* implemented here,
>>
>> Yes, that's what I'm thinking about.
>>
>>> Marc and Tomoya can also add their suggestions so that I can finalize
>> V3 a.s.a.p.
>>
>> That would be nice, indeed. Also have a look to Tomoya's PCH driver,
>> which also looks very good in the meantime.
>
> I am having a look at Tomoya's PCH driver, but as I mentioned in
> RFC V1 patch, I would rather like to have a bus sensitive `c_can` driver
What do you mean by a "bus sensitive" driver?
> on top of which we can have the platform driver `c_can_platform` which
> essentially caters to the details of registers mapping/arch differences.
> Any other functionality like USB/PCI should be present in a separate file
> like `usb_c_can.c` or `pci_c_can.c`
Sounds like the sja1000 approach, which is a good choice.
> If you agree I will try to circulate V3 a.s.ap.
go ahead.
regards, Marc
--
Pengutronix e.K. | Marc Kleine-Budde |
Industrial Linux Solutions | Phone: +49-231-2826-924 |
Vertretung West/Dortmund | Fax: +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686 | http://www.pengutronix.de |
[-- Attachment #1.2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 262 bytes --]
[-- Attachment #2: Type: text/plain, Size: 188 bytes --]
_______________________________________________
Socketcan-core mailing list
Socketcan-core-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org
https://lists.berlios.de/mailman/listinfo/socketcan-core
^ permalink raw reply
* Re: IPV6 loopback bound socket succeeds connecting to remote host
From: Shan Wei @ 2010-12-22 7:06 UTC (permalink / raw)
To: David Miller; +Cc: albertpretorius, netdev, yoshfuji, pekkas, jmorris
In-Reply-To: <20101219.224350.260095315.davem@davemloft.net>
David Miller wrote, at 12/20/2010 02:43 PM:
> From: Shan Wei <shanwei@cn.fujitsu.com>
> Date: Mon, 20 Dec 2010 14:31:28 +0800
>
>> David Miller wrote, at 12/17/2010 04:18 AM:
>>> Your approach will only modify socket based route handling, it will
>>> not handle the ipv6 forwarding case which as per the quoted RFC
>>> sections must be handled too.
>>
>> For the ipv6 forwarding case, we have done the check in ip6_forward().
>>
>> 493 int addrtype = ipv6_addr_type(&hdr->saddr);
>> 494
>> 495 /* This check is security critical. */
>> 496 if (addrtype == IPV6_ADDR_ANY ||
>> 497 addrtype & (IPV6_ADDR_MULTICAST | IPV6_ADDR_LOOPBACK))
>> 498 goto error;
>
> Indeed, thanks for pointing this out.
Notice that the state in patchwork is “Changes Requested”, what should i do
now? I have no idead which part of this patch should be changed.
--
Best Regards
-----
Shan Wei
^ permalink raw reply
* Re: [PATCH V7 1/8] ntp: add ADJ_SETOFFSET mode bit
From: Richard Cochran @ 2010-12-22 7:11 UTC (permalink / raw)
To: john stultz
Cc: Kuwahara,T., Alexander Gordeev, Rodolfo Giometti,
linux-kernel-u79uwXL29TY76Z2rM5mHXA,
linux-api-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
Alan Cox, Arnd Bergmann, Christoph Lameter, David Miller,
Krzysztof Halasa, Peter Zijlstra, Rodolfo Giometti,
Thomas Gleixner
In-Reply-To: <1292968784.2618.51.camel@work-vm>
On Tue, Dec 21, 2010 at 01:59:44PM -0800, john stultz wrote:
> However, your concern does bring up a good point: 0x40 is MOD_PPSMAX in
> BSD, and we should at-least check to make sure that the PPS code that is
> currently floating around on the lists and is in akpm's tree hasn't
> already reserved that bit.
If the concern is only about that bit, then we can choose another one
from the range 0x0f00 instead.
It seems to me that the ability to jump the clock is a very useful
feature, at least when implementing a clock servo for PTP in user
space, if not in other cases, too. I wonder why that option is missing
from the interface in the first place.
I am not aware of any serious attempt to get any kind of standardized
PTP clock support into any other OS. As far as compatibility with NTP
is concerned, why can't we let Linux set an example and let
NTP/BSD/*nix follow us?
Richard
^ permalink raw reply
* Re: [PATCH V7 1/8] ntp: add ADJ_SETOFFSET mode bit
From: Richard Cochran @ 2010-12-22 7:13 UTC (permalink / raw)
To: john stultz
Cc: Kuwahara,T., linux-kernel, linux-api, netdev, Alan Cox,
Arnd Bergmann, Christoph Lameter, David Miller, Krzysztof Halasa,
Peter Zijlstra, Rodolfo Giometti, Thomas Gleixner
In-Reply-To: <1292970355.2618.76.camel@work-vm>
On Tue, Dec 21, 2010 at 02:25:55PM -0800, john stultz wrote:
> On Wed, 2010-12-22 at 05:57 +0900, Kuwahara,T. wrote:
>
> > How about this?
> >
> > if (txc->modes & ADJ_OFFSET) {
> > if (txc->constant == INT32_MIN) {
> > /* step time */
> > } else {
> > /* slew time */
> > }
> > }
>
> This looks like magic behavior. Sort of a "knock twice and then say the
> password" interface. I don't see why that would be better then adding a
> clear new mode flag?
I have to agree with John on this one. Looks very hacky to me.
Richard
^ permalink raw reply
* Re: Using ethernet device as efficient small packet generator
From: juice @ 2010-12-22 7:30 UTC (permalink / raw)
To: Stephen Hemminger, netdev
> On Tue, 21 Dec 2010 11:56:42 +0200 shemminger wrote:
> I regularly get full 1G line rate of 64 byte packets using old Opteron
box and pktgen. It does require some tuning of IRQ's and interrupt
mitigation but
> no patches. Did you remember to do the basic stuff like setting IRQ
affinity
> and not enabling debugging or tracing in the kernel? This is on sky2,
but
> also using e1000 and tg3. Others have reported 7M packets per second
over
> 10G cards.
> The r8169 hardware is low end consumer hardware and doesn't work as
well.
> It is possible to get close to 1G line rate forwarding with a single
core
> with current
> generation processors. Actual rate depends on hardware and configuration
(size of route
> table, firewalling, etc). Much better performance with multi-queue
hardware to spread load
> over multiple cores.
I did my testing on two kinds of boxes we use in our lab, an older Pomi
Supermicro with e1000 and a newer Dell T3500 with tg3 and r8169.
Both computers have dual-core 2.4G Xeon Cpus, but with somewhat different
model and stepping.
Both boxes are running the same OS, Ubuntu 2.6.32-26-generic #48.
Could you share some information on the required interrupt tuning? It
would certainly be easiest if the full line rate can be achieved without
any patching of drivers or hindering normal eth/ip interface operation.
Yours, Jussi Ohenoja
^ permalink raw reply
* Re: [PATCH net-next-2.6] sfq: fix sfq stats handling
From: Jarek Poplawski @ 2010-12-22 7:32 UTC (permalink / raw)
To: Eric Dumazet; +Cc: David Miller, netdev
In-Reply-To: <1292998499.4317.13.camel@edumazet-laptop>
On Wed, Dec 22, 2010 at 07:14:59AM +0100, Eric Dumazet wrote:
> David, since this one fixes a bug, could you apply it before the pending
> SFQ patch (sch_sfq: allow big packets and be fair), I'll respin this one
> if necessary.
>
> Thanks
>
> [PATCH net-next-2.6] sfq: fix sfq class stats handling
>
> sfq_walk() runs without qdisc lock. By the time it selects a non empty
> hash slot and sfq_dump_class_stats() is run (with lock held), slot might
> have been freed : We then access q->slots[SFQ_EMPTY_SLOT], out of
> bounds, and crash in slot_queue_walk()
>
> On previous kernels, bug is here but out of bounds qs[SFQ_DEPTH] and
> allot[SFQ_DEPTH] are located in struct sfq_sched_data, so no illegal
> memory access happens, only possibly wrong data reported to user.
Yes, nice catch!
> Also, slot_dequeue_tail() should make sure slot skb chain is correctly
> terminated, or sfq_dump_class_stats() can access freed skbs.
...and a good hint for code reusing ;-)
Thanks,
Jarek P.
>
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
> Cc: Jarek Poplawski <jarkao2@gmail.com>
> ---
> net/sched/sch_sfq.c | 16 +++++++++++-----
> 1 files changed, 11 insertions(+), 5 deletions(-)
>
> diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c
> index 13322e8..6a2f88f 100644
> --- a/net/sched/sch_sfq.c
> +++ b/net/sched/sch_sfq.c
> @@ -281,6 +281,7 @@ static inline struct sk_buff *slot_dequeue_tail(struct sfq_slot *slot)
> struct sk_buff *skb = slot->skblist_prev;
>
> slot->skblist_prev = skb->prev;
> + skb->prev->next = (struct sk_buff *)slot;
> skb->next = skb->prev = NULL;
> return skb;
> }
> @@ -608,14 +609,19 @@ static int sfq_dump_class_stats(struct Qdisc *sch, unsigned long cl,
> struct gnet_dump *d)
> {
> struct sfq_sched_data *q = qdisc_priv(sch);
> - const struct sfq_slot *slot = &q->slots[q->ht[cl - 1]];
> - struct gnet_stats_queue qs = { .qlen = slot->qlen };
> - struct tc_sfq_xstats xstats = { .allot = slot->allot };
> + sfq_index idx = q->ht[cl - 1];
> + struct gnet_stats_queue qs = { 0 };
> + struct tc_sfq_xstats xstats = { 0 };
> struct sk_buff *skb;
>
> - slot_queue_walk(slot, skb)
> - qs.backlog += qdisc_pkt_len(skb);
> + if (idx != SFQ_EMPTY_SLOT) {
> + const struct sfq_slot *slot = &q->slots[idx];
>
> + xstats.allot = slot->allot;
> + qs.qlen = slot->qlen;
> + slot_queue_walk(slot, skb)
> + qs.backlog += qdisc_pkt_len(skb);
> + }
> if (gnet_stats_copy_queue(d, &qs) < 0)
> return -1;
> return gnet_stats_copy_app(d, &xstats, sizeof(xstats));
>
>
^ permalink raw reply
* [PATCH 04/10] bna: Enable pure priority tagged packet reception and rxf uninit cleanup fix
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Change details:
- Enable reception of pure priority tagged packets by default by turning on VLAN Id = 0
- Clear the promiscuous mode, all multicast mode flags when bna_rxf_uninit is called
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bna_txrx.c | 25 +++++++++++++++++++++++++
1 files changed, 25 insertions(+), 0 deletions(-)
diff --git a/drivers/net/bna/bna_txrx.c b/drivers/net/bna/bna_txrx.c
index fb6cf1f..5076618 100644
--- a/drivers/net/bna/bna_txrx.c
+++ b/drivers/net/bna/bna_txrx.c
@@ -1441,12 +1441,16 @@ bna_rxf_init(struct bna_rxf *rxf,
memset(rxf->vlan_filter_table, 0,
(sizeof(u32) * ((BFI_MAX_VLAN + 1) / 32)));
+ /* Set up VLAN 0 for pure priority tagged packets */
+ rxf->vlan_filter_table[0] |= 1;
+
bfa_fsm_set_state(rxf, bna_rxf_sm_stopped);
}
static void
bna_rxf_uninit(struct bna_rxf *rxf)
{
+ struct bna *bna = rxf->rx->bna;
struct bna_mac *mac;
bna_rit_mod_seg_put(&rxf->rx->bna->rit_mod, rxf->rit_segment);
@@ -1473,6 +1477,27 @@ bna_rxf_uninit(struct bna_rxf *rxf)
bna_mcam_mod_mac_put(&rxf->rx->bna->mcam_mod, mac);
}
+ /* Turn off pending promisc mode */
+ if (is_promisc_enable(rxf->rxmode_pending,
+ rxf->rxmode_pending_bitmask)) {
+ /* system promisc state should be pending */
+ BUG_ON(!(bna->rxf_promisc_id == rxf->rxf_id));
+ promisc_inactive(rxf->rxmode_pending,
+ rxf->rxmode_pending_bitmask);
+ bna->rxf_promisc_id = BFI_MAX_RXF;
+ }
+ /* Promisc mode should not be active */
+ BUG_ON(rxf->rxmode_active & BNA_RXMODE_PROMISC);
+
+ /* Turn off pending all-multi mode */
+ if (is_allmulti_enable(rxf->rxmode_pending,
+ rxf->rxmode_pending_bitmask)) {
+ allmulti_inactive(rxf->rxmode_pending,
+ rxf->rxmode_pending_bitmask);
+ }
+ /* Allmulti mode should not be active */
+ BUG_ON(rxf->rxmode_active & BNA_RXMODE_ALLMULTI);
+
rxf->rx = NULL;
}
--
1.7.1
^ permalink raw reply related
* [PATCH 08/10] bna: Restore VLAN filter table
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Change details:
- Retrieve the VLAN configuration from the networking stack and apply it to the base interface during ifconfig up
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bnad.c | 29 +++++++++++++++++++++++++++--
1 files changed, 27 insertions(+), 2 deletions(-)
diff --git a/drivers/net/bna/bnad.c b/drivers/net/bna/bnad.c
index 79d84f2..200df60 100644
--- a/drivers/net/bna/bnad.c
+++ b/drivers/net/bna/bnad.c
@@ -566,7 +566,8 @@ bnad_enable_rx_irq(struct bnad *bnad, struct bna_ccb *ccb)
{
unsigned long flags;
- spin_lock_irqsave(&bnad->bna_lock, flags); /* Because of polling context */
+ /* Because of polling context */
+ spin_lock_irqsave(&bnad->bna_lock, flags);
bnad_enable_rx_irq_unsafe(ccb);
spin_unlock_irqrestore(&bnad->bna_lock, flags);
}
@@ -1962,6 +1963,27 @@ bnad_enable_default_bcast(struct bnad *bnad)
return 0;
}
+/* Called with bnad_conf_lock() held */
+static void
+bnad_restore_vlans(struct bnad *bnad, u32 rx_id)
+{
+ u16 vlan_id;
+ unsigned long flags;
+
+ if (!bnad->vlan_grp)
+ return;
+
+ BUG_ON(!(VLAN_N_VID == (BFI_MAX_VLAN + 1)));
+
+ for (vlan_id = 0; vlan_id < VLAN_N_VID; vlan_id++) {
+ if (!vlan_group_get_device(bnad->vlan_grp, vlan_id))
+ continue;
+ spin_lock_irqsave(&bnad->bna_lock, flags);
+ bna_rx_vlan_add(bnad->rx_info[rx_id].rx, vlan_id);
+ spin_unlock_irqrestore(&bnad->bna_lock, flags);
+ }
+}
+
/* Statistics utilities */
void
bnad_netdev_qstats_fill(struct bnad *bnad, struct rtnl_link_stats64 *stats)
@@ -2337,6 +2359,9 @@ bnad_open(struct net_device *netdev)
/* Enable broadcast */
bnad_enable_default_bcast(bnad);
+ /* Restore VLANs, if any */
+ bnad_restore_vlans(bnad, 0);
+
/* Set the UCAST address */
spin_lock_irqsave(&bnad->bna_lock, flags);
bnad_mac_addr_set_locked(bnad, netdev->dev_addr);
@@ -3021,7 +3046,7 @@ static int __devinit
bnad_pci_probe(struct pci_dev *pdev,
const struct pci_device_id *pcidev_id)
{
- bool using_dac;
+ bool using_dac = false;
int err;
struct bnad *bnad;
struct bna *bna;
--
1.7.1
^ permalink raw reply related
* [PATCH 01/10] bna: TxRx and datapath fix
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Chnage Details:
- Check HW ready condition before accessing h/w register in data-path
- Postpone clean-up of data buffers to the data-path restart path and wait in the cleanup routines for in-flight DMA to complete
- Separate out Tx completion processing from Rx poll routine
- Macro for ia64 udelay
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bnad.c | 353 ++++++++++++++++++++++-------------------------
drivers/net/bna/bnad.h | 28 +++--
2 files changed, 184 insertions(+), 197 deletions(-)
diff --git a/drivers/net/bna/bnad.c b/drivers/net/bna/bnad.c
index 7e839b9..7bf7977 100644
--- a/drivers/net/bna/bnad.c
+++ b/drivers/net/bna/bnad.c
@@ -70,6 +70,8 @@ do { \
(sizeof(struct bnad_skb_unmap) * ((_depth) - 1)); \
} while (0)
+#define BNAD_TXRX_SYNC_UDELAY (1000 * 250) /* 250 msecs, unit in usecs */
+
/*
* Reinitialize completions in CQ, once Rx is taken down
*/
@@ -130,7 +132,9 @@ bnad_free_all_txbufs(struct bnad *bnad,
PCI_DMA_TODEVICE);
pci_unmap_addr_set(&unmap_array[unmap_cons], dma_addr, 0);
- unmap_cons++;
+ if (++unmap_cons >= unmap_q->q_depth)
+ break;
+
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
pci_unmap_page(bnad->pcidev,
pci_unmap_addr(&unmap_array[unmap_cons],
@@ -139,7 +143,8 @@ bnad_free_all_txbufs(struct bnad *bnad,
PCI_DMA_TODEVICE);
pci_unmap_addr_set(&unmap_array[unmap_cons], dma_addr,
0);
- unmap_cons++;
+ if (++unmap_cons >= unmap_q->q_depth)
+ break;
}
dev_kfree_skb_any(skb);
}
@@ -167,11 +172,11 @@ bnad_free_txbufs(struct bnad *bnad,
/*
* Just return if TX is stopped. This check is useful
* when bnad_free_txbufs() runs out of a tasklet scheduled
- * before bnad_cb_tx_cleanup() cleared BNAD_RF_TX_STARTED bit
+ * before bnad_cb_tx_cleanup() cleared BNAD_TXQ_TX_STARTED bit
* but this routine runs actually after the cleanup has been
* executed.
*/
- if (!test_bit(BNAD_RF_TX_STARTED, &bnad->run_flags))
+ if (!test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags))
return 0;
updated_hw_cons = *(tcb->hw_consumer_index);
@@ -252,7 +257,9 @@ bnad_tx_free_tasklet(unsigned long bnad_ptr)
(!test_and_set_bit(BNAD_TXQ_FREE_SENT,
&tcb->flags))) {
acked = bnad_free_txbufs(bnad, tcb);
- bna_ib_ack(tcb->i_dbell, acked);
+ if (likely(test_bit(BNAD_TXQ_TX_STARTED,
+ &tcb->flags)))
+ bna_ib_ack(tcb->i_dbell, acked);
smp_mb__before_clear_bit();
clear_bit(BNAD_TXQ_FREE_SENT, &tcb->flags);
}
@@ -264,7 +271,7 @@ static u32
bnad_tx(struct bnad *bnad, struct bna_tcb *tcb)
{
struct net_device *netdev = bnad->netdev;
- u32 sent;
+ u32 sent = 0;
if (test_and_set_bit(BNAD_TXQ_FREE_SENT, &tcb->flags))
return 0;
@@ -275,12 +282,15 @@ bnad_tx(struct bnad *bnad, struct bna_tcb *tcb)
netif_carrier_ok(netdev) &&
BNA_QE_FREE_CNT(tcb, tcb->q_depth) >=
BNAD_NETIF_WAKE_THRESHOLD) {
- netif_wake_queue(netdev);
- BNAD_UPDATE_CTR(bnad, netif_queue_wakeup);
+ if (test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags)) {
+ netif_wake_queue(netdev);
+ BNAD_UPDATE_CTR(bnad, netif_queue_wakeup);
+ }
}
+ }
+
+ if (likely(test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags)))
bna_ib_ack(tcb->i_dbell, sent);
- } else
- bna_ib_ack(tcb->i_dbell, 0);
smp_mb__before_clear_bit();
clear_bit(BNAD_TXQ_FREE_SENT, &tcb->flags);
@@ -313,25 +323,26 @@ bnad_reset_rcb(struct bnad *bnad, struct bna_rcb *rcb)
}
static void
-bnad_free_rxbufs(struct bnad *bnad, struct bna_rcb *rcb)
+bnad_free_all_rxbufs(struct bnad *bnad, struct bna_rcb *rcb)
{
struct bnad_unmap_q *unmap_q;
struct sk_buff *skb;
+ int unmap_cons;
unmap_q = rcb->unmap_q;
- while (BNA_QE_IN_USE_CNT(unmap_q, unmap_q->q_depth)) {
- skb = unmap_q->unmap_array[unmap_q->consumer_index].skb;
- BUG_ON(!(skb));
- unmap_q->unmap_array[unmap_q->consumer_index].skb = NULL;
+ for (unmap_cons = 0; unmap_cons < unmap_q->q_depth; unmap_cons++) {
+ skb = unmap_q->unmap_array[unmap_cons].skb;
+ if (!skb)
+ continue;
+ BUG_ON(!(pci_unmap_addr(
+ &unmap_q->unmap_array[unmap_cons], dma_addr)));
+ unmap_q->unmap_array[unmap_cons].skb = NULL;
pci_unmap_single(bnad->pcidev, pci_unmap_addr(&unmap_q->
- unmap_array[unmap_q->consumer_index],
- dma_addr), rcb->rxq->buffer_size +
- NET_IP_ALIGN, PCI_DMA_FROMDEVICE);
+ unmap_array[unmap_cons],
+ dma_addr), rcb->rxq->buffer_size,
+ PCI_DMA_FROMDEVICE);
dev_kfree_skb(skb);
- BNA_QE_INDX_ADD(unmap_q->consumer_index, 1, unmap_q->q_depth);
- BNA_QE_INDX_ADD(rcb->consumer_index, 1, rcb->q_depth);
}
-
bnad_reset_rcb(bnad, rcb);
}
@@ -385,41 +396,9 @@ finishing:
unmap_q->producer_index = unmap_prod;
rcb->producer_index = unmap_prod;
smp_mb();
- bna_rxq_prod_indx_doorbell(rcb);
- }
-}
-
-/*
- * Locking is required in the enable path
- * because it is called from a napi poll
- * context, where the bna_lock is not held
- * unlike the IRQ context.
- */
-static void
-bnad_enable_txrx_irqs(struct bnad *bnad)
-{
- struct bna_tcb *tcb;
- struct bna_ccb *ccb;
- int i, j;
- unsigned long flags;
-
- spin_lock_irqsave(&bnad->bna_lock, flags);
- for (i = 0; i < bnad->num_tx; i++) {
- for (j = 0; j < bnad->num_txq_per_tx; j++) {
- tcb = bnad->tx_info[i].tcb[j];
- bna_ib_coalescing_timer_set(tcb->i_dbell,
- tcb->txq->ib->ib_config.coalescing_timeo);
- bna_ib_ack(tcb->i_dbell, 0);
- }
- }
-
- for (i = 0; i < bnad->num_rx; i++) {
- for (j = 0; j < bnad->num_rxp_per_rx; j++) {
- ccb = bnad->rx_info[i].rx_ctrl[j].ccb;
- bnad_enable_rx_irq_unsafe(ccb);
- }
+ if (likely(test_bit(BNAD_RXQ_STARTED, &rcb->flags)))
+ bna_rxq_prod_indx_doorbell(rcb);
}
- spin_unlock_irqrestore(&bnad->bna_lock, flags);
}
static inline void
@@ -448,6 +427,9 @@ bnad_poll_cq(struct bnad *bnad, struct bna_ccb *ccb, int budget)
u32 qid0 = ccb->rcb[0]->rxq->rxq_id;
struct bna_pkt_rate *pkt_rt = &ccb->pkt_rate;
+ if (!test_bit(BNAD_RXQ_STARTED, &ccb->rcb[0]->flags))
+ return 0;
+
prefetch(bnad->netdev);
BNA_CQ_QPGE_PTR_GET(ccb->producer_index, ccb->sw_qpt, cmpl,
wi_range);
@@ -544,12 +526,15 @@ next:
BNA_QE_INDX_ADD(ccb->producer_index, wis, ccb->q_depth);
if (likely(ccb)) {
- bna_ib_ack(ccb->i_dbell, packets);
+ if (likely(test_bit(BNAD_RXQ_STARTED, &ccb->rcb[0]->flags)))
+ bna_ib_ack(ccb->i_dbell, packets);
bnad_refill_rxq(bnad, ccb->rcb[0]);
if (ccb->rcb[1])
bnad_refill_rxq(bnad, ccb->rcb[1]);
- } else
- bna_ib_ack(ccb->i_dbell, 0);
+ } else {
+ if (likely(test_bit(BNAD_RXQ_STARTED, &ccb->rcb[0]->flags)))
+ bna_ib_ack(ccb->i_dbell, 0);
+ }
return packets;
}
@@ -557,6 +542,9 @@ next:
static void
bnad_disable_rx_irq(struct bnad *bnad, struct bna_ccb *ccb)
{
+ if (unlikely(!test_bit(BNAD_RXQ_STARTED, &ccb->rcb[0]->flags)))
+ return;
+
bna_ib_coalescing_timer_set(ccb->i_dbell, 0);
bna_ib_ack(ccb->i_dbell, 0);
}
@@ -575,9 +563,11 @@ static void
bnad_netif_rx_schedule_poll(struct bnad *bnad, struct bna_ccb *ccb)
{
struct bnad_rx_ctrl *rx_ctrl = (struct bnad_rx_ctrl *)(ccb->ctrl);
- if (likely(napi_schedule_prep((&rx_ctrl->napi)))) {
+ struct napi_struct *napi = &rx_ctrl->napi;
+
+ if (likely(napi_schedule_prep(napi))) {
bnad_disable_rx_irq(bnad, ccb);
- __napi_schedule((&rx_ctrl->napi));
+ __napi_schedule(napi);
}
BNAD_UPDATE_CTR(bnad, netif_rx_schedule);
}
@@ -602,12 +592,11 @@ bnad_msix_mbox_handler(int irq, void *data)
{
u32 intr_status;
unsigned long flags;
- struct net_device *netdev = data;
- struct bnad *bnad;
+ struct bnad *bnad = (struct bnad *)data;
- bnad = netdev_priv(netdev);
+ if (unlikely(test_bit(BNAD_RF_MBOX_IRQ_DISABLED, &bnad->run_flags)))
+ return IRQ_HANDLED;
- /* BNA_ISR_GET(bnad); Inc Ref count */
spin_lock_irqsave(&bnad->bna_lock, flags);
bna_intr_status_get(&bnad->bna, intr_status);
@@ -617,7 +606,6 @@ bnad_msix_mbox_handler(int irq, void *data)
spin_unlock_irqrestore(&bnad->bna_lock, flags);
- /* BNAD_ISR_PUT(bnad); Dec Ref count */
return IRQ_HANDLED;
}
@@ -627,8 +615,7 @@ bnad_isr(int irq, void *data)
int i, j;
u32 intr_status;
unsigned long flags;
- struct net_device *netdev = data;
- struct bnad *bnad = netdev_priv(netdev);
+ struct bnad *bnad = (struct bnad *)data;
struct bnad_rx_info *rx_info;
struct bnad_rx_ctrl *rx_ctrl;
@@ -642,16 +629,21 @@ bnad_isr(int irq, void *data)
spin_lock_irqsave(&bnad->bna_lock, flags);
- if (BNA_IS_MBOX_ERR_INTR(intr_status)) {
+ if (BNA_IS_MBOX_ERR_INTR(intr_status))
bna_mbox_handler(&bnad->bna, intr_status);
- if (!BNA_IS_INTX_DATA_INTR(intr_status)) {
- spin_unlock_irqrestore(&bnad->bna_lock, flags);
- goto done;
- }
- }
+
spin_unlock_irqrestore(&bnad->bna_lock, flags);
+ if (!BNA_IS_INTX_DATA_INTR(intr_status))
+ return IRQ_HANDLED;
+
/* Process data interrupts */
+ /* Tx processing */
+ for (i = 0; i < bnad->num_tx; i++) {
+ for (j = 0; j < bnad->num_txq_per_tx; j++)
+ bnad_tx(bnad, bnad->tx_info[i].tcb[j]);
+ }
+ /* Rx processing */
for (i = 0; i < bnad->num_rx; i++) {
rx_info = &bnad->rx_info[i];
if (!rx_info->rx)
@@ -663,7 +655,6 @@ bnad_isr(int irq, void *data)
rx_ctrl->ccb);
}
}
-done:
return IRQ_HANDLED;
}
@@ -674,11 +665,7 @@ done:
static void
bnad_enable_mbox_irq(struct bnad *bnad)
{
- int irq = BNAD_GET_MBOX_IRQ(bnad);
-
- if (test_and_clear_bit(BNAD_RF_MBOX_IRQ_DISABLED, &bnad->run_flags))
- if (bnad->cfg_flags & BNAD_CF_MSIX)
- enable_irq(irq);
+ clear_bit(BNAD_RF_MBOX_IRQ_DISABLED, &bnad->run_flags);
BNAD_UPDATE_CTR(bnad, mbox_intr_enabled);
}
@@ -690,14 +677,19 @@ bnad_enable_mbox_irq(struct bnad *bnad)
static void
bnad_disable_mbox_irq(struct bnad *bnad)
{
- int irq = BNAD_GET_MBOX_IRQ(bnad);
+ set_bit(BNAD_RF_MBOX_IRQ_DISABLED, &bnad->run_flags);
+ BNAD_UPDATE_CTR(bnad, mbox_intr_disabled);
+}
- if (!test_and_set_bit(BNAD_RF_MBOX_IRQ_DISABLED, &bnad->run_flags))
- if (bnad->cfg_flags & BNAD_CF_MSIX)
- disable_irq_nosync(irq);
+static void
+bnad_set_netdev_perm_addr(struct bnad *bnad)
+{
+ struct net_device *netdev = bnad->netdev;
- BNAD_UPDATE_CTR(bnad, mbox_intr_disabled);
+ memcpy(netdev->perm_addr, &bnad->perm_addr, netdev->addr_len);
+ if (is_zero_ether_addr(netdev->dev_addr))
+ memcpy(netdev->dev_addr, &bnad->perm_addr, netdev->addr_len);
}
/* Control Path Handlers */
@@ -755,11 +747,14 @@ bnad_cb_port_link_status(struct bnad *bnad,
if (link_up) {
if (!netif_carrier_ok(bnad->netdev)) {
+ struct bna_tcb *tcb = bnad->tx_info[0].tcb[0];
+ if (!tcb)
+ return;
pr_warn("bna: %s link up\n",
bnad->netdev->name);
netif_carrier_on(bnad->netdev);
BNAD_UPDATE_CTR(bnad, link_toggle);
- if (test_bit(BNAD_RF_TX_STARTED, &bnad->run_flags)) {
+ if (test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags)) {
/* Force an immediate Transmit Schedule */
pr_info("bna: %s TX_STARTED\n",
bnad->netdev->name);
@@ -807,6 +802,18 @@ bnad_cb_tcb_destroy(struct bnad *bnad, struct bna_tcb *tcb)
{
struct bnad_tx_info *tx_info =
(struct bnad_tx_info *)tcb->txq->tx->priv;
+ struct bnad_unmap_q *unmap_q = tcb->unmap_q;
+
+ while (test_and_set_bit(BNAD_TXQ_FREE_SENT, &tcb->flags))
+ cpu_relax();
+
+ bnad_free_all_txbufs(bnad, tcb);
+
+ unmap_q->producer_index = 0;
+ unmap_q->consumer_index = 0;
+
+ smp_mb__before_clear_bit();
+ clear_bit(BNAD_TXQ_FREE_SENT, &tcb->flags);
tx_info->tcb[tcb->id] = NULL;
}
@@ -822,6 +829,12 @@ bnad_cb_rcb_setup(struct bnad *bnad, struct bna_rcb *rcb)
}
static void
+bnad_cb_rcb_destroy(struct bnad *bnad, struct bna_rcb *rcb)
+{
+ bnad_free_all_rxbufs(bnad, rcb);
+}
+
+static void
bnad_cb_ccb_setup(struct bnad *bnad, struct bna_ccb *ccb)
{
struct bnad_rx_info *rx_info =
@@ -849,7 +862,7 @@ bnad_cb_tx_stall(struct bnad *bnad, struct bna_tcb *tcb)
if (tx_info != &bnad->tx_info[0])
return;
- clear_bit(BNAD_RF_TX_STARTED, &bnad->run_flags);
+ clear_bit(BNAD_TXQ_TX_STARTED, &tcb->flags);
netif_stop_queue(bnad->netdev);
pr_info("bna: %s TX_STOPPED\n", bnad->netdev->name);
}
@@ -857,30 +870,15 @@ bnad_cb_tx_stall(struct bnad *bnad, struct bna_tcb *tcb)
static void
bnad_cb_tx_resume(struct bnad *bnad, struct bna_tcb *tcb)
{
- if (test_and_set_bit(BNAD_RF_TX_STARTED, &bnad->run_flags))
- return;
-
- if (netif_carrier_ok(bnad->netdev)) {
- pr_info("bna: %s TX_STARTED\n", bnad->netdev->name);
- netif_wake_queue(bnad->netdev);
- BNAD_UPDATE_CTR(bnad, netif_queue_wakeup);
- }
-}
-
-static void
-bnad_cb_tx_cleanup(struct bnad *bnad, struct bna_tcb *tcb)
-{
- struct bnad_unmap_q *unmap_q;
+ struct bnad_unmap_q *unmap_q = tcb->unmap_q;
- if (!tcb || (!tcb->unmap_q))
+ if (test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags))
return;
- unmap_q = tcb->unmap_q;
- if (!unmap_q->unmap_array)
- return;
+ clear_bit(BNAD_RF_TX_SHUTDOWN_DELAYED, &bnad->run_flags);
- if (test_and_set_bit(BNAD_TXQ_FREE_SENT, &tcb->flags))
- return;
+ while (test_and_set_bit(BNAD_TXQ_FREE_SENT, &tcb->flags))
+ cpu_relax();
bnad_free_all_txbufs(bnad, tcb);
@@ -889,21 +887,45 @@ bnad_cb_tx_cleanup(struct bnad *bnad, struct bna_tcb *tcb)
smp_mb__before_clear_bit();
clear_bit(BNAD_TXQ_FREE_SENT, &tcb->flags);
+
+ /*
+ * Workaround for first device enable failure & we
+ * get a 0 MAC address. We try to get the MAC address
+ * again here.
+ */
+ if (is_zero_ether_addr(&bnad->perm_addr.mac[0])) {
+ bna_port_mac_get(&bnad->bna.port, &bnad->perm_addr);
+ bnad_set_netdev_perm_addr(bnad);
+ }
+
+ set_bit(BNAD_TXQ_TX_STARTED, &tcb->flags);
+
+ if (netif_carrier_ok(bnad->netdev)) {
+ pr_info("bna: %s TX_STARTED\n", bnad->netdev->name);
+ netif_wake_queue(bnad->netdev);
+ BNAD_UPDATE_CTR(bnad, netif_queue_wakeup);
+ }
+}
+
+static void
+bnad_cb_tx_cleanup(struct bnad *bnad, struct bna_tcb *tcb)
+{
+ /* Delay only once for the whole Tx Path Shutdown */
+ if (!test_and_set_bit(BNAD_RF_TX_SHUTDOWN_DELAYED, &bnad->run_flags))
+ bnad_udelay(BNAD_TXRX_SYNC_UDELAY);
}
static void
bnad_cb_rx_cleanup(struct bnad *bnad,
struct bna_ccb *ccb)
{
- bnad_cq_cmpl_init(bnad, ccb);
-
- bnad_free_rxbufs(bnad, ccb->rcb[0]);
clear_bit(BNAD_RXQ_STARTED, &ccb->rcb[0]->flags);
- if (ccb->rcb[1]) {
- bnad_free_rxbufs(bnad, ccb->rcb[1]);
+ if (ccb->rcb[1])
clear_bit(BNAD_RXQ_STARTED, &ccb->rcb[1]->flags);
- }
+
+ if (!test_and_set_bit(BNAD_RF_RX_SHUTDOWN_DELAYED, &bnad->run_flags))
+ bnad_udelay(BNAD_TXRX_SYNC_UDELAY);
}
static void
@@ -911,6 +933,13 @@ bnad_cb_rx_post(struct bnad *bnad, struct bna_rcb *rcb)
{
struct bnad_unmap_q *unmap_q = rcb->unmap_q;
+ clear_bit(BNAD_RF_RX_SHUTDOWN_DELAYED, &bnad->run_flags);
+
+ if (rcb == rcb->cq->ccb->rcb[0])
+ bnad_cq_cmpl_init(bnad, rcb->cq->ccb);
+
+ bnad_free_all_rxbufs(bnad, rcb);
+
set_bit(BNAD_RXQ_STARTED, &rcb->flags);
/* Now allocate & post buffers for this RCB */
@@ -1047,7 +1076,7 @@ bnad_mbox_irq_free(struct bnad *bnad,
spin_unlock_irqrestore(&bnad->bna_lock, flags);
irq = BNAD_GET_MBOX_IRQ(bnad);
- free_irq(irq, bnad->netdev);
+ free_irq(irq, bnad);
kfree(intr_info->idl);
}
@@ -1061,7 +1090,7 @@ static int
bnad_mbox_irq_alloc(struct bnad *bnad,
struct bna_intr_info *intr_info)
{
- int err;
+ int err = 0;
unsigned long flags;
u32 irq;
irq_handler_t irq_handler;
@@ -1096,22 +1125,17 @@ bnad_mbox_irq_alloc(struct bnad *bnad,
*/
set_bit(BNAD_RF_MBOX_IRQ_DISABLED, &bnad->run_flags);
+ BNAD_UPDATE_CTR(bnad, mbox_intr_disabled);
+
err = request_irq(irq, irq_handler, flags,
- bnad->mbox_irq_name, bnad->netdev);
+ bnad->mbox_irq_name, bnad);
if (err) {
kfree(intr_info->idl);
intr_info->idl = NULL;
- return err;
}
- spin_lock_irqsave(&bnad->bna_lock, flags);
-
- if (bnad->cfg_flags & BNAD_CF_MSIX)
- disable_irq_nosync(irq);
-
- spin_unlock_irqrestore(&bnad->bna_lock, flags);
- return 0;
+ return err;
}
static void
@@ -1555,62 +1579,19 @@ poll_exit:
return rcvd;
}
-static int
-bnad_napi_poll_txrx(struct napi_struct *napi, int budget)
-{
- struct bnad_rx_ctrl *rx_ctrl =
- container_of(napi, struct bnad_rx_ctrl, napi);
- struct bna_ccb *ccb;
- struct bnad *bnad;
- int rcvd = 0;
- int i, j;
-
- ccb = rx_ctrl->ccb;
-
- bnad = ccb->bnad;
-
- if (!netif_carrier_ok(bnad->netdev))
- goto poll_exit;
-
- /* Handle Tx Completions, if any */
- for (i = 0; i < bnad->num_tx; i++) {
- for (j = 0; j < bnad->num_txq_per_tx; j++)
- bnad_tx(bnad, bnad->tx_info[i].tcb[j]);
- }
-
- /* Handle Rx Completions */
- rcvd = bnad_poll_cq(bnad, ccb, budget);
- if (rcvd == budget)
- return rcvd;
-poll_exit:
- napi_complete((napi));
-
- BNAD_UPDATE_CTR(bnad, netif_rx_complete);
-
- bnad_enable_txrx_irqs(bnad);
- return rcvd;
-}
-
static void
bnad_napi_enable(struct bnad *bnad, u32 rx_id)
{
- int (*napi_poll) (struct napi_struct *, int);
struct bnad_rx_ctrl *rx_ctrl;
int i;
- unsigned long flags;
-
- spin_lock_irqsave(&bnad->bna_lock, flags);
- if (bnad->cfg_flags & BNAD_CF_MSIX)
- napi_poll = bnad_napi_poll_rx;
- else
- napi_poll = bnad_napi_poll_txrx;
- spin_unlock_irqrestore(&bnad->bna_lock, flags);
/* Initialize & enable NAPI */
for (i = 0; i < bnad->num_rxp_per_rx; i++) {
rx_ctrl = &bnad->rx_info[rx_id].rx_ctrl[i];
+
netif_napi_add(bnad->netdev, &rx_ctrl->napi,
- napi_poll, 64);
+ bnad_napi_poll_rx, 64);
+
napi_enable(&rx_ctrl->napi);
}
}
@@ -1825,6 +1806,7 @@ bnad_setup_rx(struct bnad *bnad, uint rx_id)
/* Initialize the Rx event handlers */
rx_cbfn.rcb_setup_cbfn = bnad_cb_rcb_setup;
+ rx_cbfn.rcb_destroy_cbfn = bnad_cb_rcb_destroy;
rx_cbfn.rcb_destroy_cbfn = NULL;
rx_cbfn.ccb_setup_cbfn = bnad_cb_ccb_setup;
rx_cbfn.ccb_destroy_cbfn = bnad_cb_ccb_destroy;
@@ -2152,16 +2134,6 @@ bnad_q_num_adjust(struct bnad *bnad, int msix_vectors)
bnad->num_rxp_per_rx = 1;
}
-static void
-bnad_set_netdev_perm_addr(struct bnad *bnad)
-{
- struct net_device *netdev = bnad->netdev;
-
- memcpy(netdev->perm_addr, &bnad->perm_addr, netdev->addr_len);
- if (is_zero_ether_addr(netdev->dev_addr))
- memcpy(netdev->dev_addr, &bnad->perm_addr, netdev->addr_len);
-}
-
/* Enable / disable device */
static void
bnad_device_disable(struct bnad *bnad)
@@ -2433,21 +2405,21 @@ bnad_start_xmit(struct sk_buff *skb, struct net_device *netdev)
return NETDEV_TX_OK;
}
+ tx_id = 0;
+
+ tx_info = &bnad->tx_info[tx_id];
+ tcb = tx_info->tcb[tx_id];
+ unmap_q = tcb->unmap_q;
+
/*
* Takes care of the Tx that is scheduled between clearing the flag
* and the netif_stop_queue() call.
*/
- if (unlikely(!test_bit(BNAD_RF_TX_STARTED, &bnad->run_flags))) {
+ if (unlikely(!test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags))) {
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
- tx_id = 0;
-
- tx_info = &bnad->tx_info[tx_id];
- tcb = tx_info->tcb[tx_id];
- unmap_q = tcb->unmap_q;
-
vectors = 1 + skb_shinfo(skb)->nr_frags;
if (vectors > BFI_TX_MAX_VECTORS_PER_PKT) {
dev_kfree_skb(skb);
@@ -2462,7 +2434,8 @@ bnad_start_xmit(struct sk_buff *skb, struct net_device *netdev)
tcb->consumer_index &&
!test_and_set_bit(BNAD_TXQ_FREE_SENT, &tcb->flags)) {
acked = bnad_free_txbufs(bnad, tcb);
- bna_ib_ack(tcb->i_dbell, acked);
+ if (likely(test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags)))
+ bna_ib_ack(tcb->i_dbell, acked);
smp_mb__before_clear_bit();
clear_bit(BNAD_TXQ_FREE_SENT, &tcb->flags);
} else {
@@ -2624,6 +2597,10 @@ bnad_start_xmit(struct sk_buff *skb, struct net_device *netdev)
tcb->producer_index = txq_prod;
smp_mb();
+
+ if (unlikely(!test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags)))
+ return NETDEV_TX_OK;
+
bna_txq_prod_indx_doorbell(tcb);
if ((u16) (*tcb->hw_consumer_index) != tcb->consumer_index)
@@ -3066,7 +3043,7 @@ bnad_pci_probe(struct pci_dev *pdev,
/*
* PCI initialization
* Output : using_dac = 1 for 64 bit DMA
- * = 0 for 32 bit DMA
+ * = 0 for 32 bit DMA
*/
err = bnad_pci_init(bnad, pdev, &using_dac);
if (err)
diff --git a/drivers/net/bna/bnad.h b/drivers/net/bna/bnad.h
index ebc3a90..92768e2 100644
--- a/drivers/net/bna/bnad.h
+++ b/drivers/net/bna/bnad.h
@@ -51,6 +51,7 @@
*/
struct bnad_rx_ctrl {
struct bna_ccb *ccb;
+ unsigned long flags;
struct napi_struct napi;
};
@@ -82,6 +83,7 @@ struct bnad_rx_ctrl {
/* Bit positions for tcb->flags */
#define BNAD_TXQ_FREE_SENT 0
+#define BNAD_TXQ_TX_STARTED 1
/* Bit positions for rcb->flags */
#define BNAD_RXQ_REFILL 0
@@ -199,12 +201,12 @@ struct bnad_unmap_q {
/* Set, tested & cleared using xxx_bit() functions */
/* Values indicated bit positions */
#define BNAD_RF_CEE_RUNNING 1
-#define BNAD_RF_HW_ERROR 2
-#define BNAD_RF_MBOX_IRQ_DISABLED 3
-#define BNAD_RF_TX_STARTED 4
-#define BNAD_RF_RX_STARTED 5
-#define BNAD_RF_DIM_TIMER_RUNNING 6
-#define BNAD_RF_STATS_TIMER_RUNNING 7
+#define BNAD_RF_MBOX_IRQ_DISABLED 2
+#define BNAD_RF_RX_STARTED 3
+#define BNAD_RF_DIM_TIMER_RUNNING 4
+#define BNAD_RF_STATS_TIMER_RUNNING 5
+#define BNAD_RF_TX_SHUTDOWN_DELAYED 6
+#define BNAD_RF_RX_SHUTDOWN_DELAYED 7
struct bnad {
struct net_device *netdev;
@@ -320,13 +322,21 @@ extern void bnad_netdev_hwstats_fill(struct bnad *bnad, struct rtnl_link_stats64
#define bnad_enable_rx_irq_unsafe(_ccb) \
{ \
- bna_ib_coalescing_timer_set((_ccb)->i_dbell, \
- (_ccb)->rx_coalescing_timeo); \
- bna_ib_ack((_ccb)->i_dbell, 0); \
+ if (likely(test_bit(BNAD_RXQ_STARTED, &ccb->rcb[0]->flags))) {\
+ bna_ib_coalescing_timer_set((_ccb)->i_dbell, \
+ (_ccb)->rx_coalescing_timeo); \
+ bna_ib_ack((_ccb)->i_dbell, 0); \
+ } \
}
#define bnad_dim_timer_running(_bnad) \
(((_bnad)->cfg_flags & BNAD_CF_DIM_ENABLED) && \
(test_bit(BNAD_RF_DIM_TIMER_RUNNING, &((_bnad)->run_flags))))
+#if defined(__ia64__)
+#define bnad_udelay udelay
+#else
+#define bnad_udelay __udelay
+#endif
+
#endif /* __BNAD_H__ */
--
1.7.1
^ permalink raw reply related
* [PATCH 10/10] bna: Update the driver version to 2.3.2.3
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bnad.h | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/bna/bnad.h b/drivers/net/bna/bnad.h
index 06d9ea2..24fd983 100644
--- a/drivers/net/bna/bnad.h
+++ b/drivers/net/bna/bnad.h
@@ -65,7 +65,7 @@ struct bnad_rx_ctrl {
#define BNAD_NAME "bna"
#define BNAD_NAME_LEN 64
-#define BNAD_VERSION "2.3.2.0"
+#define BNAD_VERSION "2.3.2.3"
#define BNAD_MAILBOX_MSIX_VECTORS 1
--
1.7.1
^ permalink raw reply related
* [PATCH 03/10] bna: Fix ethtool register dump and reordered an API
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Change details:
- Removed semaphore register dump from ethtool
- Moved netif_carrier_off() call to before calling bna_init()
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bnad.c | 8 +++-----
drivers/net/bna/bnad_ethtool.c | 4 ----
2 files changed, 3 insertions(+), 9 deletions(-)
diff --git a/drivers/net/bna/bnad.c b/drivers/net/bna/bnad.c
index 7bf7977..348ebf9 100644
--- a/drivers/net/bna/bnad.c
+++ b/drivers/net/bna/bnad.c
@@ -3061,6 +3061,9 @@ bnad_pci_probe(struct pci_dev *pdev,
/* Initialize netdev structure, set up ethtool ops */
bnad_netdev_init(bnad, using_dac);
+ /* Set link to down state */
+ netif_carrier_off(netdev);
+
bnad_enable_msix(bnad);
/* Get resource requirement form bna */
@@ -3114,11 +3117,6 @@ bnad_pci_probe(struct pci_dev *pdev,
mutex_unlock(&bnad->conf_mutex);
- /*
- * Make sure the link appears down to the stack
- */
- netif_carrier_off(netdev);
-
/* Finally, reguister with net_device layer */
err = register_netdev(netdev);
if (err) {
diff --git a/drivers/net/bna/bnad_ethtool.c b/drivers/net/bna/bnad_ethtool.c
index 11fa2ea..3011110 100644
--- a/drivers/net/bna/bnad_ethtool.c
+++ b/drivers/net/bna/bnad_ethtool.c
@@ -330,10 +330,6 @@ do { \
BNAD_GET_REG(PCIE_MISC_REG);
- BNAD_GET_REG(HOST_SEM0_REG);
- BNAD_GET_REG(HOST_SEM1_REG);
- BNAD_GET_REG(HOST_SEM2_REG);
- BNAD_GET_REG(HOST_SEM3_REG);
BNAD_GET_REG(HOST_SEM0_INFO_REG);
BNAD_GET_REG(HOST_SEM1_INFO_REG);
BNAD_GET_REG(HOST_SEM2_INFO_REG);
--
1.7.1
^ permalink raw reply related
* [PATCH 05/10] bna: Fix for TX queue
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Chnage details:
- Call netif_wake_queue() if we have freed up sufficient elements at the end of completion processing
- Add netif_queue_stopped counter back to bnad_drv_stats {}
- Get netif_queue_stopped value from stack
- Remove BUG_ON() on value returned by pci_unmap_addr()
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bnad.c | 20 ++++++++++++++++----
drivers/net/bna/bnad.h | 7 +++++--
drivers/net/bna/bnad_ethtool.c | 4 ++++
3 files changed, 25 insertions(+), 6 deletions(-)
diff --git a/drivers/net/bna/bnad.c b/drivers/net/bna/bnad.c
index 348ebf9..79d84f2 100644
--- a/drivers/net/bna/bnad.c
+++ b/drivers/net/bna/bnad.c
@@ -109,7 +109,7 @@ static void
bnad_free_all_txbufs(struct bnad *bnad,
struct bna_tcb *tcb)
{
- u16 unmap_cons;
+ u32 unmap_cons;
struct bnad_unmap_q *unmap_q = tcb->unmap_q;
struct bnad_skb_unmap *unmap_array;
struct sk_buff *skb = NULL;
@@ -244,7 +244,7 @@ bnad_tx_free_tasklet(unsigned long bnad_ptr)
{
struct bnad *bnad = (struct bnad *)bnad_ptr;
struct bna_tcb *tcb;
- u32 acked;
+ u32 acked = 0;
int i, j;
for (i = 0; i < bnad->num_tx; i++) {
@@ -263,6 +263,20 @@ bnad_tx_free_tasklet(unsigned long bnad_ptr)
smp_mb__before_clear_bit();
clear_bit(BNAD_TXQ_FREE_SENT, &tcb->flags);
}
+ if (unlikely(!test_bit(BNAD_TXQ_TX_STARTED,
+ &tcb->flags)))
+ continue;
+ if (netif_queue_stopped(bnad->netdev)) {
+ if (acked && netif_carrier_ok(bnad->netdev) &&
+ BNA_QE_FREE_CNT(tcb, tcb->q_depth) >=
+ BNAD_NETIF_WAKE_THRESHOLD) {
+ netif_wake_queue(bnad->netdev);
+ /* TODO */
+ /* Counters for individual TxQs? */
+ BNAD_UPDATE_CTR(bnad,
+ netif_queue_wakeup);
+ }
+ }
}
}
}
@@ -334,8 +348,6 @@ bnad_free_all_rxbufs(struct bnad *bnad, struct bna_rcb *rcb)
skb = unmap_q->unmap_array[unmap_cons].skb;
if (!skb)
continue;
- BUG_ON(!(pci_unmap_addr(
- &unmap_q->unmap_array[unmap_cons], dma_addr)));
unmap_q->unmap_array[unmap_cons].skb = NULL;
pci_unmap_single(bnad->pcidev, pci_unmap_addr(&unmap_q->
unmap_array[unmap_cons],
diff --git a/drivers/net/bna/bnad.h b/drivers/net/bna/bnad.h
index 92768e2..06d9ea2 100644
--- a/drivers/net/bna/bnad.h
+++ b/drivers/net/bna/bnad.h
@@ -126,6 +126,7 @@ struct bnad_completion {
struct bnad_drv_stats {
u64 netif_queue_stop;
u64 netif_queue_wakeup;
+ u64 netif_queue_stopped;
u64 tso4;
u64 tso6;
u64 tso_err;
@@ -308,8 +309,10 @@ extern void bnad_cleanup_rx(struct bnad *bnad, uint rx_id);
extern void bnad_dim_timer_start(struct bnad *bnad);
/* Statistics */
-extern void bnad_netdev_qstats_fill(struct bnad *bnad, struct rtnl_link_stats64 *stats);
-extern void bnad_netdev_hwstats_fill(struct bnad *bnad, struct rtnl_link_stats64 *stats);
+extern void bnad_netdev_qstats_fill(struct bnad *bnad,
+ struct rtnl_link_stats64 *stats);
+extern void bnad_netdev_hwstats_fill(struct bnad *bnad,
+ struct rtnl_link_stats64 *stats);
/**
* MACROS
diff --git a/drivers/net/bna/bnad_ethtool.c b/drivers/net/bna/bnad_ethtool.c
index 3011110..99be5ae 100644
--- a/drivers/net/bna/bnad_ethtool.c
+++ b/drivers/net/bna/bnad_ethtool.c
@@ -68,6 +68,7 @@ static char *bnad_net_stats_strings[BNAD_ETHTOOL_STATS_NUM] = {
"netif_queue_stop",
"netif_queue_wakeup",
+ "netif_queue_stopped",
"tso4",
"tso6",
"tso_err",
@@ -1180,6 +1181,9 @@ bnad_get_ethtool_stats(struct net_device *netdev, struct ethtool_stats *stats,
bi = sizeof(*net_stats64) / sizeof(u64);
+ /* Get netif_queue_stopped from stack */
+ bnad->stats.drv_stats.netif_queue_stopped = netif_queue_stopped(netdev);
+
/* Fill driver stats into ethtool buffers */
stats64 = (u64 *)&bnad->stats.drv_stats;
for (i = 0; i < sizeof(struct bnad_drv_stats) / sizeof(u64); i++)
--
1.7.1
^ permalink raw reply related
* [PATCH 06/10] bna: IOC uninit check and misc cleanup
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Chnage details:
- Added a check in ioc firmware lock function to see if IOC is in BFI_IOC_UNINIT state or not. If it is not in UNINIT
state and the last IOC boot was not done by OS driver, force IOC state to BFI_IOC_UNINIT
- Unused macro and API cleanup
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bfa_defs_mfg_comm.h | 22 -----------------
drivers/net/bna/bfa_ioc.c | 45 +---------------------------------
2 files changed, 2 insertions(+), 65 deletions(-)
diff --git a/drivers/net/bna/bfa_defs_mfg_comm.h b/drivers/net/bna/bfa_defs_mfg_comm.h
index 987978f..fdd6776 100644
--- a/drivers/net/bna/bfa_defs_mfg_comm.h
+++ b/drivers/net/bna/bfa_defs_mfg_comm.h
@@ -95,28 +95,6 @@ enum {
(type) == BFA_MFG_TYPE_CNA10P1 || \
bfa_mfg_is_mezz(type)))
-/**
- * Check if the card having old wwn/mac handling
- */
-#define bfa_mfg_is_old_wwn_mac_model(type) (( \
- (type) == BFA_MFG_TYPE_FC8P2 || \
- (type) == BFA_MFG_TYPE_FC8P1 || \
- (type) == BFA_MFG_TYPE_FC4P2 || \
- (type) == BFA_MFG_TYPE_FC4P1 || \
- (type) == BFA_MFG_TYPE_CNA10P2 || \
- (type) == BFA_MFG_TYPE_CNA10P1 || \
- (type) == BFA_MFG_TYPE_JAYHAWK || \
- (type) == BFA_MFG_TYPE_WANCHESE))
-
-#define bfa_mfg_increment_wwn_mac(m, i) \
-do { \
- u32 t = ((m)[0] << 16) | ((m)[1] << 8) | (m)[2]; \
- t += (i); \
- (m)[0] = (t >> 16) & 0xFF; \
- (m)[1] = (t >> 8) & 0xFF; \
- (m)[2] = t & 0xFF; \
-} while (0)
-
#define bfa_mfg_adapter_prop_init_flash(card_type, prop) \
do { \
switch ((card_type)) { \
diff --git a/drivers/net/bna/bfa_ioc.c b/drivers/net/bna/bfa_ioc.c
index e94e5aa..8ed147e 100644
--- a/drivers/net/bna/bfa_ioc.c
+++ b/drivers/net/bna/bfa_ioc.c
@@ -58,9 +58,6 @@
#define bfa_ioc_notify_hbfail(__ioc) \
((__ioc)->ioc_hwif->ioc_notify_hbfail(__ioc))
-#define bfa_ioc_is_optrom(__ioc) \
- (bfa_cb_image_get_size(BFA_IOC_FWIMG_TYPE(__ioc)) < BFA_IOC_FWIMG_MINSZ)
-
#define bfa_ioc_mbox_cmd_pending(__ioc) \
(!list_empty(&((__ioc)->mbox_mod.cmd_q)) || \
readl((__ioc)->ioc_regs.hfn_mbox_cmd))
@@ -101,7 +98,6 @@ static void bfa_ioc_get_adapter_manufacturer(struct bfa_ioc *ioc,
char *manufacturer);
static void bfa_ioc_get_adapter_model(struct bfa_ioc *ioc, char *model);
static u64 bfa_ioc_get_pwwn(struct bfa_ioc *ioc);
-static mac_t bfa_ioc_get_mfg_mac(struct bfa_ioc *ioc);
/**
* IOC state machine events
@@ -865,12 +861,6 @@ bfa_ioc_fwver_valid(struct bfa_ioc *ioc)
{
struct bfi_ioc_image_hdr fwhdr, *drv_fwhdr;
- /**
- * If bios/efi boot (flash based) -- return true
- */
- if (bfa_ioc_is_optrom(ioc))
- return true;
-
bfa_nw_ioc_fwver_get(ioc, &fwhdr);
drv_fwhdr = (struct bfi_ioc_image_hdr *)
bfa_cb_image_get_chunk(BFA_IOC_FWIMG_TYPE(ioc), 0);
@@ -934,13 +924,8 @@ bfa_ioc_hwinit(struct bfa_ioc *ioc, bool force)
/**
* If IOC function is disabled and firmware version is same,
* just re-enable IOC.
- *
- * If option rom, IOC must not be in operational state. With
- * convergence, IOC will be in operational state when 2nd driver
- * is loaded.
*/
- if (ioc_fwstate == BFI_IOC_DISABLED ||
- (!bfa_ioc_is_optrom(ioc) && ioc_fwstate == BFI_IOC_OP)) {
+ if (ioc_fwstate == BFI_IOC_DISABLED || ioc_fwstate == BFI_IOC_OP) {
/**
* When using MSI-X any pending firmware ready event should
* be flushed. Otherwise MSI-X interrupts are not delivered.
@@ -1078,11 +1063,6 @@ bfa_ioc_download_fw(struct bfa_ioc *ioc, u32 boot_type,
*/
bfa_ioc_lmem_init(ioc);
- /**
- * Flash based firmware boot
- */
- if (bfa_ioc_is_optrom(ioc))
- boot_type = BFI_BOOT_TYPE_FLASH;
fwimg = bfa_cb_image_get_chunk(BFA_IOC_FWIMG_TYPE(ioc), chunkno);
pgnum = bfa_ioc_smem_pgnum(ioc, loff);
@@ -1689,28 +1669,7 @@ bfa_ioc_get_pwwn(struct bfa_ioc *ioc)
mac_t
bfa_nw_ioc_get_mac(struct bfa_ioc *ioc)
{
- /*
- * Currently mfg mac is used as FCoE enode mac (not configured by PBC)
- */
- if (bfa_ioc_get_type(ioc) == BFA_IOC_TYPE_FCoE)
- return bfa_ioc_get_mfg_mac(ioc);
- else
- return ioc->attr->mac;
-}
-
-static mac_t
-bfa_ioc_get_mfg_mac(struct bfa_ioc *ioc)
-{
- mac_t m;
-
- m = ioc->attr->mfg_mac;
- if (bfa_mfg_is_old_wwn_mac_model(ioc->attr->card_type))
- m.mac[MAC_ADDRLEN - 1] += bfa_ioc_pcifn(ioc);
- else
- bfa_mfg_increment_wwn_mac(&(m.mac[MAC_ADDRLEN-3]),
- bfa_ioc_pcifn(ioc));
-
- return m;
+ return ioc->attr->mac;
}
/**
--
1.7.1
^ permalink raw reply related
* [PATCH 07/10] bna: Removed unused code
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Change details:
- Remove unused APIs and code cleanup
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bna_ctrl.c | 241 +------------------------------------------
drivers/net/bna/bna_txrx.c | 11 +--
drivers/net/bna/bna_types.h | 4 +-
3 files changed, 3 insertions(+), 253 deletions(-)
diff --git a/drivers/net/bna/bna_ctrl.c b/drivers/net/bna/bna_ctrl.c
index 68e4c5e..e152747 100644
--- a/drivers/net/bna/bna_ctrl.c
+++ b/drivers/net/bna/bna_ctrl.c
@@ -2132,37 +2132,6 @@ rxf_fltr_mbox_cmd(struct bna_rxf *rxf, u8 cmd, enum bna_status status)
bna_mbox_send(rxf->rx->bna, &rxf->mbox_qe);
}
-static void
-__rxf_default_function_config(struct bna_rxf *rxf, enum bna_status status)
-{
- struct bna_rx_fndb_ram *rx_fndb_ram;
- u32 ctrl_flags;
- int i;
-
- rx_fndb_ram = (struct bna_rx_fndb_ram *)
- BNA_GET_MEM_BASE_ADDR(rxf->rx->bna->pcidev.pci_bar_kva,
- RX_FNDB_RAM_BASE_OFFSET);
-
- for (i = 0; i < BFI_MAX_RXF; i++) {
- if (status == BNA_STATUS_T_ENABLED) {
- if (i == rxf->rxf_id)
- continue;
-
- ctrl_flags =
- readl(&rx_fndb_ram[i].control_flags);
- ctrl_flags |= BNA_RXF_CF_DEFAULT_FUNCTION_ENABLE;
- writel(ctrl_flags,
- &rx_fndb_ram[i].control_flags);
- } else {
- ctrl_flags =
- readl(&rx_fndb_ram[i].control_flags);
- ctrl_flags &= ~BNA_RXF_CF_DEFAULT_FUNCTION_ENABLE;
- writel(ctrl_flags,
- &rx_fndb_ram[i].control_flags);
- }
- }
-}
-
int
rxf_process_packet_filter_ucast(struct bna_rxf *rxf)
{
@@ -2229,46 +2198,6 @@ rxf_process_packet_filter_promisc(struct bna_rxf *rxf)
}
int
-rxf_process_packet_filter_default(struct bna_rxf *rxf)
-{
- struct bna *bna = rxf->rx->bna;
-
- /* Enable/disable default mode */
- if (is_default_enable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask)) {
- /* move default configuration from pending -> active */
- default_inactive(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask);
- rxf->rxmode_active |= BNA_RXMODE_DEFAULT;
-
- /* Disable VLAN filter to allow all VLANs */
- __rxf_vlan_filter_set(rxf, BNA_STATUS_T_DISABLED);
- /* Redirect all other RxF vlan filtering to this one */
- __rxf_default_function_config(rxf, BNA_STATUS_T_ENABLED);
- rxf_fltr_mbox_cmd(rxf, BFI_LL_H2I_RXF_DEFAULT_SET_REQ,
- BNA_STATUS_T_ENABLED);
- return 1;
- } else if (is_default_disable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask)) {
- /* move default configuration from pending -> active */
- default_inactive(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask);
- rxf->rxmode_active &= ~BNA_RXMODE_DEFAULT;
- bna->rxf_default_id = BFI_MAX_RXF;
-
- /* Revert VLAN filter */
- __rxf_vlan_filter_set(rxf, rxf->vlan_filter_status);
- /* Stop RxF vlan filter table redirection */
- __rxf_default_function_config(rxf, BNA_STATUS_T_DISABLED);
- rxf_fltr_mbox_cmd(rxf, BFI_LL_H2I_RXF_DEFAULT_SET_REQ,
- BNA_STATUS_T_DISABLED);
- return 1;
- }
-
- return 0;
-}
-
-int
rxf_process_packet_filter_allmulti(struct bna_rxf *rxf)
{
/* Enable/disable allmulti mode */
@@ -2365,48 +2294,6 @@ rxf_clear_packet_filter_promisc(struct bna_rxf *rxf)
}
int
-rxf_clear_packet_filter_default(struct bna_rxf *rxf)
-{
- struct bna *bna = rxf->rx->bna;
-
- /* 8. Execute pending default mode disable command */
- if (is_default_disable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask)) {
- /* move default configuration from pending -> active */
- default_inactive(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask);
- rxf->rxmode_active &= ~BNA_RXMODE_DEFAULT;
- bna->rxf_default_id = BFI_MAX_RXF;
-
- /* Revert VLAN filter */
- __rxf_vlan_filter_set(rxf, rxf->vlan_filter_status);
- /* Stop RxF vlan filter table redirection */
- __rxf_default_function_config(rxf, BNA_STATUS_T_DISABLED);
- rxf_fltr_mbox_cmd(rxf, BFI_LL_H2I_RXF_DEFAULT_SET_REQ,
- BNA_STATUS_T_DISABLED);
- return 1;
- }
-
- /* 9. Clear active default mode; move it to pending enable */
- if (rxf->rxmode_active & BNA_RXMODE_DEFAULT) {
- /* move default configuration from active -> pending */
- default_enable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask);
- rxf->rxmode_active &= ~BNA_RXMODE_DEFAULT;
-
- /* Revert VLAN filter */
- __rxf_vlan_filter_set(rxf, rxf->vlan_filter_status);
- /* Stop RxF vlan filter table redirection */
- __rxf_default_function_config(rxf, BNA_STATUS_T_DISABLED);
- rxf_fltr_mbox_cmd(rxf, BFI_LL_H2I_RXF_DEFAULT_SET_REQ,
- BNA_STATUS_T_DISABLED);
- return 1;
- }
-
- return 0;
-}
-
-int
rxf_clear_packet_filter_allmulti(struct bna_rxf *rxf)
{
/* 10. Execute pending allmulti mode disable command */
@@ -2481,28 +2368,6 @@ rxf_reset_packet_filter_promisc(struct bna_rxf *rxf)
}
void
-rxf_reset_packet_filter_default(struct bna_rxf *rxf)
-{
- struct bna *bna = rxf->rx->bna;
-
- /* 8. Clear pending default mode disable */
- if (is_default_disable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask)) {
- default_inactive(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask);
- rxf->rxmode_active &= ~BNA_RXMODE_DEFAULT;
- bna->rxf_default_id = BFI_MAX_RXF;
- }
-
- /* 9. Move default mode config from active -> pending */
- if (rxf->rxmode_active & BNA_RXMODE_DEFAULT) {
- default_enable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask);
- rxf->rxmode_active &= ~BNA_RXMODE_DEFAULT;
- }
-}
-
-void
rxf_reset_packet_filter_allmulti(struct bna_rxf *rxf)
{
/* 10. Clear pending allmulti mode disable */
@@ -2599,76 +2464,6 @@ rxf_promisc_disable(struct bna_rxf *rxf)
* 1 = need h/w change
*/
static int
-rxf_default_enable(struct bna_rxf *rxf)
-{
- struct bna *bna = rxf->rx->bna;
- int ret = 0;
-
- /* There can not be any pending disable command */
-
- /* Do nothing if pending enable or already enabled */
- if (is_default_enable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask) ||
- (rxf->rxmode_active & BNA_RXMODE_DEFAULT)) {
- /* Schedule enable */
- } else {
- /* Default mode should not be active in the system */
- default_enable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask);
- bna->rxf_default_id = rxf->rxf_id;
- ret = 1;
- }
-
- return ret;
-}
-
-/**
- * Should only be called by bna_rxf_mode_set.
- * Helps deciding if h/w configuration is needed or not.
- * Returns:
- * 0 = no h/w change
- * 1 = need h/w change
- */
-static int
-rxf_default_disable(struct bna_rxf *rxf)
-{
- struct bna *bna = rxf->rx->bna;
- int ret = 0;
-
- /* There can not be any pending disable */
-
- /* Turn off pending enable command , if any */
- if (is_default_enable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask)) {
- /* Promisc mode should not be active */
- /* system default state should be pending */
- default_inactive(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask);
- /* Remove the default state from the system */
- bna->rxf_default_id = BFI_MAX_RXF;
-
- /* Schedule disable */
- } else if (rxf->rxmode_active & BNA_RXMODE_DEFAULT) {
- /* Default mode should be active in the system */
- default_disable(rxf->rxmode_pending,
- rxf->rxmode_pending_bitmask);
- ret = 1;
-
- /* Do nothing if already disabled */
- } else {
- }
-
- return ret;
-}
-
-/**
- * Should only be called by bna_rxf_mode_set.
- * Helps deciding if h/w configuration is needed or not.
- * Returns:
- * 0 = no h/w change
- * 1 = need h/w change
- */
-static int
rxf_allmulti_enable(struct bna_rxf *rxf)
{
int ret = 0;
@@ -2730,38 +2525,13 @@ bna_rx_mode_set(struct bna_rx *rx, enum bna_rxmode new_mode,
struct bna_rxf *rxf = &rx->rxf;
int need_hw_config = 0;
- /* Error checks */
+ /* Process the commands */
if (is_promisc_enable(new_mode, bitmask)) {
/* If promisc mode is already enabled elsewhere in the system */
if ((rx->bna->rxf_promisc_id != BFI_MAX_RXF) &&
(rx->bna->rxf_promisc_id != rxf->rxf_id))
goto err_return;
-
- /* If default mode is already enabled in the system */
- if (rx->bna->rxf_default_id != BFI_MAX_RXF)
- goto err_return;
-
- /* Trying to enable promiscuous and default mode together */
- if (is_default_enable(new_mode, bitmask))
- goto err_return;
- }
-
- if (is_default_enable(new_mode, bitmask)) {
- /* If default mode is already enabled elsewhere in the system */
- if ((rx->bna->rxf_default_id != BFI_MAX_RXF) &&
- (rx->bna->rxf_default_id != rxf->rxf_id)) {
- goto err_return;
- }
-
- /* If promiscuous mode is already enabled in the system */
- if (rx->bna->rxf_promisc_id != BFI_MAX_RXF)
- goto err_return;
- }
-
- /* Process the commands */
-
- if (is_promisc_enable(new_mode, bitmask)) {
if (rxf_promisc_enable(rxf))
need_hw_config = 1;
} else if (is_promisc_disable(new_mode, bitmask)) {
@@ -2769,14 +2539,6 @@ bna_rx_mode_set(struct bna_rx *rx, enum bna_rxmode new_mode,
need_hw_config = 1;
}
- if (is_default_enable(new_mode, bitmask)) {
- if (rxf_default_enable(rxf))
- need_hw_config = 1;
- } else if (is_default_disable(new_mode, bitmask)) {
- if (rxf_default_disable(rxf))
- need_hw_config = 1;
- }
-
if (is_allmulti_enable(new_mode, bitmask)) {
if (rxf_allmulti_enable(rxf))
need_hw_config = 1;
@@ -3202,7 +2964,6 @@ bna_init(struct bna *bna, struct bnad *bnad, struct bfa_pcidev *pcidev,
bna_mcam_mod_init(&bna->mcam_mod, bna, res_info);
- bna->rxf_default_id = BFI_MAX_RXF;
bna->rxf_promisc_id = BFI_MAX_RXF;
/* Mbox q element for posting stat request to f/w */
diff --git a/drivers/net/bna/bna_txrx.c b/drivers/net/bna/bna_txrx.c
index 5076618..58c7664 100644
--- a/drivers/net/bna/bna_txrx.c
+++ b/drivers/net/bna/bna_txrx.c
@@ -1226,8 +1226,7 @@ rxf_process_packet_filter_vlan(struct bna_rxf *rxf)
/* Apply the VLAN filter */
if (rxf->rxf_flags & BNA_RXF_FL_VLAN_CONFIG_PENDING) {
rxf->rxf_flags &= ~BNA_RXF_FL_VLAN_CONFIG_PENDING;
- if (!(rxf->rxmode_active & BNA_RXMODE_PROMISC) &&
- !(rxf->rxmode_active & BNA_RXMODE_DEFAULT))
+ if (!(rxf->rxmode_active & BNA_RXMODE_PROMISC))
__rxf_vlan_filter_set(rxf, rxf->vlan_filter_status);
}
@@ -1276,9 +1275,6 @@ rxf_process_packet_filter(struct bna_rxf *rxf)
if (rxf_process_packet_filter_promisc(rxf))
return 1;
- if (rxf_process_packet_filter_default(rxf))
- return 1;
-
if (rxf_process_packet_filter_allmulti(rxf))
return 1;
@@ -1340,9 +1336,6 @@ rxf_clear_packet_filter(struct bna_rxf *rxf)
if (rxf_clear_packet_filter_promisc(rxf))
return 1;
- if (rxf_clear_packet_filter_default(rxf))
- return 1;
-
if (rxf_clear_packet_filter_allmulti(rxf))
return 1;
@@ -1389,8 +1382,6 @@ rxf_reset_packet_filter(struct bna_rxf *rxf)
rxf_reset_packet_filter_promisc(rxf);
- rxf_reset_packet_filter_default(rxf);
-
rxf_reset_packet_filter_allmulti(rxf);
}
diff --git a/drivers/net/bna/bna_types.h b/drivers/net/bna/bna_types.h
index d79fd98..b9c134f 100644
--- a/drivers/net/bna/bna_types.h
+++ b/drivers/net/bna/bna_types.h
@@ -165,8 +165,7 @@ enum bna_rxp_type {
enum bna_rxmode {
BNA_RXMODE_PROMISC = 1,
- BNA_RXMODE_DEFAULT = 2,
- BNA_RXMODE_ALLMULTI = 4
+ BNA_RXMODE_ALLMULTI = 2
};
enum bna_rx_event {
@@ -1118,7 +1117,6 @@ struct bna {
struct bna_rit_mod rit_mod;
- int rxf_default_id;
int rxf_promisc_id;
struct bna_mbox_qe mbox_qe;
--
1.7.1
^ permalink raw reply related
* [PATCH 09/10] bna: IOC failure auto recovery fix
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Change details:
- Made IOC auto_recovery synchronized and not timer based.
- Only one PCI function will attempt to recover and reinitialize the ASIC on a failure, that too after all the active
PCI functions acknowledge the IOC failure.
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bfa_defs.h | 22 +-
drivers/net/bna/bfa_ioc.c | 1174 ++++++++++++++++++++++++++++++++----------
drivers/net/bna/bfa_ioc.h | 49 +-
drivers/net/bna/bfa_ioc_ct.c | 102 ++++-
drivers/net/bna/bfi_ctreg.h | 41 +-
drivers/net/bna/bna.h | 2 -
drivers/net/bna/bnad.c | 21 +-
7 files changed, 1061 insertions(+), 350 deletions(-)
diff --git a/drivers/net/bna/bfa_defs.h b/drivers/net/bna/bfa_defs.h
index 29c1b8d..2ea0dfe 100644
--- a/drivers/net/bna/bfa_defs.h
+++ b/drivers/net/bna/bfa_defs.h
@@ -112,16 +112,18 @@ struct bfa_ioc_pci_attr {
* IOC states
*/
enum bfa_ioc_state {
- BFA_IOC_RESET = 1, /*!< IOC is in reset state */
- BFA_IOC_SEMWAIT = 2, /*!< Waiting for IOC h/w semaphore */
- BFA_IOC_HWINIT = 3, /*!< IOC h/w is being initialized */
- BFA_IOC_GETATTR = 4, /*!< IOC is being configured */
- BFA_IOC_OPERATIONAL = 5, /*!< IOC is operational */
- BFA_IOC_INITFAIL = 6, /*!< IOC hardware failure */
- BFA_IOC_HBFAIL = 7, /*!< IOC heart-beat failure */
- BFA_IOC_DISABLING = 8, /*!< IOC is being disabled */
- BFA_IOC_DISABLED = 9, /*!< IOC is disabled */
- BFA_IOC_FWMISMATCH = 10, /*!< IOC f/w different from drivers */
+ BFA_IOC_UNINIT = 1, /*!< IOC is in uninit state */
+ BFA_IOC_RESET = 2, /*!< IOC is in reset state */
+ BFA_IOC_SEMWAIT = 3, /*!< Waiting for IOC h/w semaphore */
+ BFA_IOC_HWINIT = 4, /*!< IOC h/w is being initialized */
+ BFA_IOC_GETATTR = 5, /*!< IOC is being configured */
+ BFA_IOC_OPERATIONAL = 6, /*!< IOC is operational */
+ BFA_IOC_INITFAIL = 7, /*!< IOC hardware failure */
+ BFA_IOC_FAIL = 8, /*!< IOC heart-beat failure */
+ BFA_IOC_DISABLING = 9, /*!< IOC is being disabled */
+ BFA_IOC_DISABLED = 10, /*!< IOC is disabled */
+ BFA_IOC_FWMISMATCH = 11, /*!< IOC f/w different from drivers */
+ BFA_IOC_ENABLING = 12, /*!< IOC is being enabled */
};
/**
diff --git a/drivers/net/bna/bfa_ioc.c b/drivers/net/bna/bfa_ioc.c
index 8ed147e..34933cb 100644
--- a/drivers/net/bna/bfa_ioc.c
+++ b/drivers/net/bna/bfa_ioc.c
@@ -26,25 +26,6 @@
* IOC local definitions
*/
-#define bfa_ioc_timer_start(__ioc) \
- mod_timer(&(__ioc)->ioc_timer, jiffies + \
- msecs_to_jiffies(BFA_IOC_TOV))
-#define bfa_ioc_timer_stop(__ioc) del_timer(&(__ioc)->ioc_timer)
-
-#define bfa_ioc_recovery_timer_start(__ioc) \
- mod_timer(&(__ioc)->ioc_timer, jiffies + \
- msecs_to_jiffies(BFA_IOC_TOV_RECOVER))
-
-#define bfa_sem_timer_start(__ioc) \
- mod_timer(&(__ioc)->sem_timer, jiffies + \
- msecs_to_jiffies(BFA_IOC_HWSEM_TOV))
-#define bfa_sem_timer_stop(__ioc) del_timer(&(__ioc)->sem_timer)
-
-#define bfa_hb_timer_start(__ioc) \
- mod_timer(&(__ioc)->hb_timer, jiffies + \
- msecs_to_jiffies(BFA_IOC_HB_TOV))
-#define bfa_hb_timer_stop(__ioc) del_timer(&(__ioc)->hb_timer)
-
/**
* Asic specific macros : see bfa_hw_cb.c and bfa_hw_ct.c for details.
*/
@@ -55,8 +36,16 @@
((__ioc)->ioc_hwif->ioc_firmware_unlock(__ioc))
#define bfa_ioc_reg_init(__ioc) ((__ioc)->ioc_hwif->ioc_reg_init(__ioc))
#define bfa_ioc_map_port(__ioc) ((__ioc)->ioc_hwif->ioc_map_port(__ioc))
-#define bfa_ioc_notify_hbfail(__ioc) \
- ((__ioc)->ioc_hwif->ioc_notify_hbfail(__ioc))
+#define bfa_ioc_notify_fail(__ioc) \
+ ((__ioc)->ioc_hwif->ioc_notify_fail(__ioc))
+#define bfa_ioc_sync_join(__ioc) \
+ ((__ioc)->ioc_hwif->ioc_sync_join(__ioc))
+#define bfa_ioc_sync_leave(__ioc) \
+ ((__ioc)->ioc_hwif->ioc_sync_leave(__ioc))
+#define bfa_ioc_sync_ack(__ioc) \
+ ((__ioc)->ioc_hwif->ioc_sync_ack(__ioc))
+#define bfa_ioc_sync_complete(__ioc) \
+ ((__ioc)->ioc_hwif->ioc_sync_complete(__ioc))
#define bfa_ioc_mbox_cmd_pending(__ioc) \
(!list_empty(&((__ioc)->mbox_mod.cmd_q)) || \
@@ -82,6 +71,12 @@ static void bfa_ioc_recover(struct bfa_ioc *ioc);
static void bfa_ioc_check_attr_wwns(struct bfa_ioc *ioc);
static void bfa_ioc_disable_comp(struct bfa_ioc *ioc);
static void bfa_ioc_lpu_stop(struct bfa_ioc *ioc);
+static void bfa_ioc_fail_notify(struct bfa_ioc *ioc);
+static void bfa_ioc_pf_enabled(struct bfa_ioc *ioc);
+static void bfa_ioc_pf_disabled(struct bfa_ioc *ioc);
+static void bfa_ioc_pf_initfailed(struct bfa_ioc *ioc);
+static void bfa_ioc_pf_failed(struct bfa_ioc *ioc);
+static void bfa_ioc_pf_fwmismatch(struct bfa_ioc *ioc);
static void bfa_ioc_boot(struct bfa_ioc *ioc, u32 boot_type,
u32 boot_param);
static u32 bfa_ioc_smem_pgnum(struct bfa_ioc *ioc, u32 fmaddr);
@@ -100,69 +95,171 @@ static void bfa_ioc_get_adapter_model(struct bfa_ioc *ioc, char *model);
static u64 bfa_ioc_get_pwwn(struct bfa_ioc *ioc);
/**
- * IOC state machine events
+ * IOC state machine definitions/declarations
*/
enum ioc_event {
- IOC_E_ENABLE = 1, /*!< IOC enable request */
- IOC_E_DISABLE = 2, /*!< IOC disable request */
- IOC_E_TIMEOUT = 3, /*!< f/w response timeout */
- IOC_E_FWREADY = 4, /*!< f/w initialization done */
- IOC_E_FWRSP_GETATTR = 5, /*!< IOC get attribute response */
- IOC_E_FWRSP_ENABLE = 6, /*!< enable f/w response */
- IOC_E_FWRSP_DISABLE = 7, /*!< disable f/w response */
- IOC_E_HBFAIL = 8, /*!< heartbeat failure */
- IOC_E_HWERROR = 9, /*!< hardware error interrupt */
- IOC_E_SEMLOCKED = 10, /*!< h/w semaphore is locked */
- IOC_E_DETACH = 11, /*!< driver detach cleanup */
+ IOC_E_RESET = 1, /*!< IOC reset request */
+ IOC_E_ENABLE = 2, /*!< IOC enable request */
+ IOC_E_DISABLE = 3, /*!< IOC disable request */
+ IOC_E_DETACH = 4, /*!< driver detach cleanup */
+ IOC_E_ENABLED = 5, /*!< f/w enabled */
+ IOC_E_FWRSP_GETATTR = 6, /*!< IOC get attribute response */
+ IOC_E_DISABLED = 7, /*!< f/w disabled */
+ IOC_E_INITFAILED = 8, /*!< failure notice by iocpf sm */
+ IOC_E_PFAILED = 9, /*!< failure notice by iocpf sm */
+ IOC_E_HBFAIL = 10, /*!< heartbeat failure */
+ IOC_E_HWERROR = 11, /*!< hardware error interrupt */
+ IOC_E_TIMEOUT = 12, /*!< timeout */
};
+bfa_fsm_state_decl(bfa_ioc, uninit, struct bfa_ioc, enum ioc_event);
bfa_fsm_state_decl(bfa_ioc, reset, struct bfa_ioc, enum ioc_event);
-bfa_fsm_state_decl(bfa_ioc, fwcheck, struct bfa_ioc, enum ioc_event);
-bfa_fsm_state_decl(bfa_ioc, mismatch, struct bfa_ioc, enum ioc_event);
-bfa_fsm_state_decl(bfa_ioc, semwait, struct bfa_ioc, enum ioc_event);
-bfa_fsm_state_decl(bfa_ioc, hwinit, struct bfa_ioc, enum ioc_event);
bfa_fsm_state_decl(bfa_ioc, enabling, struct bfa_ioc, enum ioc_event);
bfa_fsm_state_decl(bfa_ioc, getattr, struct bfa_ioc, enum ioc_event);
bfa_fsm_state_decl(bfa_ioc, op, struct bfa_ioc, enum ioc_event);
-bfa_fsm_state_decl(bfa_ioc, initfail, struct bfa_ioc, enum ioc_event);
-bfa_fsm_state_decl(bfa_ioc, hbfail, struct bfa_ioc, enum ioc_event);
+bfa_fsm_state_decl(bfa_ioc, fail_retry, struct bfa_ioc, enum ioc_event);
+bfa_fsm_state_decl(bfa_ioc, fail, struct bfa_ioc, enum ioc_event);
bfa_fsm_state_decl(bfa_ioc, disabling, struct bfa_ioc, enum ioc_event);
bfa_fsm_state_decl(bfa_ioc, disabled, struct bfa_ioc, enum ioc_event);
static struct bfa_sm_table ioc_sm_table[] = {
+ {BFA_SM(bfa_ioc_sm_uninit), BFA_IOC_UNINIT},
{BFA_SM(bfa_ioc_sm_reset), BFA_IOC_RESET},
- {BFA_SM(bfa_ioc_sm_fwcheck), BFA_IOC_FWMISMATCH},
- {BFA_SM(bfa_ioc_sm_mismatch), BFA_IOC_FWMISMATCH},
- {BFA_SM(bfa_ioc_sm_semwait), BFA_IOC_SEMWAIT},
- {BFA_SM(bfa_ioc_sm_hwinit), BFA_IOC_HWINIT},
- {BFA_SM(bfa_ioc_sm_enabling), BFA_IOC_HWINIT},
+ {BFA_SM(bfa_ioc_sm_enabling), BFA_IOC_ENABLING},
{BFA_SM(bfa_ioc_sm_getattr), BFA_IOC_GETATTR},
{BFA_SM(bfa_ioc_sm_op), BFA_IOC_OPERATIONAL},
- {BFA_SM(bfa_ioc_sm_initfail), BFA_IOC_INITFAIL},
- {BFA_SM(bfa_ioc_sm_hbfail), BFA_IOC_HBFAIL},
+ {BFA_SM(bfa_ioc_sm_fail_retry), BFA_IOC_INITFAIL},
+ {BFA_SM(bfa_ioc_sm_fail), BFA_IOC_FAIL},
{BFA_SM(bfa_ioc_sm_disabling), BFA_IOC_DISABLING},
{BFA_SM(bfa_ioc_sm_disabled), BFA_IOC_DISABLED},
};
/**
+ * IOCPF state machine definitions/declarations
+ */
+
+/*
+ * Forward declareations for iocpf state machine
+ */
+static void bfa_iocpf_enable(struct bfa_ioc *ioc);
+static void bfa_iocpf_disable(struct bfa_ioc *ioc);
+static void bfa_iocpf_fail(struct bfa_ioc *ioc);
+static void bfa_iocpf_initfail(struct bfa_ioc *ioc);
+static void bfa_iocpf_getattrfail(struct bfa_ioc *ioc);
+static void bfa_iocpf_stop(struct bfa_ioc *ioc);
+
+/**
+ * IOCPF state machine events
+ */
+enum iocpf_event {
+ IOCPF_E_ENABLE = 1, /*!< IOCPF enable request */
+ IOCPF_E_DISABLE = 2, /*!< IOCPF disable request */
+ IOCPF_E_STOP = 3, /*!< stop on driver detach */
+ IOCPF_E_FWREADY = 4, /*!< f/w initialization done */
+ IOCPF_E_FWRSP_ENABLE = 5, /*!< enable f/w response */
+ IOCPF_E_FWRSP_DISABLE = 6, /*!< disable f/w response */
+ IOCPF_E_FAIL = 7, /*!< failure notice by ioc sm */
+ IOCPF_E_INITFAIL = 8, /*!< init fail notice by ioc sm */
+ IOCPF_E_GETATTRFAIL = 9, /*!< init fail notice by ioc sm */
+ IOCPF_E_SEMLOCKED = 10, /*!< h/w semaphore is locked */
+ IOCPF_E_TIMEOUT = 11, /*!< f/w response timeout */
+};
+
+/**
+ * IOCPF states
+ */
+enum bfa_iocpf_state {
+ BFA_IOCPF_RESET = 1, /*!< IOC is in reset state */
+ BFA_IOCPF_SEMWAIT = 2, /*!< Waiting for IOC h/w semaphore */
+ BFA_IOCPF_HWINIT = 3, /*!< IOC h/w is being initialized */
+ BFA_IOCPF_READY = 4, /*!< IOCPF is initialized */
+ BFA_IOCPF_INITFAIL = 5, /*!< IOCPF failed */
+ BFA_IOCPF_FAIL = 6, /*!< IOCPF failed */
+ BFA_IOCPF_DISABLING = 7, /*!< IOCPF is being disabled */
+ BFA_IOCPF_DISABLED = 8, /*!< IOCPF is disabled */
+ BFA_IOCPF_FWMISMATCH = 9, /*!< IOC f/w different from drivers */
+};
+
+bfa_fsm_state_decl(bfa_iocpf, reset, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, fwcheck, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, mismatch, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, semwait, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, hwinit, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, enabling, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, ready, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, initfail_sync, struct bfa_iocpf,
+ enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, initfail, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, fail_sync, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, fail, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, disabling, struct bfa_iocpf, enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, disabling_sync, struct bfa_iocpf,
+ enum iocpf_event);
+bfa_fsm_state_decl(bfa_iocpf, disabled, struct bfa_iocpf, enum iocpf_event);
+
+static struct bfa_sm_table iocpf_sm_table[] = {
+ {BFA_SM(bfa_iocpf_sm_reset), BFA_IOCPF_RESET},
+ {BFA_SM(bfa_iocpf_sm_fwcheck), BFA_IOCPF_FWMISMATCH},
+ {BFA_SM(bfa_iocpf_sm_mismatch), BFA_IOCPF_FWMISMATCH},
+ {BFA_SM(bfa_iocpf_sm_semwait), BFA_IOCPF_SEMWAIT},
+ {BFA_SM(bfa_iocpf_sm_hwinit), BFA_IOCPF_HWINIT},
+ {BFA_SM(bfa_iocpf_sm_enabling), BFA_IOCPF_HWINIT},
+ {BFA_SM(bfa_iocpf_sm_ready), BFA_IOCPF_READY},
+ {BFA_SM(bfa_iocpf_sm_initfail_sync), BFA_IOCPF_INITFAIL},
+ {BFA_SM(bfa_iocpf_sm_initfail), BFA_IOCPF_INITFAIL},
+ {BFA_SM(bfa_iocpf_sm_fail_sync), BFA_IOCPF_FAIL},
+ {BFA_SM(bfa_iocpf_sm_fail), BFA_IOCPF_FAIL},
+ {BFA_SM(bfa_iocpf_sm_disabling), BFA_IOCPF_DISABLING},
+ {BFA_SM(bfa_iocpf_sm_disabling_sync), BFA_IOCPF_DISABLING},
+ {BFA_SM(bfa_iocpf_sm_disabled), BFA_IOCPF_DISABLED},
+};
+
+/**
+ * IOC State Machine
+ */
+
+/**
+ * Beginning state. IOC uninit state.
+ */
+static void
+bfa_ioc_sm_uninit_entry(struct bfa_ioc *ioc)
+{
+}
+
+/**
+ * IOC is in uninit state.
+ */
+static void
+bfa_ioc_sm_uninit(struct bfa_ioc *ioc, enum ioc_event event)
+{
+ switch (event) {
+ case IOC_E_RESET:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_reset);
+ break;
+
+ default:
+ bfa_sm_fault(ioc, event);
+ }
+}
+
+/**
* Reset entry actions -- initialize state machine
*/
static void
bfa_ioc_sm_reset_entry(struct bfa_ioc *ioc)
{
- ioc->retry_count = 0;
- ioc->auto_recover = bfa_nw_auto_recover;
+ bfa_fsm_set_state(&ioc->iocpf, bfa_iocpf_sm_reset);
}
/**
- * Beginning state. IOC is in reset state.
+ * IOC is in reset state.
*/
static void
bfa_ioc_sm_reset(struct bfa_ioc *ioc, enum ioc_event event)
{
switch (event) {
case IOC_E_ENABLE:
- bfa_fsm_set_state(ioc, bfa_ioc_sm_fwcheck);
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_enabling);
break;
case IOC_E_DISABLE:
@@ -170,6 +267,7 @@ bfa_ioc_sm_reset(struct bfa_ioc *ioc, enum ioc_event event)
break;
case IOC_E_DETACH:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_uninit);
break;
default:
@@ -177,42 +275,43 @@ bfa_ioc_sm_reset(struct bfa_ioc *ioc, enum ioc_event event)
}
}
-/**
- * Semaphore should be acquired for version check.
- */
static void
-bfa_ioc_sm_fwcheck_entry(struct bfa_ioc *ioc)
+bfa_ioc_sm_enabling_entry(struct bfa_ioc *ioc)
{
- bfa_ioc_hw_sem_get(ioc);
+ bfa_iocpf_enable(ioc);
}
/**
- * Awaiting h/w semaphore to continue with version check.
+ * Host IOC function is being enabled, awaiting response from firmware.
+ * Semaphore is acquired.
*/
static void
-bfa_ioc_sm_fwcheck(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_ioc_sm_enabling(struct bfa_ioc *ioc, enum ioc_event event)
{
switch (event) {
- case IOC_E_SEMLOCKED:
- if (bfa_ioc_firmware_lock(ioc)) {
- ioc->retry_count = 0;
- bfa_fsm_set_state(ioc, bfa_ioc_sm_hwinit);
- } else {
- bfa_nw_ioc_hw_sem_release(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_mismatch);
- }
+ case IOC_E_ENABLED:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_getattr);
+ break;
+
+ case IOC_E_PFAILED:
+ /* !!! fall through !!! */
+ case IOC_E_HWERROR:
+ ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_IOC_FAILURE);
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_fail_retry);
+ if (event != IOC_E_PFAILED)
+ bfa_iocpf_initfail(ioc);
break;
case IOC_E_DISABLE:
- bfa_ioc_disable_comp(ioc);
- /* fall through */
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_disabling);
+ break;
case IOC_E_DETACH:
- bfa_ioc_hw_sem_get_cancel(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_reset);
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_uninit);
+ bfa_iocpf_stop(ioc);
break;
- case IOC_E_FWREADY:
+ case IOC_E_ENABLE:
break;
default:
@@ -221,41 +320,85 @@ bfa_ioc_sm_fwcheck(struct bfa_ioc *ioc, enum ioc_event event)
}
/**
- * Notify enable completion callback and generate mismatch AEN.
+ * Semaphore should be acquired for version check.
*/
static void
-bfa_ioc_sm_mismatch_entry(struct bfa_ioc *ioc)
+bfa_ioc_sm_getattr_entry(struct bfa_ioc *ioc)
{
- /**
- * Provide enable completion callback and AEN notification only once.
- */
- if (ioc->retry_count == 0)
- ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_IOC_FAILURE);
- ioc->retry_count++;
- bfa_ioc_timer_start(ioc);
+ mod_timer(&ioc->ioc_timer, jiffies +
+ msecs_to_jiffies(BFA_IOC_TOV));
+ bfa_ioc_send_getattr(ioc);
}
/**
- * Awaiting firmware version match.
+ * IOC configuration in progress. Timer is active.
*/
static void
-bfa_ioc_sm_mismatch(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_ioc_sm_getattr(struct bfa_ioc *ioc, enum ioc_event event)
{
switch (event) {
+ case IOC_E_FWRSP_GETATTR:
+ del_timer(&ioc->ioc_timer);
+ bfa_ioc_check_attr_wwns(ioc);
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_op);
+ break;
+
+ case IOC_E_PFAILED:
+ case IOC_E_HWERROR:
+ del_timer(&ioc->ioc_timer);
+ /* fall through */
case IOC_E_TIMEOUT:
- bfa_fsm_set_state(ioc, bfa_ioc_sm_fwcheck);
+ ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_IOC_FAILURE);
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_fail_retry);
+ if (event != IOC_E_PFAILED)
+ bfa_iocpf_getattrfail(ioc);
break;
case IOC_E_DISABLE:
- bfa_ioc_disable_comp(ioc);
- /* fall through */
+ del_timer(&ioc->ioc_timer);
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_disabling);
+ break;
- case IOC_E_DETACH:
- bfa_ioc_timer_stop(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_reset);
+ case IOC_E_ENABLE:
+ break;
+
+ default:
+ bfa_sm_fault(ioc, event);
+ }
+}
+
+static void
+bfa_ioc_sm_op_entry(struct bfa_ioc *ioc)
+{
+ ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_OK);
+ bfa_ioc_hb_monitor(ioc);
+}
+
+static void
+bfa_ioc_sm_op(struct bfa_ioc *ioc, enum ioc_event event)
+{
+ switch (event) {
+ case IOC_E_ENABLE:
+ break;
+
+ case IOC_E_DISABLE:
+ bfa_ioc_hb_stop(ioc);
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_disabling);
break;
- case IOC_E_FWREADY:
+ case IOC_E_PFAILED:
+ case IOC_E_HWERROR:
+ bfa_ioc_hb_stop(ioc);
+ /* !!! fall through !!! */
+ case IOC_E_HBFAIL:
+ bfa_ioc_fail_notify(ioc);
+ if (ioc->iocpf.auto_recover)
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_fail_retry);
+ else
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_fail);
+
+ if (event != IOC_E_PFAILED)
+ bfa_iocpf_fail(ioc);
break;
default:
@@ -263,30 +406,61 @@ bfa_ioc_sm_mismatch(struct bfa_ioc *ioc, enum ioc_event event)
}
}
+static void
+bfa_ioc_sm_disabling_entry(struct bfa_ioc *ioc)
+{
+ bfa_iocpf_disable(ioc);
+}
+
/**
- * Request for semaphore.
+ * IOC is being desabled
*/
static void
-bfa_ioc_sm_semwait_entry(struct bfa_ioc *ioc)
+bfa_ioc_sm_disabling(struct bfa_ioc *ioc, enum ioc_event event)
{
- bfa_ioc_hw_sem_get(ioc);
+ switch (event) {
+ case IOC_E_DISABLED:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_disabled);
+ break;
+
+ case IOC_E_HWERROR:
+ /*
+ * No state change. Will move to disabled state
+ * after iocpf sm completes failure processing and
+ * moves to disabled state.
+ */
+ bfa_iocpf_fail(ioc);
+ break;
+
+ default:
+ bfa_sm_fault(ioc, event);
+ }
}
/**
- * Awaiting semaphore for h/w initialzation.
+ * IOC desable completion entry.
*/
static void
-bfa_ioc_sm_semwait(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_ioc_sm_disabled_entry(struct bfa_ioc *ioc)
+{
+ bfa_ioc_disable_comp(ioc);
+}
+
+static void
+bfa_ioc_sm_disabled(struct bfa_ioc *ioc, enum ioc_event event)
{
switch (event) {
- case IOC_E_SEMLOCKED:
- ioc->retry_count = 0;
- bfa_fsm_set_state(ioc, bfa_ioc_sm_hwinit);
+ case IOC_E_ENABLE:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_enabling);
break;
case IOC_E_DISABLE:
- bfa_ioc_hw_sem_get_cancel(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_disabled);
+ ioc->cbfn->disable_cbfn(ioc->bfa);
+ break;
+
+ case IOC_E_DETACH:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_uninit);
+ bfa_iocpf_stop(ioc);
break;
default:
@@ -295,46 +469,45 @@ bfa_ioc_sm_semwait(struct bfa_ioc *ioc, enum ioc_event event)
}
static void
-bfa_ioc_sm_hwinit_entry(struct bfa_ioc *ioc)
+bfa_ioc_sm_fail_retry_entry(struct bfa_ioc *ioc)
{
- bfa_ioc_timer_start(ioc);
- bfa_ioc_reset(ioc, false);
}
/**
- * @brief
- * Hardware is being initialized. Interrupts are enabled.
- * Holding hardware semaphore lock.
+ * Hardware initialization retry.
*/
static void
-bfa_ioc_sm_hwinit(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_ioc_sm_fail_retry(struct bfa_ioc *ioc, enum ioc_event event)
{
switch (event) {
- case IOC_E_FWREADY:
- bfa_ioc_timer_stop(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_enabling);
+ case IOC_E_ENABLED:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_getattr);
break;
+ case IOC_E_PFAILED:
case IOC_E_HWERROR:
- bfa_ioc_timer_stop(ioc);
- /* fall through */
+ /**
+ * Initialization retry failed.
+ */
+ ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_IOC_FAILURE);
+ if (event != IOC_E_PFAILED)
+ bfa_iocpf_initfail(ioc);
+ break;
- case IOC_E_TIMEOUT:
- ioc->retry_count++;
- if (ioc->retry_count < BFA_IOC_HWINIT_MAX) {
- bfa_ioc_timer_start(ioc);
- bfa_ioc_reset(ioc, true);
- break;
- }
+ case IOC_E_INITFAILED:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_fail);
+ break;
- bfa_nw_ioc_hw_sem_release(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_initfail);
+ case IOC_E_ENABLE:
break;
case IOC_E_DISABLE:
- bfa_nw_ioc_hw_sem_release(ioc);
- bfa_ioc_timer_stop(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_disabled);
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_disabling);
+ break;
+
+ case IOC_E_DETACH:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_uninit);
+ bfa_iocpf_stop(ioc);
break;
default:
@@ -343,51 +516,248 @@ bfa_ioc_sm_hwinit(struct bfa_ioc *ioc, enum ioc_event event)
}
static void
-bfa_ioc_sm_enabling_entry(struct bfa_ioc *ioc)
+bfa_ioc_sm_fail_entry(struct bfa_ioc *ioc)
{
- bfa_ioc_timer_start(ioc);
- bfa_ioc_send_enable(ioc);
}
/**
- * Host IOC function is being enabled, awaiting response from firmware.
- * Semaphore is acquired.
+ * IOC failure.
*/
static void
-bfa_ioc_sm_enabling(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_ioc_sm_fail(struct bfa_ioc *ioc, enum ioc_event event)
{
switch (event) {
- case IOC_E_FWRSP_ENABLE:
- bfa_ioc_timer_stop(ioc);
- bfa_nw_ioc_hw_sem_release(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_getattr);
+ case IOC_E_ENABLE:
+ ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_IOC_FAILURE);
+ break;
+
+ case IOC_E_DISABLE:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_disabling);
+ break;
+
+ case IOC_E_DETACH:
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_uninit);
+ bfa_iocpf_stop(ioc);
break;
case IOC_E_HWERROR:
- bfa_ioc_timer_stop(ioc);
- /* fall through */
+ /* HB failure notification, ignore. */
+ break;
- case IOC_E_TIMEOUT:
- ioc->retry_count++;
- if (ioc->retry_count < BFA_IOC_HWINIT_MAX) {
- writel(BFI_IOC_UNINIT,
- ioc->ioc_regs.ioc_fwstate);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_hwinit);
- break;
+ default:
+ bfa_sm_fault(ioc, event);
+ }
+}
+
+/**
+ * IOCPF State Machine
+ */
+
+/**
+ * Reset entry actions -- initialize state machine
+ */
+static void
+bfa_iocpf_sm_reset_entry(struct bfa_iocpf *iocpf)
+{
+ iocpf->retry_count = 0;
+ iocpf->auto_recover = bfa_nw_auto_recover;
+}
+
+/**
+ * Beginning state. IOC is in reset state.
+ */
+static void
+bfa_iocpf_sm_reset(struct bfa_iocpf *iocpf, enum iocpf_event event)
+{
+ switch (event) {
+ case IOCPF_E_ENABLE:
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_fwcheck);
+ break;
+
+ case IOCPF_E_STOP:
+ break;
+
+ default:
+ bfa_sm_fault(iocpf->ioc, event);
+ }
+}
+
+/**
+ * Semaphore should be acquired for version check.
+ */
+static void
+bfa_iocpf_sm_fwcheck_entry(struct bfa_iocpf *iocpf)
+{
+ bfa_ioc_hw_sem_get(iocpf->ioc);
+}
+
+/**
+ * Awaiting h/w semaphore to continue with version check.
+ */
+static void
+bfa_iocpf_sm_fwcheck(struct bfa_iocpf *iocpf, enum iocpf_event event)
+{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
+ switch (event) {
+ case IOCPF_E_SEMLOCKED:
+ if (bfa_ioc_firmware_lock(ioc)) {
+ if (bfa_ioc_sync_complete(ioc)) {
+ iocpf->retry_count = 0;
+ bfa_ioc_sync_join(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_hwinit);
+ } else {
+ bfa_ioc_firmware_unlock(ioc);
+ bfa_nw_ioc_hw_sem_release(ioc);
+ mod_timer(&ioc->sem_timer, jiffies +
+ msecs_to_jiffies(BFA_IOC_HWSEM_TOV));
+ }
+ } else {
+ bfa_nw_ioc_hw_sem_release(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_mismatch);
}
+ break;
- bfa_nw_ioc_hw_sem_release(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_initfail);
+ case IOCPF_E_DISABLE:
+ bfa_ioc_hw_sem_get_cancel(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_reset);
+ bfa_ioc_pf_disabled(ioc);
break;
- case IOC_E_DISABLE:
- bfa_ioc_timer_stop(ioc);
+ case IOCPF_E_STOP:
+ bfa_ioc_hw_sem_get_cancel(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_reset);
+ break;
+
+ default:
+ bfa_sm_fault(ioc, event);
+ }
+}
+
+/**
+ * Notify enable completion callback
+ */
+static void
+bfa_iocpf_sm_mismatch_entry(struct bfa_iocpf *iocpf)
+{
+ /* Call only the first time sm enters fwmismatch state. */
+ if (iocpf->retry_count == 0)
+ bfa_ioc_pf_fwmismatch(iocpf->ioc);
+
+ iocpf->retry_count++;
+ mod_timer(&(iocpf->ioc)->iocpf_timer, jiffies +
+ msecs_to_jiffies(BFA_IOC_TOV));
+}
+
+/**
+ * Awaiting firmware version match.
+ */
+static void
+bfa_iocpf_sm_mismatch(struct bfa_iocpf *iocpf, enum iocpf_event event)
+{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
+ switch (event) {
+ case IOCPF_E_TIMEOUT:
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_fwcheck);
+ break;
+
+ case IOCPF_E_DISABLE:
+ del_timer(&ioc->iocpf_timer);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_reset);
+ bfa_ioc_pf_disabled(ioc);
+ break;
+
+ case IOCPF_E_STOP:
+ del_timer(&ioc->iocpf_timer);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_reset);
+ break;
+
+ default:
+ bfa_sm_fault(ioc, event);
+ }
+}
+
+/**
+ * Request for semaphore.
+ */
+static void
+bfa_iocpf_sm_semwait_entry(struct bfa_iocpf *iocpf)
+{
+ bfa_ioc_hw_sem_get(iocpf->ioc);
+}
+
+/**
+ * Awaiting semaphore for h/w initialzation.
+ */
+static void
+bfa_iocpf_sm_semwait(struct bfa_iocpf *iocpf, enum iocpf_event event)
+{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
+ switch (event) {
+ case IOCPF_E_SEMLOCKED:
+ if (bfa_ioc_sync_complete(ioc)) {
+ bfa_ioc_sync_join(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_hwinit);
+ } else {
+ bfa_nw_ioc_hw_sem_release(ioc);
+ mod_timer(&ioc->sem_timer, jiffies +
+ msecs_to_jiffies(BFA_IOC_HWSEM_TOV));
+ }
+ break;
+
+ case IOCPF_E_DISABLE:
+ bfa_ioc_hw_sem_get_cancel(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabling_sync);
+ break;
+
+ default:
+ bfa_sm_fault(ioc, event);
+ }
+}
+
+static void
+bfa_iocpf_sm_hwinit_entry(struct bfa_iocpf *iocpf)
+{
+ mod_timer(&(iocpf->ioc)->iocpf_timer, jiffies +
+ msecs_to_jiffies(BFA_IOC_TOV));
+ bfa_ioc_reset(iocpf->ioc, 0);
+}
+
+/**
+ * Hardware is being initialized. Interrupts are enabled.
+ * Holding hardware semaphore lock.
+ */
+static void
+bfa_iocpf_sm_hwinit(struct bfa_iocpf *iocpf, enum iocpf_event event)
+{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
+ switch (event) {
+ case IOCPF_E_FWREADY:
+ del_timer(&ioc->iocpf_timer);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_enabling);
+ break;
+
+ case IOCPF_E_INITFAIL:
+ del_timer(&ioc->iocpf_timer);
+ /*
+ * !!! fall through !!!
+ */
+
+ case IOCPF_E_TIMEOUT:
bfa_nw_ioc_hw_sem_release(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_disabled);
+ if (event == IOCPF_E_TIMEOUT)
+ bfa_ioc_pf_failed(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_initfail_sync);
break;
- case IOC_E_FWREADY:
- bfa_ioc_send_enable(ioc);
+ case IOCPF_E_DISABLE:
+ del_timer(&ioc->iocpf_timer);
+ bfa_ioc_sync_leave(ioc);
+ bfa_nw_ioc_hw_sem_release(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabled);
break;
default:
@@ -396,37 +766,49 @@ bfa_ioc_sm_enabling(struct bfa_ioc *ioc, enum ioc_event event)
}
static void
-bfa_ioc_sm_getattr_entry(struct bfa_ioc *ioc)
+bfa_iocpf_sm_enabling_entry(struct bfa_iocpf *iocpf)
{
- bfa_ioc_timer_start(ioc);
- bfa_ioc_send_getattr(ioc);
+ mod_timer(&(iocpf->ioc)->iocpf_timer, jiffies +
+ msecs_to_jiffies(BFA_IOC_TOV));
+ bfa_ioc_send_enable(iocpf->ioc);
}
/**
- * @brief
- * IOC configuration in progress. Timer is active.
+ * Host IOC function is being enabled, awaiting response from firmware.
+ * Semaphore is acquired.
*/
static void
-bfa_ioc_sm_getattr(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_iocpf_sm_enabling(struct bfa_iocpf *iocpf, enum iocpf_event event)
{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
switch (event) {
- case IOC_E_FWRSP_GETATTR:
- bfa_ioc_timer_stop(ioc);
- bfa_ioc_check_attr_wwns(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_op);
+ case IOCPF_E_FWRSP_ENABLE:
+ del_timer(&ioc->iocpf_timer);
+ bfa_nw_ioc_hw_sem_release(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_ready);
break;
- case IOC_E_HWERROR:
- bfa_ioc_timer_stop(ioc);
- /* fall through */
+ case IOCPF_E_INITFAIL:
+ del_timer(&ioc->iocpf_timer);
+ /*
+ * !!! fall through !!!
+ */
+ case IOCPF_E_TIMEOUT:
+ bfa_nw_ioc_hw_sem_release(ioc);
+ if (event == IOCPF_E_TIMEOUT)
+ bfa_ioc_pf_failed(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_initfail_sync);
+ break;
- case IOC_E_TIMEOUT:
- bfa_fsm_set_state(ioc, bfa_ioc_sm_initfail);
+ case IOCPF_E_DISABLE:
+ del_timer(&ioc->iocpf_timer);
+ bfa_nw_ioc_hw_sem_release(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabling);
break;
- case IOC_E_DISABLE:
- bfa_ioc_timer_stop(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_disabled);
+ case IOCPF_E_FWREADY:
+ bfa_ioc_send_enable(ioc);
break;
default:
@@ -434,36 +816,42 @@ bfa_ioc_sm_getattr(struct bfa_ioc *ioc, enum ioc_event event)
}
}
+static bool
+bfa_nw_ioc_is_operational(struct bfa_ioc *ioc)
+{
+ return bfa_fsm_cmp_state(ioc, bfa_ioc_sm_op);
+}
+
static void
-bfa_ioc_sm_op_entry(struct bfa_ioc *ioc)
+bfa_iocpf_sm_ready_entry(struct bfa_iocpf *iocpf)
{
- ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_OK);
- bfa_ioc_hb_monitor(ioc);
+ bfa_ioc_pf_enabled(iocpf->ioc);
}
static void
-bfa_ioc_sm_op(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_iocpf_sm_ready(struct bfa_iocpf *iocpf, enum iocpf_event event)
{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
switch (event) {
- case IOC_E_ENABLE:
+ case IOCPF_E_DISABLE:
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabling);
break;
- case IOC_E_DISABLE:
- bfa_ioc_hb_stop(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_disabling);
+ case IOCPF_E_GETATTRFAIL:
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_initfail_sync);
break;
- case IOC_E_HWERROR:
- case IOC_E_FWREADY:
- /**
- * Hard error or IOC recovery by other function.
- * Treat it same as heartbeat failure.
- */
- bfa_ioc_hb_stop(ioc);
- /* !!! fall through !!! */
+ case IOCPF_E_FAIL:
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_fail_sync);
+ break;
- case IOC_E_HBFAIL:
- bfa_fsm_set_state(ioc, bfa_ioc_sm_hbfail);
+ case IOCPF_E_FWREADY:
+ bfa_ioc_pf_failed(ioc);
+ if (bfa_nw_ioc_is_operational(ioc))
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_fail_sync);
+ else
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_initfail_sync);
break;
default:
@@ -472,33 +860,40 @@ bfa_ioc_sm_op(struct bfa_ioc *ioc, enum ioc_event event)
}
static void
-bfa_ioc_sm_disabling_entry(struct bfa_ioc *ioc)
+bfa_iocpf_sm_disabling_entry(struct bfa_iocpf *iocpf)
{
- bfa_ioc_timer_start(ioc);
- bfa_ioc_send_disable(ioc);
+ mod_timer(&(iocpf->ioc)->iocpf_timer, jiffies +
+ msecs_to_jiffies(BFA_IOC_TOV));
+ bfa_ioc_send_disable(iocpf->ioc);
}
/**
* IOC is being disabled
*/
static void
-bfa_ioc_sm_disabling(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_iocpf_sm_disabling(struct bfa_iocpf *iocpf, enum iocpf_event event)
{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
switch (event) {
- case IOC_E_FWRSP_DISABLE:
- bfa_ioc_timer_stop(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_disabled);
+ case IOCPF_E_FWRSP_DISABLE:
+ case IOCPF_E_FWREADY:
+ del_timer(&ioc->iocpf_timer);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabling_sync);
break;
- case IOC_E_HWERROR:
- bfa_ioc_timer_stop(ioc);
+ case IOCPF_E_FAIL:
+ del_timer(&ioc->iocpf_timer);
/*
* !!! fall through !!!
*/
- case IOC_E_TIMEOUT:
+ case IOCPF_E_TIMEOUT:
writel(BFI_IOC_FAIL, ioc->ioc_regs.ioc_fwstate);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_disabled);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabling_sync);
+ break;
+
+ case IOCPF_E_FWRSP_ENABLE:
break;
default:
@@ -506,33 +901,58 @@ bfa_ioc_sm_disabling(struct bfa_ioc *ioc, enum ioc_event event)
}
}
-/**
- * IOC disable completion entry.
- */
static void
-bfa_ioc_sm_disabled_entry(struct bfa_ioc *ioc)
+bfa_iocpf_sm_disabling_sync_entry(struct bfa_iocpf *iocpf)
{
- bfa_ioc_disable_comp(ioc);
+ bfa_ioc_hw_sem_get(iocpf->ioc);
}
+/**
+ * IOC hb ack request is being removed.
+ */
static void
-bfa_ioc_sm_disabled(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_iocpf_sm_disabling_sync(struct bfa_iocpf *iocpf, enum iocpf_event event)
{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
switch (event) {
- case IOC_E_ENABLE:
- bfa_fsm_set_state(ioc, bfa_ioc_sm_semwait);
+ case IOCPF_E_SEMLOCKED:
+ bfa_ioc_sync_leave(ioc);
+ bfa_nw_ioc_hw_sem_release(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabled);
break;
- case IOC_E_DISABLE:
- ioc->cbfn->disable_cbfn(ioc->bfa);
+ case IOCPF_E_FAIL:
break;
- case IOC_E_FWREADY:
+ default:
+ bfa_sm_fault(ioc, event);
+ }
+}
+
+/**
+ * IOC disable completion entry.
+ */
+static void
+bfa_iocpf_sm_disabled_entry(struct bfa_iocpf *iocpf)
+{
+ bfa_ioc_pf_disabled(iocpf->ioc);
+}
+
+static void
+bfa_iocpf_sm_disabled(struct bfa_iocpf *iocpf, enum iocpf_event event)
+{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
+ switch (event) {
+ case IOCPF_E_ENABLE:
+ iocpf->retry_count = 0;
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_semwait);
break;
- case IOC_E_DETACH:
+ case IOCPF_E_STOP:
bfa_ioc_firmware_unlock(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_reset);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_reset);
break;
default:
@@ -541,33 +961,50 @@ bfa_ioc_sm_disabled(struct bfa_ioc *ioc, enum ioc_event event)
}
static void
-bfa_ioc_sm_initfail_entry(struct bfa_ioc *ioc)
+bfa_iocpf_sm_initfail_sync_entry(struct bfa_iocpf *iocpf)
{
- ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_IOC_FAILURE);
- bfa_ioc_timer_start(ioc);
+ bfa_ioc_hw_sem_get(iocpf->ioc);
}
/**
- * @brief
* Hardware initialization failed.
*/
static void
-bfa_ioc_sm_initfail(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_iocpf_sm_initfail_sync(struct bfa_iocpf *iocpf, enum iocpf_event event)
{
+ struct bfa_ioc *ioc = iocpf->ioc;
+
switch (event) {
- case IOC_E_DISABLE:
- bfa_ioc_timer_stop(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_disabled);
+ case IOCPF_E_SEMLOCKED:
+ bfa_ioc_notify_fail(ioc);
+ bfa_ioc_sync_ack(ioc);
+ iocpf->retry_count++;
+ if (iocpf->retry_count >= BFA_IOC_HWINIT_MAX) {
+ bfa_ioc_sync_leave(ioc);
+ bfa_nw_ioc_hw_sem_release(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_initfail);
+ } else {
+ if (bfa_ioc_sync_complete(ioc))
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_hwinit);
+ else {
+ bfa_nw_ioc_hw_sem_release(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_semwait);
+ }
+ }
break;
- case IOC_E_DETACH:
- bfa_ioc_timer_stop(ioc);
+ case IOCPF_E_DISABLE:
+ bfa_ioc_hw_sem_get_cancel(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabling_sync);
+ break;
+
+ case IOCPF_E_STOP:
+ bfa_ioc_hw_sem_get_cancel(ioc);
bfa_ioc_firmware_unlock(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_reset);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_reset);
break;
- case IOC_E_TIMEOUT:
- bfa_fsm_set_state(ioc, bfa_ioc_sm_semwait);
+ case IOCPF_E_FAIL:
break;
default:
@@ -576,80 +1013,108 @@ bfa_ioc_sm_initfail(struct bfa_ioc *ioc, enum ioc_event event)
}
static void
-bfa_ioc_sm_hbfail_entry(struct bfa_ioc *ioc)
+bfa_iocpf_sm_initfail_entry(struct bfa_iocpf *iocpf)
{
- struct list_head *qe;
- struct bfa_ioc_hbfail_notify *notify;
+ bfa_ioc_pf_initfailed(iocpf->ioc);
+}
- /**
- * Mark IOC as failed in hardware and stop firmware.
- */
- bfa_ioc_lpu_stop(ioc);
- writel(BFI_IOC_FAIL, ioc->ioc_regs.ioc_fwstate);
+/**
+ * Hardware initialization failed.
+ */
+static void
+bfa_iocpf_sm_initfail(struct bfa_iocpf *iocpf, enum iocpf_event event)
+{
+ struct bfa_ioc *ioc = iocpf->ioc;
- /**
- * Notify other functions on HB failure.
- */
- bfa_ioc_notify_hbfail(ioc);
+ switch (event) {
+ case IOCPF_E_DISABLE:
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabled);
+ break;
- /**
- * Notify driver and common modules registered for notification.
- */
- ioc->cbfn->hbfail_cbfn(ioc->bfa);
- list_for_each(qe, &ioc->hb_notify_q) {
- notify = (struct bfa_ioc_hbfail_notify *) qe;
- notify->cbfn(notify->cbarg);
+ case IOCPF_E_STOP:
+ bfa_ioc_firmware_unlock(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_reset);
+ break;
+
+ default:
+ bfa_sm_fault(ioc, event);
}
+}
+static void
+bfa_iocpf_sm_fail_sync_entry(struct bfa_iocpf *iocpf)
+{
/**
- * Flush any queued up mailbox requests.
+ * Mark IOC as failed in hardware and stop firmware.
*/
- bfa_ioc_mbox_hbfail(ioc);
+ bfa_ioc_lpu_stop(iocpf->ioc);
/**
- * Trigger auto-recovery after a delay.
+ * Flush any queued up mailbox requests.
*/
- if (ioc->auto_recover)
- mod_timer(&ioc->ioc_timer, jiffies +
- msecs_to_jiffies(BFA_IOC_TOV_RECOVER));
+ bfa_ioc_mbox_hbfail(iocpf->ioc);
+ bfa_ioc_hw_sem_get(iocpf->ioc);
}
/**
- * @brief
- * IOC heartbeat failure.
+ * IOC is in failed state.
*/
static void
-bfa_ioc_sm_hbfail(struct bfa_ioc *ioc, enum ioc_event event)
+bfa_iocpf_sm_fail_sync(struct bfa_iocpf *iocpf, enum iocpf_event event)
{
- switch (event) {
+ struct bfa_ioc *ioc = iocpf->ioc;
- case IOC_E_ENABLE:
- ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_IOC_FAILURE);
+ switch (event) {
+ case IOCPF_E_SEMLOCKED:
+ iocpf->retry_count = 0;
+ bfa_ioc_sync_ack(ioc);
+ bfa_ioc_notify_fail(ioc);
+ if (!iocpf->auto_recover) {
+ bfa_ioc_sync_leave(ioc);
+ bfa_nw_ioc_hw_sem_release(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_fail);
+ } else {
+ if (bfa_ioc_sync_complete(ioc))
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_hwinit);
+ else {
+ bfa_nw_ioc_hw_sem_release(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_semwait);
+ }
+ }
break;
- case IOC_E_DISABLE:
- if (ioc->auto_recover)
- bfa_ioc_timer_stop(ioc);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_disabled);
+ case IOCPF_E_DISABLE:
+ bfa_ioc_hw_sem_get_cancel(ioc);
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabling_sync);
break;
- case IOC_E_TIMEOUT:
- bfa_fsm_set_state(ioc, bfa_ioc_sm_semwait);
+ case IOCPF_E_FAIL:
break;
- case IOC_E_FWREADY:
- /**
- * Recovery is already initiated by other function.
- */
- break;
+ default:
+ bfa_sm_fault(ioc, event);
+ }
+}
- case IOC_E_HWERROR:
- /*
- * HB failure notification, ignore.
- */
+static void
+bfa_iocpf_sm_fail_entry(struct bfa_iocpf *iocpf)
+{
+}
+
+/**
+ * @brief
+ * IOC is in failed state.
+ */
+static void
+bfa_iocpf_sm_fail(struct bfa_iocpf *iocpf, enum iocpf_event event)
+{
+ switch (event) {
+ case IOCPF_E_DISABLE:
+ bfa_fsm_set_state(iocpf, bfa_iocpf_sm_disabled);
break;
+
default:
- bfa_sm_fault(ioc, event);
+ bfa_sm_fault(iocpf->ioc, event);
}
}
@@ -674,14 +1139,6 @@ bfa_ioc_disable_comp(struct bfa_ioc *ioc)
}
}
-void
-bfa_nw_ioc_sem_timeout(void *ioc_arg)
-{
- struct bfa_ioc *ioc = (struct bfa_ioc *) ioc_arg;
-
- bfa_ioc_hw_sem_get(ioc);
-}
-
bool
bfa_nw_ioc_sem_get(void __iomem *sem_reg)
{
@@ -721,7 +1178,7 @@ bfa_ioc_hw_sem_get(struct bfa_ioc *ioc)
*/
r32 = readl(ioc->ioc_regs.ioc_sem_reg);
if (r32 == 0) {
- bfa_fsm_send_event(ioc, IOC_E_SEMLOCKED);
+ bfa_fsm_send_event(&ioc->iocpf, IOCPF_E_SEMLOCKED);
return;
}
@@ -932,7 +1389,7 @@ bfa_ioc_hwinit(struct bfa_ioc *ioc, bool force)
*/
bfa_ioc_msgflush(ioc);
ioc->cbfn->reset_cbfn(ioc->bfa);
- bfa_fsm_send_event(ioc, IOC_E_FWREADY);
+ bfa_fsm_send_event(&ioc->iocpf, IOCPF_E_FWREADY);
return;
}
@@ -1018,7 +1475,6 @@ bfa_nw_ioc_hb_check(void *cbarg)
hb_count = readl(ioc->ioc_regs.heartbeat);
if (ioc->hb_count == hb_count) {
- pr_crit("Firmware heartbeat failure at %d", hb_count);
bfa_ioc_recover(ioc);
return;
} else {
@@ -1189,6 +1645,55 @@ bfa_ioc_mbox_hbfail(struct bfa_ioc *ioc)
bfa_q_deq(&mod->cmd_q, &cmd);
}
+static void
+bfa_ioc_fail_notify(struct bfa_ioc *ioc)
+{
+ struct list_head *qe;
+ struct bfa_ioc_hbfail_notify *notify;
+
+ /**
+ * Notify driver and common modules registered for notification.
+ */
+ ioc->cbfn->hbfail_cbfn(ioc->bfa);
+ list_for_each(qe, &ioc->hb_notify_q) {
+ notify = (struct bfa_ioc_hbfail_notify *) qe;
+ notify->cbfn(notify->cbarg);
+ }
+}
+
+static void
+bfa_ioc_pf_enabled(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(ioc, IOC_E_ENABLED);
+}
+
+static void
+bfa_ioc_pf_disabled(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(ioc, IOC_E_DISABLED);
+}
+
+static void
+bfa_ioc_pf_initfailed(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(ioc, IOC_E_INITFAILED);
+}
+
+static void
+bfa_ioc_pf_failed(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(ioc, IOC_E_PFAILED);
+}
+
+static void
+bfa_ioc_pf_fwmismatch(struct bfa_ioc *ioc)
+{
+ /**
+ * Provide enable completion callback and AEN notification.
+ */
+ ioc->cbfn->enable_cbfn(ioc->bfa, BFA_STATUS_IOC_FAILURE);
+}
+
/**
* IOC public
*/
@@ -1284,6 +1789,7 @@ static void
bfa_ioc_isr(struct bfa_ioc *ioc, struct bfi_mbmsg *m)
{
union bfi_ioc_i2h_msg_u *msg;
+ struct bfa_iocpf *iocpf = &ioc->iocpf;
msg = (union bfi_ioc_i2h_msg_u *) m;
@@ -1294,15 +1800,15 @@ bfa_ioc_isr(struct bfa_ioc *ioc, struct bfi_mbmsg *m)
break;
case BFI_IOC_I2H_READY_EVENT:
- bfa_fsm_send_event(ioc, IOC_E_FWREADY);
+ bfa_fsm_send_event(iocpf, IOCPF_E_FWREADY);
break;
case BFI_IOC_I2H_ENABLE_REPLY:
- bfa_fsm_send_event(ioc, IOC_E_FWRSP_ENABLE);
+ bfa_fsm_send_event(iocpf, IOCPF_E_FWRSP_ENABLE);
break;
case BFI_IOC_I2H_DISABLE_REPLY:
- bfa_fsm_send_event(ioc, IOC_E_FWRSP_DISABLE);
+ bfa_fsm_send_event(iocpf, IOCPF_E_FWRSP_DISABLE);
break;
case BFI_IOC_I2H_GETATTR_REPLY:
@@ -1328,11 +1834,13 @@ bfa_nw_ioc_attach(struct bfa_ioc *ioc, void *bfa, struct bfa_ioc_cbfn *cbfn)
ioc->fcmode = false;
ioc->pllinit = false;
ioc->dbg_fwsave_once = true;
+ ioc->iocpf.ioc = ioc;
bfa_ioc_mbox_attach(ioc);
INIT_LIST_HEAD(&ioc->hb_notify_q);
- bfa_fsm_set_state(ioc, bfa_ioc_sm_reset);
+ bfa_fsm_set_state(ioc, bfa_ioc_sm_uninit);
+ bfa_fsm_send_event(ioc, IOC_E_RESET);
}
/**
@@ -1637,7 +2145,40 @@ bfa_ioc_get_adapter_model(struct bfa_ioc *ioc, char *model)
static enum bfa_ioc_state
bfa_ioc_get_state(struct bfa_ioc *ioc)
{
- return bfa_sm_to_state(ioc_sm_table, ioc->fsm);
+ enum bfa_iocpf_state iocpf_st;
+ enum bfa_ioc_state ioc_st = bfa_sm_to_state(ioc_sm_table, ioc->fsm);
+
+ if (ioc_st == BFA_IOC_ENABLING ||
+ ioc_st == BFA_IOC_FAIL || ioc_st == BFA_IOC_INITFAIL) {
+
+ iocpf_st = bfa_sm_to_state(iocpf_sm_table, ioc->iocpf.fsm);
+
+ switch (iocpf_st) {
+ case BFA_IOCPF_SEMWAIT:
+ ioc_st = BFA_IOC_SEMWAIT;
+ break;
+
+ case BFA_IOCPF_HWINIT:
+ ioc_st = BFA_IOC_HWINIT;
+ break;
+
+ case BFA_IOCPF_FWMISMATCH:
+ ioc_st = BFA_IOC_FWMISMATCH;
+ break;
+
+ case BFA_IOCPF_FAIL:
+ ioc_st = BFA_IOC_FAIL;
+ break;
+
+ case BFA_IOCPF_INITFAIL:
+ ioc_st = BFA_IOC_INITFAIL;
+ break;
+
+ default:
+ break;
+ }
+ }
+ return ioc_st;
}
void
@@ -1678,8 +2219,13 @@ bfa_nw_ioc_get_mac(struct bfa_ioc *ioc)
static void
bfa_ioc_recover(struct bfa_ioc *ioc)
{
- bfa_ioc_stats(ioc, ioc_hbfails);
- bfa_fsm_send_event(ioc, IOC_E_HBFAIL);
+ u16 bdf;
+
+ bdf = (ioc->pcidev.pci_slot << 8 | ioc->pcidev.pci_func << 3 |
+ ioc->pcidev.device_id);
+
+ pr_crit("Firmware heartbeat failure at %d", bdf);
+ BUG_ON(1);
}
static void
@@ -1687,5 +2233,61 @@ bfa_ioc_check_attr_wwns(struct bfa_ioc *ioc)
{
if (bfa_ioc_get_type(ioc) == BFA_IOC_TYPE_LL)
return;
+}
+
+/**
+ * @dg hal_iocpf_pvt BFA IOC PF private functions
+ * @{
+ */
+
+static void
+bfa_iocpf_enable(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(&ioc->iocpf, IOCPF_E_ENABLE);
+}
+static void
+bfa_iocpf_disable(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(&ioc->iocpf, IOCPF_E_DISABLE);
+}
+
+static void
+bfa_iocpf_fail(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(&ioc->iocpf, IOCPF_E_FAIL);
+}
+
+static void
+bfa_iocpf_initfail(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(&ioc->iocpf, IOCPF_E_INITFAIL);
+}
+
+static void
+bfa_iocpf_getattrfail(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(&ioc->iocpf, IOCPF_E_GETATTRFAIL);
+}
+
+static void
+bfa_iocpf_stop(struct bfa_ioc *ioc)
+{
+ bfa_fsm_send_event(&ioc->iocpf, IOCPF_E_STOP);
+}
+
+void
+bfa_nw_iocpf_timeout(void *ioc_arg)
+{
+ struct bfa_ioc *ioc = (struct bfa_ioc *) ioc_arg;
+
+ bfa_fsm_send_event(&ioc->iocpf, IOCPF_E_TIMEOUT);
+}
+
+void
+bfa_nw_iocpf_sem_timeout(void *ioc_arg)
+{
+ struct bfa_ioc *ioc = (struct bfa_ioc *) ioc_arg;
+
+ bfa_ioc_hw_sem_get(ioc);
}
diff --git a/drivers/net/bna/bfa_ioc.h b/drivers/net/bna/bfa_ioc.h
index a73d84e..e4974bc 100644
--- a/drivers/net/bna/bfa_ioc.h
+++ b/drivers/net/bna/bfa_ioc.h
@@ -26,16 +26,7 @@
#define BFA_IOC_TOV 3000 /* msecs */
#define BFA_IOC_HWSEM_TOV 500 /* msecs */
#define BFA_IOC_HB_TOV 500 /* msecs */
-#define BFA_IOC_HWINIT_MAX 2
-#define BFA_IOC_TOV_RECOVER BFA_IOC_HB_TOV
-
-/**
- * Generic Scatter Gather Element used by driver
- */
-struct bfa_sge {
- u32 sg_len;
- void *sg_addr;
-};
+#define BFA_IOC_HWINIT_MAX 5
/**
* PCI device information required by IOC
@@ -65,19 +56,6 @@ struct bfa_dma {
#define BFI_SMEM_CT_SIZE 0x280000U /* ! 2.5MB for catapult */
/**
- * @brief BFA dma address assignment macro
- */
-#define bfa_dma_addr_set(dma_addr, pa) \
- __bfa_dma_addr_set(&dma_addr, (u64)pa)
-
-static inline void
-__bfa_dma_addr_set(union bfi_addr_u *dma_addr, u64 pa)
-{
- dma_addr->a32.addr_lo = (u32) pa;
- dma_addr->a32.addr_hi = (u32) (upper_32_bits(pa));
-}
-
-/**
* @brief BFA dma address assignment macro. (big endian format)
*/
#define bfa_dma_be_addr_set(dma_addr, pa) \
@@ -105,8 +83,11 @@ struct bfa_ioc_regs {
void __iomem *host_page_num_fn;
void __iomem *heartbeat;
void __iomem *ioc_fwstate;
+ void __iomem *alt_ioc_fwstate;
void __iomem *ll_halt;
+ void __iomem *alt_ll_halt;
void __iomem *err_set;
+ void __iomem *ioc_fail_sync;
void __iomem *shirq_isr_next;
void __iomem *shirq_msk_next;
void __iomem *smem_page_start;
@@ -165,16 +146,22 @@ struct bfa_ioc_hbfail_notify {
(__notify)->cbarg = (__cbarg); \
} while (0)
+struct bfa_iocpf {
+ bfa_fsm_t fsm;
+ struct bfa_ioc *ioc;
+ u32 retry_count;
+ bool auto_recover;
+};
+
struct bfa_ioc {
bfa_fsm_t fsm;
struct bfa *bfa;
struct bfa_pcidev pcidev;
- struct bfa_timer_mod *timer_mod;
struct timer_list ioc_timer;
+ struct timer_list iocpf_timer;
struct timer_list sem_timer;
struct timer_list hb_timer;
u32 hb_count;
- u32 retry_count;
struct list_head hb_notify_q;
void *dbg_fwsave;
int dbg_fwsave_len;
@@ -182,7 +169,6 @@ struct bfa_ioc {
enum bfi_mclass ioc_mc;
struct bfa_ioc_regs ioc_regs;
struct bfa_ioc_drv_stats stats;
- bool auto_recover;
bool fcmode;
bool ctdev;
bool cna;
@@ -195,6 +181,7 @@ struct bfa_ioc {
struct bfa_ioc_cbfn *cbfn;
struct bfa_ioc_mbox_mod mbox_mod;
struct bfa_ioc_hwif *ioc_hwif;
+ struct bfa_iocpf iocpf;
};
struct bfa_ioc_hwif {
@@ -205,8 +192,12 @@ struct bfa_ioc_hwif {
void (*ioc_map_port) (struct bfa_ioc *ioc);
void (*ioc_isr_mode_set) (struct bfa_ioc *ioc,
bool msix);
- void (*ioc_notify_hbfail) (struct bfa_ioc *ioc);
+ void (*ioc_notify_fail) (struct bfa_ioc *ioc);
void (*ioc_ownership_reset) (struct bfa_ioc *ioc);
+ void (*ioc_sync_join) (struct bfa_ioc *ioc);
+ void (*ioc_sync_leave) (struct bfa_ioc *ioc);
+ void (*ioc_sync_ack) (struct bfa_ioc *ioc);
+ bool (*ioc_sync_complete) (struct bfa_ioc *ioc);
};
#define bfa_ioc_pcifn(__ioc) ((__ioc)->pcidev.pci_func)
@@ -271,7 +262,6 @@ void bfa_nw_ioc_enable(struct bfa_ioc *ioc);
void bfa_nw_ioc_disable(struct bfa_ioc *ioc);
void bfa_nw_ioc_error_isr(struct bfa_ioc *ioc);
-
void bfa_nw_ioc_get_attr(struct bfa_ioc *ioc, struct bfa_ioc_attr *ioc_attr);
void bfa_nw_ioc_hbfail_register(struct bfa_ioc *ioc,
struct bfa_ioc_hbfail_notify *notify);
@@ -289,7 +279,8 @@ mac_t bfa_nw_ioc_get_mac(struct bfa_ioc *ioc);
*/
void bfa_nw_ioc_timeout(void *ioc);
void bfa_nw_ioc_hb_check(void *ioc);
-void bfa_nw_ioc_sem_timeout(void *ioc);
+void bfa_nw_iocpf_timeout(void *ioc);
+void bfa_nw_iocpf_sem_timeout(void *ioc);
/*
* F/W Image Size & Chunk
diff --git a/drivers/net/bna/bfa_ioc_ct.c b/drivers/net/bna/bfa_ioc_ct.c
index 121cfd6..469997c 100644
--- a/drivers/net/bna/bfa_ioc_ct.c
+++ b/drivers/net/bna/bfa_ioc_ct.c
@@ -22,6 +22,15 @@
#include "bfi_ctreg.h"
#include "bfa_defs.h"
+#define bfa_ioc_ct_sync_pos(__ioc) \
+ ((u32) (1 << bfa_ioc_pcifn(__ioc)))
+#define BFA_IOC_SYNC_REQD_SH 16
+#define bfa_ioc_ct_get_sync_ackd(__val) (__val & 0x0000ffff)
+#define bfa_ioc_ct_clear_sync_ackd(__val) (__val & 0xffff0000)
+#define bfa_ioc_ct_get_sync_reqd(__val) (__val >> BFA_IOC_SYNC_REQD_SH)
+#define bfa_ioc_ct_sync_reqd_pos(__ioc) \
+ (bfa_ioc_ct_sync_pos(__ioc) << BFA_IOC_SYNC_REQD_SH)
+
/*
* forward declarations
*/
@@ -30,8 +39,12 @@ static void bfa_ioc_ct_firmware_unlock(struct bfa_ioc *ioc);
static void bfa_ioc_ct_reg_init(struct bfa_ioc *ioc);
static void bfa_ioc_ct_map_port(struct bfa_ioc *ioc);
static void bfa_ioc_ct_isr_mode_set(struct bfa_ioc *ioc, bool msix);
-static void bfa_ioc_ct_notify_hbfail(struct bfa_ioc *ioc);
+static void bfa_ioc_ct_notify_fail(struct bfa_ioc *ioc);
static void bfa_ioc_ct_ownership_reset(struct bfa_ioc *ioc);
+static void bfa_ioc_ct_sync_join(struct bfa_ioc *ioc);
+static void bfa_ioc_ct_sync_leave(struct bfa_ioc *ioc);
+static void bfa_ioc_ct_sync_ack(struct bfa_ioc *ioc);
+static bool bfa_ioc_ct_sync_complete(struct bfa_ioc *ioc);
static enum bfa_status bfa_ioc_ct_pll_init(void __iomem *rb, bool fcmode);
static struct bfa_ioc_hwif nw_hwif_ct;
@@ -48,8 +61,12 @@ bfa_nw_ioc_set_ct_hwif(struct bfa_ioc *ioc)
nw_hwif_ct.ioc_reg_init = bfa_ioc_ct_reg_init;
nw_hwif_ct.ioc_map_port = bfa_ioc_ct_map_port;
nw_hwif_ct.ioc_isr_mode_set = bfa_ioc_ct_isr_mode_set;
- nw_hwif_ct.ioc_notify_hbfail = bfa_ioc_ct_notify_hbfail;
+ nw_hwif_ct.ioc_notify_fail = bfa_ioc_ct_notify_fail;
nw_hwif_ct.ioc_ownership_reset = bfa_ioc_ct_ownership_reset;
+ nw_hwif_ct.ioc_sync_join = bfa_ioc_ct_sync_join;
+ nw_hwif_ct.ioc_sync_leave = bfa_ioc_ct_sync_leave;
+ nw_hwif_ct.ioc_sync_ack = bfa_ioc_ct_sync_ack;
+ nw_hwif_ct.ioc_sync_complete = bfa_ioc_ct_sync_complete;
ioc->ioc_hwif = &nw_hwif_ct;
}
@@ -86,6 +103,7 @@ bfa_ioc_ct_firmware_lock(struct bfa_ioc *ioc)
if (usecnt == 0) {
writel(1, ioc->ioc_regs.ioc_usage_reg);
bfa_nw_ioc_sem_release(ioc->ioc_regs.ioc_usage_sem_reg);
+ writel(0, ioc->ioc_regs.ioc_fail_sync);
return true;
}
@@ -149,12 +167,14 @@ bfa_ioc_ct_firmware_unlock(struct bfa_ioc *ioc)
* Notify other functions on HB failure.
*/
static void
-bfa_ioc_ct_notify_hbfail(struct bfa_ioc *ioc)
+bfa_ioc_ct_notify_fail(struct bfa_ioc *ioc)
{
if (ioc->cna) {
writel(__FW_INIT_HALT_P, ioc->ioc_regs.ll_halt);
+ writel(__FW_INIT_HALT_P, ioc->ioc_regs.alt_ll_halt);
/* Wait for halt to take effect */
readl(ioc->ioc_regs.ll_halt);
+ readl(ioc->ioc_regs.alt_ll_halt);
} else {
writel(__PSS_ERR_STATUS_SET, ioc->ioc_regs.err_set);
readl(ioc->ioc_regs.err_set);
@@ -206,15 +226,19 @@ bfa_ioc_ct_reg_init(struct bfa_ioc *ioc)
if (ioc->port_id == 0) {
ioc->ioc_regs.heartbeat = rb + BFA_IOC0_HBEAT_REG;
ioc->ioc_regs.ioc_fwstate = rb + BFA_IOC0_STATE_REG;
+ ioc->ioc_regs.alt_ioc_fwstate = rb + BFA_IOC1_STATE_REG;
ioc->ioc_regs.hfn_mbox_cmd = rb + iocreg_mbcmd_p0[pcifn].hfn;
ioc->ioc_regs.lpu_mbox_cmd = rb + iocreg_mbcmd_p0[pcifn].lpu;
ioc->ioc_regs.ll_halt = rb + FW_INIT_HALT_P0;
+ ioc->ioc_regs.alt_ll_halt = rb + FW_INIT_HALT_P1;
} else {
ioc->ioc_regs.heartbeat = (rb + BFA_IOC1_HBEAT_REG);
ioc->ioc_regs.ioc_fwstate = (rb + BFA_IOC1_STATE_REG);
+ ioc->ioc_regs.alt_ioc_fwstate = rb + BFA_IOC0_STATE_REG;
ioc->ioc_regs.hfn_mbox_cmd = rb + iocreg_mbcmd_p1[pcifn].hfn;
ioc->ioc_regs.lpu_mbox_cmd = rb + iocreg_mbcmd_p1[pcifn].lpu;
ioc->ioc_regs.ll_halt = rb + FW_INIT_HALT_P1;
+ ioc->ioc_regs.alt_ll_halt = rb + FW_INIT_HALT_P0;
}
/*
@@ -232,6 +256,7 @@ bfa_ioc_ct_reg_init(struct bfa_ioc *ioc)
ioc->ioc_regs.ioc_usage_sem_reg = (rb + HOST_SEM1_REG);
ioc->ioc_regs.ioc_init_sem_reg = (rb + HOST_SEM2_REG);
ioc->ioc_regs.ioc_usage_reg = (rb + BFA_FW_USE_COUNT);
+ ioc->ioc_regs.ioc_fail_sync = (rb + BFA_IOC_FAIL_SYNC);
/**
* sram memory access
@@ -317,6 +342,77 @@ bfa_ioc_ct_ownership_reset(struct bfa_ioc *ioc)
bfa_nw_ioc_hw_sem_release(ioc);
}
+/**
+ * Synchronized IOC failure processing routines
+ */
+static void
+bfa_ioc_ct_sync_join(struct bfa_ioc *ioc)
+{
+ u32 r32 = readl(ioc->ioc_regs.ioc_fail_sync);
+ u32 sync_pos = bfa_ioc_ct_sync_reqd_pos(ioc);
+
+ writel((r32 | sync_pos), ioc->ioc_regs.ioc_fail_sync);
+}
+
+static void
+bfa_ioc_ct_sync_leave(struct bfa_ioc *ioc)
+{
+ u32 r32 = readl(ioc->ioc_regs.ioc_fail_sync);
+ u32 sync_msk = bfa_ioc_ct_sync_reqd_pos(ioc) |
+ bfa_ioc_ct_sync_pos(ioc);
+
+ writel((r32 & ~sync_msk), ioc->ioc_regs.ioc_fail_sync);
+}
+
+static void
+bfa_ioc_ct_sync_ack(struct bfa_ioc *ioc)
+{
+ u32 r32 = readl(ioc->ioc_regs.ioc_fail_sync);
+
+ writel((r32 | bfa_ioc_ct_sync_pos(ioc)), ioc->ioc_regs.ioc_fail_sync);
+}
+
+static bool
+bfa_ioc_ct_sync_complete(struct bfa_ioc *ioc)
+{
+ u32 r32 = readl(ioc->ioc_regs.ioc_fail_sync);
+ u32 sync_reqd = bfa_ioc_ct_get_sync_reqd(r32);
+ u32 sync_ackd = bfa_ioc_ct_get_sync_ackd(r32);
+ u32 tmp_ackd;
+
+ if (sync_ackd == 0)
+ return true;
+
+ /**
+ * The check below is to see whether any other PCI fn
+ * has reinitialized the ASIC (reset sync_ackd bits)
+ * and failed again while this IOC was waiting for hw
+ * semaphore (in bfa_iocpf_sm_semwait()).
+ */
+ tmp_ackd = sync_ackd;
+ if ((sync_reqd & bfa_ioc_ct_sync_pos(ioc)) &&
+ !(sync_ackd & bfa_ioc_ct_sync_pos(ioc)))
+ sync_ackd |= bfa_ioc_ct_sync_pos(ioc);
+
+ if (sync_reqd == sync_ackd) {
+ writel(bfa_ioc_ct_clear_sync_ackd(r32),
+ ioc->ioc_regs.ioc_fail_sync);
+ writel(BFI_IOC_FAIL, ioc->ioc_regs.ioc_fwstate);
+ writel(BFI_IOC_FAIL, ioc->ioc_regs.alt_ioc_fwstate);
+ return true;
+ }
+
+ /**
+ * If another PCI fn reinitialized and failed again while
+ * this IOC was waiting for hw sem, the sync_ackd bit for
+ * this IOC need to be set again to allow reinitialization.
+ */
+ if (tmp_ackd != sync_ackd)
+ writel((r32 | sync_ackd), ioc->ioc_regs.ioc_fail_sync);
+
+ return false;
+}
+
static enum bfa_status
bfa_ioc_ct_pll_init(void __iomem *rb, bool fcmode)
{
diff --git a/drivers/net/bna/bfi_ctreg.h b/drivers/net/bna/bfi_ctreg.h
index 404ea35..5130d79 100644
--- a/drivers/net/bna/bfi_ctreg.h
+++ b/drivers/net/bna/bfi_ctreg.h
@@ -535,6 +535,7 @@ enum {
#define BFA_IOC1_HBEAT_REG HOST_SEM2_INFO_REG
#define BFA_IOC1_STATE_REG HOST_SEM3_INFO_REG
#define BFA_FW_USE_COUNT HOST_SEM4_INFO_REG
+#define BFA_IOC_FAIL_SYNC HOST_SEM5_INFO_REG
#define CPE_DEPTH_Q(__n) \
(CPE_DEPTH_Q0 + (__n) * (CPE_DEPTH_Q1 - CPE_DEPTH_Q0))
@@ -552,22 +553,30 @@ enum {
(RME_PI_PTR_Q0 + (__n) * (RME_PI_PTR_Q1 - RME_PI_PTR_Q0))
#define RME_CI_PTR_Q(__n) \
(RME_CI_PTR_Q0 + (__n) * (RME_CI_PTR_Q1 - RME_CI_PTR_Q0))
-#define HQM_QSET_RXQ_DRBL_P0(__n) (HQM_QSET0_RXQ_DRBL_P0 + (__n) \
- * (HQM_QSET1_RXQ_DRBL_P0 - HQM_QSET0_RXQ_DRBL_P0))
-#define HQM_QSET_TXQ_DRBL_P0(__n) (HQM_QSET0_TXQ_DRBL_P0 + (__n) \
- * (HQM_QSET1_TXQ_DRBL_P0 - HQM_QSET0_TXQ_DRBL_P0))
-#define HQM_QSET_IB_DRBL_1_P0(__n) (HQM_QSET0_IB_DRBL_1_P0 + (__n) \
- * (HQM_QSET1_IB_DRBL_1_P0 - HQM_QSET0_IB_DRBL_1_P0))
-#define HQM_QSET_IB_DRBL_2_P0(__n) (HQM_QSET0_IB_DRBL_2_P0 + (__n) \
- * (HQM_QSET1_IB_DRBL_2_P0 - HQM_QSET0_IB_DRBL_2_P0))
-#define HQM_QSET_RXQ_DRBL_P1(__n) (HQM_QSET0_RXQ_DRBL_P1 + (__n) \
- * (HQM_QSET1_RXQ_DRBL_P1 - HQM_QSET0_RXQ_DRBL_P1))
-#define HQM_QSET_TXQ_DRBL_P1(__n) (HQM_QSET0_TXQ_DRBL_P1 + (__n) \
- * (HQM_QSET1_TXQ_DRBL_P1 - HQM_QSET0_TXQ_DRBL_P1))
-#define HQM_QSET_IB_DRBL_1_P1(__n) (HQM_QSET0_IB_DRBL_1_P1 + (__n) \
- * (HQM_QSET1_IB_DRBL_1_P1 - HQM_QSET0_IB_DRBL_1_P1))
-#define HQM_QSET_IB_DRBL_2_P1(__n) (HQM_QSET0_IB_DRBL_2_P1 + (__n) \
- * (HQM_QSET1_IB_DRBL_2_P1 - HQM_QSET0_IB_DRBL_2_P1))
+#define HQM_QSET_RXQ_DRBL_P0(__n) \
+ (HQM_QSET0_RXQ_DRBL_P0 + (__n) * \
+ (HQM_QSET1_RXQ_DRBL_P0 - HQM_QSET0_RXQ_DRBL_P0))
+#define HQM_QSET_TXQ_DRBL_P0(__n) \
+ (HQM_QSET0_TXQ_DRBL_P0 + (__n) * \
+ (HQM_QSET1_TXQ_DRBL_P0 - HQM_QSET0_TXQ_DRBL_P0))
+#define HQM_QSET_IB_DRBL_1_P0(__n) \
+ (HQM_QSET0_IB_DRBL_1_P0 + (__n) * \
+ (HQM_QSET1_IB_DRBL_1_P0 - HQM_QSET0_IB_DRBL_1_P0))
+#define HQM_QSET_IB_DRBL_2_P0(__n) \
+ (HQM_QSET0_IB_DRBL_2_P0 + (__n) * \
+ (HQM_QSET1_IB_DRBL_2_P0 - HQM_QSET0_IB_DRBL_2_P0))
+#define HQM_QSET_RXQ_DRBL_P1(__n) \
+ (HQM_QSET0_RXQ_DRBL_P1 + (__n) * \
+ (HQM_QSET1_RXQ_DRBL_P1 - HQM_QSET0_RXQ_DRBL_P1))
+#define HQM_QSET_TXQ_DRBL_P1(__n) \
+ (HQM_QSET0_TXQ_DRBL_P1 + (__n) * \
+ (HQM_QSET1_TXQ_DRBL_P1 - HQM_QSET0_TXQ_DRBL_P1))
+#define HQM_QSET_IB_DRBL_1_P1(__n) \
+ (HQM_QSET0_IB_DRBL_1_P1 + (__n) * \
+ (HQM_QSET1_IB_DRBL_1_P1 - HQM_QSET0_IB_DRBL_1_P1))
+#define HQM_QSET_IB_DRBL_2_P1(__n) \
+ (HQM_QSET0_IB_DRBL_2_P1 + (__n) * \
+ (HQM_QSET1_IB_DRBL_2_P1 - HQM_QSET0_IB_DRBL_2_P1))
#define CPE_Q_NUM(__fn, __q) (((__fn) << 2) + (__q))
#define RME_Q_NUM(__fn, __q) (((__fn) << 2) + (__q))
diff --git a/drivers/net/bna/bna.h b/drivers/net/bna/bna.h
index fd93f76..a287f89 100644
--- a/drivers/net/bna/bna.h
+++ b/drivers/net/bna/bna.h
@@ -32,8 +32,6 @@ extern const u32 bna_napi_dim_vector[][BNA_BIAS_T_MAX];
/* Log string size */
#define BNA_MESSAGE_SIZE 256
-#define bna_device_timer(_dev) bfa_timer_beat(&((_dev)->timer_mod))
-
/* MBOX API for PORT, TX, RX */
#define bna_mbox_qe_fill(_qe, _cmd, _cmd_len, _cbfn, _cbarg) \
do { \
diff --git a/drivers/net/bna/bnad.c b/drivers/net/bna/bnad.c
index 200df60..277dbe0 100644
--- a/drivers/net/bna/bnad.c
+++ b/drivers/net/bna/bnad.c
@@ -1425,13 +1425,24 @@ bnad_ioc_hb_check(unsigned long data)
}
static void
-bnad_ioc_sem_timeout(unsigned long data)
+bnad_iocpf_timeout(unsigned long data)
{
struct bnad *bnad = (struct bnad *)data;
unsigned long flags;
spin_lock_irqsave(&bnad->bna_lock, flags);
- bfa_nw_ioc_sem_timeout((void *) &bnad->bna.device.ioc);
+ bfa_nw_iocpf_timeout((void *) &bnad->bna.device.ioc);
+ spin_unlock_irqrestore(&bnad->bna_lock, flags);
+}
+
+static void
+bnad_iocpf_sem_timeout(unsigned long data)
+{
+ struct bnad *bnad = (struct bnad *)data;
+ unsigned long flags;
+
+ spin_lock_irqsave(&bnad->bna_lock, flags);
+ bfa_nw_iocpf_sem_timeout((void *) &bnad->bna.device.ioc);
spin_unlock_irqrestore(&bnad->bna_lock, flags);
}
@@ -3132,11 +3143,13 @@ bnad_pci_probe(struct pci_dev *pdev,
((unsigned long)bnad));
setup_timer(&bnad->bna.device.ioc.hb_timer, bnad_ioc_hb_check,
((unsigned long)bnad));
- setup_timer(&bnad->bna.device.ioc.sem_timer, bnad_ioc_sem_timeout,
+ setup_timer(&bnad->bna.device.ioc.iocpf_timer, bnad_iocpf_timeout,
+ ((unsigned long)bnad));
+ setup_timer(&bnad->bna.device.ioc.sem_timer, bnad_iocpf_sem_timeout,
((unsigned long)bnad));
/* Now start the timer before calling IOC */
- mod_timer(&bnad->bna.device.ioc.ioc_timer,
+ mod_timer(&bnad->bna.device.ioc.iocpf_timer,
jiffies + msecs_to_jiffies(BNA_IOC_TIMER_FREQ));
/*
--
1.7.1
^ permalink raw reply related
* Re: [PATCH net-next-2.6 v2 1/1] can: c_can: Added support for Bosch C_CAN controller
From: Wolfgang Grandegger @ 2010-12-22 7:50 UTC (permalink / raw)
To: Marc Kleine-Budde
Cc: Socketcan-core-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org,
netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
In-Reply-To: <4D11A039.5090006-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
On 12/22/2010 07:52 AM, Marc Kleine-Budde wrote:
> On 12/22/2010 04:36 AM, Bhupesh SHARMA wrote:
>> Hi Wolfgang,
>>
>>> Hi Bhupesh,
>>>
>>> On 12/21/2010 05:48 AM, Bhupesh SHARMA wrote:
>>>> Hi Wolfgang,
>>> ...
>>>>> In the meantime I compared the CAN chapter of the PCH manual with
>>> the
>>>>> C_CAN manual. The paragraphs I checked are *identical*. This makes
>>>>> clear, that the "pch_can" is a clone of the C_CAN CAN controller,
>>> with
>>>>> a few extensions, though. Therefore it would make sense, to
>>> implement a
>>>>> bus sensitive interface like for the SJA1000 allowing to handle both
>>>>> CAN
>>>>> controllers with one driver sooner than later. Therefore, could you
>>>>> please implement:
>>>>>
>>>>> drivers/net/can/c_can/c_can.c
>>>>> /c_can_platform.c
>>>>>
>>>>> Then an interface to the PCI based PCH CAN controller could be added
>>>>> easily, e.g. as "pch_pci.c". You already had something similar in
>>> your
>>>>> RFC version of the patch, IIRC.
>>>>
>>>> This was the approach I initially proposed in my RFC V1 patch :)
>>>> But unfortunately we could not agree to it.
>>>
>>> I know. But at that time I was not aware of any other bus used for the
>>> C_CAN controller.
>>>
>>>> So, please let me reiterate what I understood and what was present
>>>> in RFC version of the patch. Please add your comments/views:
>>>>
>>>> - drivers/net/can/c_can/c_can.c (similar on lines of
>>> sja1000.c)
>>>> i.e. a)no *probe* / *remove* functions here,
>>>> b)register read/write implemented here.
>>>>
>>>> - drivers/net/can/c_can/c_can_platform.c (similar on lines of
>>> sja1000_platform.c)
>>>> i.e. *probe* / *remove* implemented here,
>>>
>>> Yes, that's what I'm thinking about.
>>>
>>>> Marc and Tomoya can also add their suggestions so that I can finalize
>>> V3 a.s.a.p.
>>>
>>> That would be nice, indeed. Also have a look to Tomoya's PCH driver,
>>> which also looks very good in the meantime.
>>
>> I am having a look at Tomoya's PCH driver, but as I mentioned in
>> RFC V1 patch, I would rather like to have a bus sensitive `c_can` driver
>
> What do you mean by a "bus sensitive" driver?
I was thinking about a "bus independent interface" like for the SJA1000.
A bus sensitive driver would the be in c_can_platform.c.
>> on top of which we can have the platform driver `c_can_platform` which
>> essentially caters to the details of registers mapping/arch differences.
>> Any other functionality like USB/PCI should be present in a separate file
>> like `usb_c_can.c` or `pci_c_can.c`
>
> Sounds like the sja1000 approach, which is a good choice.
I fully agree.
>> If you agree I will try to circulate V3 a.s.ap.
>
> go ahead.
>
Yes, please.
Wolfgang.
^ permalink raw reply
* [PATCH 02/10] bna: Port enable disable sync and txq priority fix
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody, Debashis Dutt
In-Reply-To: <1293002596-3239-1-git-send-email-rmody@brocade.com>
Change details:
- Fixed port enable/disable sync through a change in LL port state machine.
- Change txq->priority masking to 0x7 (3 bits) from 0x3 (2 bits).
Signed-off-by: Debashis Dutt <ddutt@brocade.com>
Signed-off-by: Rasesh Mody <rmody@brocade.com>
---
drivers/net/bna/bna.h | 4 +-
drivers/net/bna/bna_ctrl.c | 136 +++++++++++++++++++++++++++++++++----------
drivers/net/bna/bna_txrx.c | 8 +-
drivers/net/bna/bna_types.h | 7 +-
4 files changed, 116 insertions(+), 39 deletions(-)
diff --git a/drivers/net/bna/bna.h b/drivers/net/bna/bna.h
index df6676b..fd93f76 100644
--- a/drivers/net/bna/bna.h
+++ b/drivers/net/bna/bna.h
@@ -390,8 +390,8 @@ void bna_mbox_send(struct bna *bna, struct bna_mbox_qe *mbox_qe);
/* API for RX */
int bna_port_mtu_get(struct bna_port *port);
-void bna_llport_admin_up(struct bna_llport *llport);
-void bna_llport_admin_down(struct bna_llport *llport);
+void bna_llport_rx_started(struct bna_llport *llport);
+void bna_llport_rx_stopped(struct bna_llport *llport);
/* API for BNAD */
void bna_port_enable(struct bna_port *port);
diff --git a/drivers/net/bna/bna_ctrl.c b/drivers/net/bna/bna_ctrl.c
index 07b2659..68e4c5e 100644
--- a/drivers/net/bna/bna_ctrl.c
+++ b/drivers/net/bna/bna_ctrl.c
@@ -59,14 +59,70 @@ bna_port_cb_link_down(struct bna_port *port, int status)
port->link_cbfn(port->bna->bnad, BNA_LINK_DOWN);
}
+static inline int
+llport_can_be_up(struct bna_llport *llport)
+{
+ int ready = 0;
+ if (llport->type == BNA_PORT_T_REGULAR)
+ ready = ((llport->flags & BNA_LLPORT_F_ADMIN_UP) &&
+ (llport->flags & BNA_LLPORT_F_RX_STARTED) &&
+ (llport->flags & BNA_LLPORT_F_PORT_ENABLED));
+ else
+ ready = ((llport->flags & BNA_LLPORT_F_ADMIN_UP) &&
+ (llport->flags & BNA_LLPORT_F_RX_STARTED) &&
+ !(llport->flags & BNA_LLPORT_F_PORT_ENABLED));
+ return ready;
+}
+
+#define llport_is_up llport_can_be_up
+
+enum bna_llport_event {
+ LLPORT_E_START = 1,
+ LLPORT_E_STOP = 2,
+ LLPORT_E_FAIL = 3,
+ LLPORT_E_UP = 4,
+ LLPORT_E_DOWN = 5,
+ LLPORT_E_FWRESP_UP_OK = 6,
+ LLPORT_E_FWRESP_UP_FAIL = 7,
+ LLPORT_E_FWRESP_DOWN = 8
+};
+
+static void
+bna_llport_cb_port_enabled(struct bna_llport *llport)
+{
+ llport->flags |= BNA_LLPORT_F_PORT_ENABLED;
+
+ if (llport_can_be_up(llport))
+ bfa_fsm_send_event(llport, LLPORT_E_UP);
+}
+
+static void
+bna_llport_cb_port_disabled(struct bna_llport *llport)
+{
+ int llport_up = llport_is_up(llport);
+
+ llport->flags &= ~BNA_LLPORT_F_PORT_ENABLED;
+
+ if (llport_up)
+ bfa_fsm_send_event(llport, LLPORT_E_DOWN);
+}
+
/**
* MBOX
*/
static int
bna_is_aen(u8 msg_id)
{
- return msg_id == BFI_LL_I2H_LINK_DOWN_AEN ||
- msg_id == BFI_LL_I2H_LINK_UP_AEN;
+ switch (msg_id) {
+ case BFI_LL_I2H_LINK_DOWN_AEN:
+ case BFI_LL_I2H_LINK_UP_AEN:
+ case BFI_LL_I2H_PORT_ENABLE_AEN:
+ case BFI_LL_I2H_PORT_DISABLE_AEN:
+ return 1;
+
+ default:
+ return 0;
+ }
}
static void
@@ -81,6 +137,12 @@ bna_mbox_aen_callback(struct bna *bna, struct bfi_mbmsg *msg)
case BFI_LL_I2H_LINK_DOWN_AEN:
bna_port_cb_link_down(&bna->port, aen->reason);
break;
+ case BFI_LL_I2H_PORT_ENABLE_AEN:
+ bna_llport_cb_port_enabled(&bna->port.llport);
+ break;
+ case BFI_LL_I2H_PORT_DISABLE_AEN:
+ bna_llport_cb_port_disabled(&bna->port.llport);
+ break;
default:
break;
}
@@ -251,16 +313,6 @@ static void bna_llport_start(struct bna_llport *llport);
static void bna_llport_stop(struct bna_llport *llport);
static void bna_llport_fail(struct bna_llport *llport);
-enum bna_llport_event {
- LLPORT_E_START = 1,
- LLPORT_E_STOP = 2,
- LLPORT_E_FAIL = 3,
- LLPORT_E_UP = 4,
- LLPORT_E_DOWN = 5,
- LLPORT_E_FWRESP_UP = 6,
- LLPORT_E_FWRESP_DOWN = 7
-};
-
enum bna_llport_state {
BNA_LLPORT_STOPPED = 1,
BNA_LLPORT_DOWN = 2,
@@ -320,7 +372,7 @@ bna_llport_sm_stopped(struct bna_llport *llport,
/* No-op */
break;
- case LLPORT_E_FWRESP_UP:
+ case LLPORT_E_FWRESP_UP_OK:
case LLPORT_E_FWRESP_DOWN:
/**
* These events are received due to flushing of mbox when
@@ -366,6 +418,7 @@ bna_llport_sm_down(struct bna_llport *llport,
static void
bna_llport_sm_up_resp_wait_entry(struct bna_llport *llport)
{
+ BUG_ON(!llport_can_be_up(llport));
/**
* NOTE: Do not call bna_fw_llport_up() here. That will over step
* mbox due to down_resp_wait -> up_resp_wait transition on event
@@ -390,10 +443,14 @@ bna_llport_sm_up_resp_wait(struct bna_llport *llport,
bfa_fsm_set_state(llport, bna_llport_sm_down_resp_wait);
break;
- case LLPORT_E_FWRESP_UP:
+ case LLPORT_E_FWRESP_UP_OK:
bfa_fsm_set_state(llport, bna_llport_sm_up);
break;
+ case LLPORT_E_FWRESP_UP_FAIL:
+ bfa_fsm_set_state(llport, bna_llport_sm_down);
+ break;
+
case LLPORT_E_FWRESP_DOWN:
/* down_resp_wait -> up_resp_wait transition on LLPORT_E_UP */
bna_fw_llport_up(llport);
@@ -431,11 +488,12 @@ bna_llport_sm_down_resp_wait(struct bna_llport *llport,
bfa_fsm_set_state(llport, bna_llport_sm_up_resp_wait);
break;
- case LLPORT_E_FWRESP_UP:
+ case LLPORT_E_FWRESP_UP_OK:
/* up_resp_wait->down_resp_wait transition on LLPORT_E_DOWN */
bna_fw_llport_down(llport);
break;
+ case LLPORT_E_FWRESP_UP_FAIL:
case LLPORT_E_FWRESP_DOWN:
bfa_fsm_set_state(llport, bna_llport_sm_down);
break;
@@ -496,11 +554,12 @@ bna_llport_sm_last_resp_wait(struct bna_llport *llport,
/* No-op */
break;
- case LLPORT_E_FWRESP_UP:
+ case LLPORT_E_FWRESP_UP_OK:
/* up_resp_wait->last_resp_wait transition on LLPORT_T_STOP */
bna_fw_llport_down(llport);
break;
+ case LLPORT_E_FWRESP_UP_FAIL:
case LLPORT_E_FWRESP_DOWN:
bfa_fsm_set_state(llport, bna_llport_sm_stopped);
break;
@@ -541,7 +600,14 @@ bna_fw_cb_llport_up(void *arg, int status)
struct bna_llport *llport = (struct bna_llport *)arg;
bfa_q_qe_init(&llport->mbox_qe.qe);
- bfa_fsm_send_event(llport, LLPORT_E_FWRESP_UP);
+ if (status == BFI_LL_CMD_FAIL) {
+ if (llport->type == BNA_PORT_T_REGULAR)
+ llport->flags &= ~BNA_LLPORT_F_PORT_ENABLED;
+ else
+ llport->flags &= ~BNA_LLPORT_F_ADMIN_UP;
+ bfa_fsm_send_event(llport, LLPORT_E_FWRESP_UP_FAIL);
+ } else
+ bfa_fsm_send_event(llport, LLPORT_E_FWRESP_UP_OK);
}
static void
@@ -588,13 +654,14 @@ bna_port_cb_llport_stopped(struct bna_port *port,
static void
bna_llport_init(struct bna_llport *llport, struct bna *bna)
{
- llport->flags |= BNA_LLPORT_F_ENABLED;
+ llport->flags |= BNA_LLPORT_F_ADMIN_UP;
+ llport->flags |= BNA_LLPORT_F_PORT_ENABLED;
llport->type = BNA_PORT_T_REGULAR;
llport->bna = bna;
llport->link_status = BNA_LINK_DOWN;
- llport->admin_up_count = 0;
+ llport->rx_started_count = 0;
llport->stop_cbfn = NULL;
@@ -606,7 +673,8 @@ bna_llport_init(struct bna_llport *llport, struct bna *bna)
static void
bna_llport_uninit(struct bna_llport *llport)
{
- llport->flags &= ~BNA_LLPORT_F_ENABLED;
+ llport->flags &= ~BNA_LLPORT_F_ADMIN_UP;
+ llport->flags &= ~BNA_LLPORT_F_PORT_ENABLED;
llport->bna = NULL;
}
@@ -628,6 +696,8 @@ bna_llport_stop(struct bna_llport *llport)
static void
bna_llport_fail(struct bna_llport *llport)
{
+ /* Reset the physical port status to enabled */
+ llport->flags |= BNA_LLPORT_F_PORT_ENABLED;
bfa_fsm_send_event(llport, LLPORT_E_FAIL);
}
@@ -638,25 +708,31 @@ bna_llport_state_get(struct bna_llport *llport)
}
void
-bna_llport_admin_up(struct bna_llport *llport)
+bna_llport_rx_started(struct bna_llport *llport)
{
- llport->admin_up_count++;
+ llport->rx_started_count++;
+
+ if (llport->rx_started_count == 1) {
+
+ llport->flags |= BNA_LLPORT_F_RX_STARTED;
- if (llport->admin_up_count == 1) {
- llport->flags |= BNA_LLPORT_F_RX_ENABLED;
- if (llport->flags & BNA_LLPORT_F_ENABLED)
+ if (llport_can_be_up(llport))
bfa_fsm_send_event(llport, LLPORT_E_UP);
}
}
void
-bna_llport_admin_down(struct bna_llport *llport)
+bna_llport_rx_stopped(struct bna_llport *llport)
{
- llport->admin_up_count--;
+ int llport_up = llport_is_up(llport);
+
+ llport->rx_started_count--;
+
+ if (llport->rx_started_count == 0) {
+
+ llport->flags &= ~BNA_LLPORT_F_RX_STARTED;
- if (llport->admin_up_count == 0) {
- llport->flags &= ~BNA_LLPORT_F_RX_ENABLED;
- if (llport->flags & BNA_LLPORT_F_ENABLED)
+ if (llport_up)
bfa_fsm_send_event(llport, LLPORT_E_DOWN);
}
}
diff --git a/drivers/net/bna/bna_txrx.c b/drivers/net/bna/bna_txrx.c
index ad93fdb..fb6cf1f 100644
--- a/drivers/net/bna/bna_txrx.c
+++ b/drivers/net/bna/bna_txrx.c
@@ -1947,7 +1947,7 @@ bna_rx_sm_started_entry(struct bna_rx *rx)
bna_ib_ack(&rxp->cq.ib->door_bell, 0);
}
- bna_llport_admin_up(&rx->bna->port.llport);
+ bna_llport_rx_started(&rx->bna->port.llport);
}
void
@@ -1955,13 +1955,13 @@ bna_rx_sm_started(struct bna_rx *rx, enum bna_rx_event event)
{
switch (event) {
case RX_E_FAIL:
- bna_llport_admin_down(&rx->bna->port.llport);
+ bna_llport_rx_stopped(&rx->bna->port.llport);
bfa_fsm_set_state(rx, bna_rx_sm_stopped);
rx_ib_fail(rx);
bna_rxf_fail(&rx->rxf);
break;
case RX_E_STOP:
- bna_llport_admin_down(&rx->bna->port.llport);
+ bna_llport_rx_stopped(&rx->bna->port.llport);
bfa_fsm_set_state(rx, bna_rx_sm_rxf_stop_wait);
break;
default:
@@ -3373,7 +3373,7 @@ __bna_txq_start(struct bna_tx *tx, struct bna_txq *txq)
txq_cfg.cns_ptr2_n_q_state = BNA_Q_IDLE_STATE;
txq_cfg.nxt_qid_n_fid_n_pri = (((tx->txf.txf_id & 0x3f) << 3) |
- (txq->priority & 0x3));
+ (txq->priority & 0x7));
txq_cfg.wvc_n_cquota_n_rquota =
((((u32)BFI_TX_MAX_WRR_QUOTA & 0xfff) << 12) |
(BFI_TX_MAX_WRR_QUOTA & 0xfff));
diff --git a/drivers/net/bna/bna_types.h b/drivers/net/bna/bna_types.h
index 6877310..d79fd98 100644
--- a/drivers/net/bna/bna_types.h
+++ b/drivers/net/bna/bna_types.h
@@ -249,8 +249,9 @@ enum bna_link_status {
};
enum bna_llport_flags {
- BNA_LLPORT_F_ENABLED = 1,
- BNA_LLPORT_F_RX_ENABLED = 2
+ BNA_LLPORT_F_ADMIN_UP = 1,
+ BNA_LLPORT_F_PORT_ENABLED = 2,
+ BNA_LLPORT_F_RX_STARTED = 4
};
enum bna_port_flags {
@@ -405,7 +406,7 @@ struct bna_llport {
enum bna_link_status link_status;
- int admin_up_count;
+ int rx_started_count;
void (*stop_cbfn)(struct bna_port *, enum bna_cb_status);
--
1.7.1
^ permalink raw reply related
* [PATCH 00/10] bna: Update Brocade BNA Ethernet driver to v2.3.2.3
From: Rasesh Mody @ 2010-12-22 7:23 UTC (permalink / raw)
To: netdev; +Cc: Rasesh Mody
The following patch set updates the Brocade BNA driver to v2.3.2.3.
The patches are generated, compiled and tested against net-next-2.6 (2.6.37-rc1).
Rasesh Mody (10):
bna: TxRx and datapath fix
bna: Port enable disable sync and txq priority fix
bna: Fix ethtool register dump and reordered an API
bna: Enable pure priority tagged packet reception and rxf uninit
cleanup fix
bna: Fix for TX queue
bna: IOC uninit check and misc cleanup
bna: Removed unused code
bna: Restore VLAN filter table
bna: IOC failure auto recovery fix
bna: Update the driver version to 2.3.2.3
drivers/net/bna/bfa_defs.h | 22 +-
drivers/net/bna/bfa_defs_mfg_comm.h | 22 -
drivers/net/bna/bfa_ioc.c | 1219 +++++++++++++++++++++++++----------
drivers/net/bna/bfa_ioc.h | 49 +-
drivers/net/bna/bfa_ioc_ct.c | 102 +++-
drivers/net/bna/bfi_ctreg.h | 41 +-
drivers/net/bna/bna.h | 6 +-
drivers/net/bna/bna_ctrl.c | 377 +++--------
drivers/net/bna/bna_txrx.c | 44 +-
drivers/net/bna/bna_types.h | 11 +-
drivers/net/bna/bnad.c | 427 +++++++------
drivers/net/bna/bnad.h | 37 +-
drivers/net/bna/bnad_ethtool.c | 8 +-
13 files changed, 1445 insertions(+), 920 deletions(-)
^ permalink raw reply
* RE: [PATCH net-next-2.6 v2 1/1] can: c_can: Added support for Bosch C_CAN controller
From: Bhupesh SHARMA @ 2010-12-22 8:03 UTC (permalink / raw)
To: Wolfgang Grandegger, Marc Kleine-Budde
Cc: Socketcan-core-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org,
netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
In-Reply-To: <4D11ADD2.40506-5Yr1BZd7O62+XT7JhA+gdA@public.gmane.org>
Hi Marc and Wolfgang,
> -----Original Message-----
> From: Wolfgang Grandegger [mailto:wg-5Yr1BZd7O62+XT7JhA+gdA@public.gmane.org]
> Sent: Wednesday, December 22, 2010 1:21 PM
> To: Marc Kleine-Budde
> Cc: Bhupesh SHARMA; Socketcan-core-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org;
> netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> Subject: Re: [PATCH net-next-2.6 v2 1/1] can: c_can: Added support for
> Bosch C_CAN controller
>
> On 12/22/2010 07:52 AM, Marc Kleine-Budde wrote:
> > On 12/22/2010 04:36 AM, Bhupesh SHARMA wrote:
> >> Hi Wolfgang,
> >>
> >>> Hi Bhupesh,
> >>>
> >>> On 12/21/2010 05:48 AM, Bhupesh SHARMA wrote:
> >>>> Hi Wolfgang,
> >>> ...
> >>>>> In the meantime I compared the CAN chapter of the PCH manual with
> >>> the
> >>>>> C_CAN manual. The paragraphs I checked are *identical*. This
> makes
> >>>>> clear, that the "pch_can" is a clone of the C_CAN CAN
> controller,
> >>> with
> >>>>> a few extensions, though. Therefore it would make sense, to
> >>> implement a
> >>>>> bus sensitive interface like for the SJA1000 allowing to handle
> both
> >>>>> CAN
> >>>>> controllers with one driver sooner than later. Therefore, could
> you
> >>>>> please implement:
> >>>>>
> >>>>> drivers/net/can/c_can/c_can.c
> >>>>> /c_can_platform.c
> >>>>>
> >>>>> Then an interface to the PCI based PCH CAN controller could be
> added
> >>>>> easily, e.g. as "pch_pci.c". You already had something similar in
> >>> your
> >>>>> RFC version of the patch, IIRC.
> >>>>
> >>>> This was the approach I initially proposed in my RFC V1 patch :)
> >>>> But unfortunately we could not agree to it.
> >>>
> >>> I know. But at that time I was not aware of any other bus used for
> the
> >>> C_CAN controller.
> >>>
> >>>> So, please let me reiterate what I understood and what was present
> >>>> in RFC version of the patch. Please add your comments/views:
> >>>>
> >>>> - drivers/net/can/c_can/c_can.c (similar on lines of
> >>> sja1000.c)
> >>>> i.e. a)no *probe* / *remove* functions here,
> >>>> b)register read/write implemented here.
> >>>>
> >>>> - drivers/net/can/c_can/c_can_platform.c (similar on lines
> of
> >>> sja1000_platform.c)
> >>>> i.e. *probe* / *remove* implemented here,
> >>>
> >>> Yes, that's what I'm thinking about.
> >>>
> >>>> Marc and Tomoya can also add their suggestions so that I can
> finalize
> >>> V3 a.s.a.p.
> >>>
> >>> That would be nice, indeed. Also have a look to Tomoya's PCH
> driver,
> >>> which also looks very good in the meantime.
> >>
> >> I am having a look at Tomoya's PCH driver, but as I mentioned in
> >> RFC V1 patch, I would rather like to have a bus sensitive `c_can`
> driver
> >
> > What do you mean by a "bus sensitive" driver?
>
> I was thinking about a "bus independent interface" like for the
> SJA1000.
> A bus sensitive driver would the be in c_can_platform.c.
>
> >> on top of which we can have the platform driver `c_can_platform`
> which
> >> essentially caters to the details of registers mapping/arch
> differences.
> >> Any other functionality like USB/PCI should be present in a separate
> file
> >> like `usb_c_can.c` or `pci_c_can.c`
> >
> > Sounds like the sja1000 approach, which is a good choice.
>
> I fully agree.
>
> >> If you agree I will try to circulate V3 a.s.ap.
> >
> > go ahead.
> >
>
> Yes, please.
Ok, so I would try to circulate V3 by tomorrow with a sja1000 *like* approach that ensures
in-order packet reception as well.
Regards,
Bhupesh
^ permalink raw reply
* Re: Using ethernet device as efficient small packet generator
From: Eric Dumazet @ 2010-12-22 8:08 UTC (permalink / raw)
To: juice; +Cc: Stephen Hemminger, netdev
In-Reply-To: <e66c1ff20e095bcc3a9a678a9935dc7e.squirrel@www.liukuma.net>
Le mercredi 22 décembre 2010 à 09:30 +0200, juice a écrit :
> > On Tue, 21 Dec 2010 11:56:42 +0200 shemminger wrote:
> > I regularly get full 1G line rate of 64 byte packets using old Opteron
> box and pktgen. It does require some tuning of IRQ's and interrupt
> mitigation but
> > no patches. Did you remember to do the basic stuff like setting IRQ
> affinity
> > and not enabling debugging or tracing in the kernel? This is on sky2,
> but
> > also using e1000 and tg3. Others have reported 7M packets per second
> over
> > 10G cards.
> > The r8169 hardware is low end consumer hardware and doesn't work as
> well.
> > It is possible to get close to 1G line rate forwarding with a single
> core
> > with current
> > generation processors. Actual rate depends on hardware and configuration
> (size of route
> > table, firewalling, etc). Much better performance with multi-queue
> hardware to spread load
> > over multiple cores.
>
> I did my testing on two kinds of boxes we use in our lab, an older Pomi
> Supermicro with e1000 and a newer Dell T3500 with tg3 and r8169.
> Both computers have dual-core 2.4G Xeon Cpus, but with somewhat different
> model and stepping.
> Both boxes are running the same OS, Ubuntu 2.6.32-26-generic #48.
>
Hmm, might be better with 10.10 ubuntu, with 2.6.35 kernels
> Could you share some information on the required interrupt tuning? It
> would certainly be easiest if the full line rate can be achieved without
> any patching of drivers or hindering normal eth/ip interface operation.
>
Thats pretty easy.
Say your card has 8 queues, do :
echo 01 >/proc/irq/*/eth1-fp-0/../smp_affinity
echo 02 >/proc/irq/*/eth1-fp-1/../smp_affinity
echo 04 >/proc/irq/*/eth1-fp-2/../smp_affinity
echo 08 >/proc/irq/*/eth1-fp-3/../smp_affinity
echo 10 >/proc/irq/*/eth1-fp-4/../smp_affinity
echo 20 >/proc/irq/*/eth1-fp-5/../smp_affinity
echo 40 >/proc/irq/*/eth1-fp-6/../smp_affinity
echo 80 >/proc/irq/*/eth1-fp-7/../smp_affinity
Then, start your pktgen threads on each queue, so that TX completion IRQ
are run on same CPU.
I confirm getting 6Mpps (or more) out of the box is OK.
I did it one year ago on ixgbe, no patches needed.
With recent kernels, it should even be faster.
^ permalink raw reply
* Re: [PATCH net-next-2.6 v2 1/1] can: c_can: Added support for Bosch C_CAN controller
From: Wolfgang Grandegger @ 2010-12-22 8:10 UTC (permalink / raw)
To: Bhupesh SHARMA
Cc: Socketcan-core-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org,
netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Marc Kleine-Budde
In-Reply-To: <D5ECB3C7A6F99444980976A8C6D896384DEAE4855F-8vAmw3ZAcdzhJTuQ9jeba9BPR1lH4CV8@public.gmane.org>
On 12/22/2010 09:03 AM, Bhupesh SHARMA wrote:
> Hi Marc and Wolfgang,
>
>> -----Original Message-----
>> From: Wolfgang Grandegger [mailto:wg-5Yr1BZd7O62+XT7JhA+gdA@public.gmane.org]
>> Sent: Wednesday, December 22, 2010 1:21 PM
>> To: Marc Kleine-Budde
>> Cc: Bhupesh SHARMA; Socketcan-core-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org;
>> netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
>> Subject: Re: [PATCH net-next-2.6 v2 1/1] can: c_can: Added support for
>> Bosch C_CAN controller
>>
>> On 12/22/2010 07:52 AM, Marc Kleine-Budde wrote:
>>> On 12/22/2010 04:36 AM, Bhupesh SHARMA wrote:
>>>> Hi Wolfgang,
>>>>
>>>>> Hi Bhupesh,
>>>>>
>>>>> On 12/21/2010 05:48 AM, Bhupesh SHARMA wrote:
>>>>>> Hi Wolfgang,
>>>>> ...
>>>>>>> In the meantime I compared the CAN chapter of the PCH manual with
>>>>> the
>>>>>>> C_CAN manual. The paragraphs I checked are *identical*. This
>> makes
>>>>>>> clear, that the "pch_can" is a clone of the C_CAN CAN
>> controller,
>>>>> with
>>>>>>> a few extensions, though. Therefore it would make sense, to
>>>>> implement a
>>>>>>> bus sensitive interface like for the SJA1000 allowing to handle
>> both
>>>>>>> CAN
>>>>>>> controllers with one driver sooner than later. Therefore, could
>> you
>>>>>>> please implement:
>>>>>>>
>>>>>>> drivers/net/can/c_can/c_can.c
>>>>>>> /c_can_platform.c
>>>>>>>
>>>>>>> Then an interface to the PCI based PCH CAN controller could be
>> added
>>>>>>> easily, e.g. as "pch_pci.c". You already had something similar in
>>>>> your
>>>>>>> RFC version of the patch, IIRC.
>>>>>>
>>>>>> This was the approach I initially proposed in my RFC V1 patch :)
>>>>>> But unfortunately we could not agree to it.
>>>>>
>>>>> I know. But at that time I was not aware of any other bus used for
>> the
>>>>> C_CAN controller.
>>>>>
>>>>>> So, please let me reiterate what I understood and what was present
>>>>>> in RFC version of the patch. Please add your comments/views:
>>>>>>
>>>>>> - drivers/net/can/c_can/c_can.c (similar on lines of
>>>>> sja1000.c)
>>>>>> i.e. a)no *probe* / *remove* functions here,
>>>>>> b)register read/write implemented here.
>>>>>>
>>>>>> - drivers/net/can/c_can/c_can_platform.c (similar on lines
>> of
>>>>> sja1000_platform.c)
>>>>>> i.e. *probe* / *remove* implemented here,
>>>>>
>>>>> Yes, that's what I'm thinking about.
>>>>>
>>>>>> Marc and Tomoya can also add their suggestions so that I can
>> finalize
>>>>> V3 a.s.a.p.
>>>>>
>>>>> That would be nice, indeed. Also have a look to Tomoya's PCH
>> driver,
>>>>> which also looks very good in the meantime.
>>>>
>>>> I am having a look at Tomoya's PCH driver, but as I mentioned in
>>>> RFC V1 patch, I would rather like to have a bus sensitive `c_can`
>> driver
>>>
>>> What do you mean by a "bus sensitive" driver?
>>
>> I was thinking about a "bus independent interface" like for the
>> SJA1000.
>> A bus sensitive driver would the be in c_can_platform.c.
>>
>>>> on top of which we can have the platform driver `c_can_platform`
>> which
>>>> essentially caters to the details of registers mapping/arch
>> differences.
>>>> Any other functionality like USB/PCI should be present in a separate
>> file
>>>> like `usb_c_can.c` or `pci_c_can.c`
>>>
>>> Sounds like the sja1000 approach, which is a good choice.
>>
>> I fully agree.
>>
>>>> If you agree I will try to circulate V3 a.s.ap.
>>>
>>> go ahead.
>>>
>>
>> Yes, please.
>
> Ok, so I would try to circulate V3 by tomorrow with a sja1000 *like* approach that ensures
> in-order packet reception as well.
That would be nice. But well, no hurry. It's Xmas time :-),
Wolfgang.
^ permalink raw reply
* Re: [PATCH V7 4/8] posix clocks: hook dynamic clocks into system calls
From: Richard Cochran @ 2010-12-22 8:21 UTC (permalink / raw)
To: Thomas Gleixner
Cc: linux-kernel, linux-api, netdev, Alan Cox, Arnd Bergmann,
Christoph Lameter, David Miller, John Stultz, Krzysztof Halasa,
Peter Zijlstra, Rodolfo Giometti
In-Reply-To: <alpine.LFD.2.00.1012171010160.12146@localhost6.localdomain6>
On Fri, Dec 17, 2010 at 11:03:46AM +0100, Thomas Gleixner wrote:
> So you could do the following:
...
> static int get_posix_clock(const clockid_t id, struct posix_clock_descr *cd)
> {
> struct file *fp = fget(CLOCKID_TO_FD(id));
> int ret;
>
> if (!fp || fp->f_op->open != posix_clock_open || !fp->private_data)
> return -ENODEV;
> ret = get_fd_clk(cd, fp);
> if (ret)
> fput(fp);
> return ret;
> }
In order to avoid leaking a fp reference, shouldn't this go like:
static int get_posix_clock(const clockid_t id, struct posix_clock_desc *cd)
{
struct file *fp = fget(CLOCKID_TO_FD(id));
int ret = -EINVAL;
if (!fp)
return ret;
if (fp->f_op->open != posix_clock_open || !fp->private_data)
goto out;
ret = get_fd_clk(cd, fp);
out:
if (ret)
fput(fp);
return ret;
}
Thanks again for your help,
Richard
^ permalink raw reply
* Re: Help: major pppoe regression since 2.6.35 (panic on first ppp conection)?
From: Andrew Morton @ 2010-12-22 8:22 UTC (permalink / raw)
To: Joel Soete; +Cc: Linux Kernel, netdev
In-Reply-To: <4D0C9BFA.4080003@scarlet.be>
(cc netdev)
The bug is still present in 2.6.37-rc6.
On Sat, 18 Dec 2010 11:33:14 +0000 Joel Soete <soete.joel@scarlet.be> wrote:
> Hello all,
>
> Since a year now, I installed my favorite Debian testing on my new X86_64 platform on which I like to run my personal kernel
> config. And all was working fine since 2.6.35 was released.
>
> The first time I experiment this issue, it was during a test of Ubuntu upgrade to the latest release 10.10 against which I
> open this bug report <https://bugs.launchpad.net/ubuntu/+source/linux/+bug/618636>: in short the system panicing on first
> connection to my ISP.
>
> This still occurs to me on the latest ubuntu stock linux image 2.6.35 as well as on my personal config from vanilla sources.
>
> As in general this kind of problem is quickly identified and solved, I do some more test with 2.6.36 and even 2.6.37-rc?
> releases, in smp or up config, with or without netfilter, with or without foreign nvidia graphical driver but this issue is
> still there for me?
>
> Obviously, I also do some researches with google but without a lot of success. So I try to figure out when this pb appears
> in the development history and find out that it was already in 2.6.35-rc1. So I do more kind of rewind in time and all was
> working fine since 2.6.34-git5 but problems appears in 2.6.34-git[67]. For my pppoe issue, there are few changes that I
> tried to revert without success?
>
> The latest related stuff I can discovered is this regression report 2.6.35 -> 2.6.36
> <http://www.spinics.net/lists/netdev/msg145824.html> and the debugging patch proposed by Jarek Poplawski effectively drops
> the ppp interface but the system doesn't panic any more, so can I get following kernel log:
> Dec 17 17:26:53 sidh2 kernel: [ 25.941424] eth1: Link up
> Dec 17 17:26:53 sidh2 kernel: [ 25.941736] eth1: Link changed: 10Mbps, half duplex
> Dec 17 17:26:53 sidh2 kernel: [ 26.266657] r8169 0000:08:00.0: eth0: link up
> Dec 17 17:26:53 sidh2 kernel: [ 26.266662] r8169 0000:08:00.0: eth0: link up
> Dec 17 17:26:53 sidh2 kernel: [ 26.829432] ip_tables: (C) 2000-2006 Netfilter Core Team
> Dec 17 17:26:58 sidh2 kernel: [ 33.017352] ppdev: user-space parallel port driver
> Dec 17 17:27:04 sidh2 kernel: [ 36.357885] eth0: no IPv6 routers present
> Dec 17 17:27:04 sidh2 kernel: [ 36.517833] eth1: no IPv6 routers present
> Dec 17 17:28:12 sidh2 kernel: [ 106.511762] BUG: unable to handle kernel NULL pointer dereference at (null)
> Dec 17 17:28:12 sidh2 kernel: [ 106.511952] IP: [<ffffffff810e405b>] put_page+0x1b/0x180
> Dec 17 17:28:12 sidh2 kernel: [ 106.512078] PGD 13c6f0067 PUD 134283067 PMD 0
> Dec 17 17:28:12 sidh2 kernel: [ 106.512293] Oops: 0000 [#1] SMP
> Dec 17 17:28:12 sidh2 kernel: [ 106.512456] last sysfs file: /sys/devices/virtual/sound/timer/uevent
> Dec 17 17:28:12 sidh2 kernel: [ 106.512533] CPU 4
> Dec 17 17:28:12 sidh2 kernel: [ 106.512584] Modules linked in: ppdev xt_TCPMSS xt_tcpmss xt_tcpudp iptable_mangle ip_tables
> x_tables lp nvidia(P) parport_pc parport agpgart i7core_edac edac_core tpm_tis tpm tpm_bios [last unloaded: scsi_wait_scan]
> Dec 17 17:28:12 sidh2 kernel: [ 106.513674]
> Dec 17 17:28:12 sidh2 kernel: [ 106.513742] Pid: 1821, comm: pppd Tainted: P 2.6.35-amd64-t3 #5 EX58-UD3R/EX58-UD3R
> Dec 17 17:28:12 sidh2 kernel: [ 106.513838] RIP: 0010:[<ffffffff810e405b>] [<ffffffff810e405b>] put_page+0x1b/0x180
> Dec 17 17:28:12 sidh2 kernel: [ 106.513982] RSP: 0018:ffff88013cecfe08 EFLAGS: 00010292
> Dec 17 17:28:12 sidh2 kernel: [ 106.514056] RAX: 0000000000000030 RBX: 0000000000000000 RCX: 0000000000001733
> Dec 17 17:28:12 sidh2 kernel: [ 106.514134] RDX: ffff88013c906640 RSI: ffff88013c906060 RDI: 0000000000000000
> Dec 17 17:28:12 sidh2 kernel: [ 106.514211] RBP: ffff88013c86e900 R08: ffffffff81342b90 R09: 00000000ffe22e08
> Dec 17 17:28:12 sidh2 kernel: [ 106.514288] R10: ffff88013cece000 R11: 0000000000000000 R12: 00000000080d5f42
> Dec 17 17:28:12 sidh2 kernel: [ 106.514366] R13: ffff88013c86f400 R14: 000000000000000a R15: 00000000000005de
> Dec 17 17:28:12 sidh2 kernel: [ 106.514444] FS: 0000000000000000(0000) GS:ffff880001f00000(0063) knlGS:00000000f75f5b20
> Dec 17 17:28:12 sidh2 kernel: [ 106.514539] CS: 0010 DS: 002b ES: 002b CR0: 000000008005003b
> Dec 17 17:28:12 sidh2 kernel: [ 106.514614] CR2: 0000000000000000 CR3: 000000013c758000 CR4: 00000000000006e0
> Dec 17 17:28:12 sidh2 kernel: [ 106.514692] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> Dec 17 17:28:12 sidh2 kernel: [ 106.514769] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
> Dec 17 17:28:12 sidh2 kernel: [ 106.514847] Process pppd (pid: 1821, threadinfo ffff88013cece000, task ffff88013c7ea120)
> Dec 17 17:28:12 sidh2 kernel: [ 106.514941] Stack:
> Dec 17 17:28:12 sidh2 kernel: [ 106.515008] 0000000000000000 0000000000000001 ffff88013c86e900 00000000080d5f42
> Dec 17 17:28:12 sidh2 kernel: [ 106.515227] <0> ffff88013c86f400 ffffffff814288b7 ffff88013c523c00 ffff88013c86e900
> Dec 17 17:28:12 sidh2 kernel: [ 106.515556] <0> ffff88013c523c00 ffffffff814283c1 ffff88013c86e900 ffffffff81342d1c
> Dec 17 17:28:12 sidh2 kernel: [ 106.515948] Call Trace:
> Dec 17 17:28:12 sidh2 kernel: [ 106.516022] [<ffffffff814288b7>] ? skb_release_data+0x77/0xd0
> Dec 17 17:28:12 sidh2 kernel: [ 106.516099] [<ffffffff814283c1>] ? __kfree_skb+0x11/0x90
> Dec 17 17:28:12 sidh2 kernel: [ 106.516176] [<ffffffff81342d1c>] ? ppp_read+0x18c/0x210
> Dec 17 17:28:12 sidh2 kernel: [ 106.516252] [<ffffffff8104df70>] ? default_wake_function+0x0/0x20
> Dec 17 17:28:12 sidh2 kernel: [ 106.516329] [<ffffffff8111f905>] ? vfs_read+0xb5/0x1a0
> Dec 17 17:28:12 sidh2 kernel: [ 106.516404] [<ffffffff8111fa3e>] ? sys_read+0x4e/0x90
> Dec 17 17:28:12 sidh2 kernel: [ 106.516479] [<ffffffff810996ff>] ? compat_sys_gettimeofday+0x9f/0xb0
> Dec 17 17:28:12 sidh2 kernel: [ 106.516558] [<ffffffff8103eb8f>] ? sysenter_dispatch+0x7/0x2e
> Dec 17 17:28:12 sidh2 kernel: [ 106.516632] Code: cc bb f8 ff 66 66 66 2e 0f 1f 84 00 00 00 00 00 48 83 ec 28 48 89 5c 24
> 08 48 89 6c 24 10 48 89 fb 4c 89 64 24 18 4c 89 6c 24 20 <48> 8b 07 66 a9 00 c0 0f 85 13 01 00 00 f0 ff 4f 08 0f 94 c0 84
> Dec 17 17:28:12 sidh2 kernel: [ 106.519823] RIP [<ffffffff810e405b>] put_page+0x1b/0x180
> Dec 17 17:28:12 sidh2 kernel: [ 106.519945] RSP <ffff88013cecfe08>
> Dec 17 17:28:12 sidh2 kernel: [ 106.520015] CR2: 0000000000000000
> Dec 17 17:28:12 sidh2 kernel: [ 106.520093] ---[ end trace bac76655d5f488ec ]---
>
>
> Sorry but I don't have enough knowledge to understand this pb, so all help will be appreciated.
>
> Tia,
> J.
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
^ permalink raw reply
* Re: [PATCH] ipv4: remove all rt cache entries on UNREGISTER event
From: Nicolas Dichtel @ 2010-12-22 8:32 UTC (permalink / raw)
To: Eric Dumazet; +Cc: netdev, Octavian Purdila
In-Reply-To: <1285751929.2615.30.camel@edumazet-laptop>
What is the status of this patch? The initial problem is still here in 2.6.37-rc5+
Regards,
Nicolas
Le 29.09.2010 11:18, Eric Dumazet a écrit :
> I found following patch was enough to avoid route being created if
> device is down. This is still racy and needs more thinking.
>
>
> diff --git a/net/ipv4/route.c b/net/ipv4/route.c
> index ac6559c..1ee0b1a 100644
> --- a/net/ipv4/route.c
> +++ b/net/ipv4/route.c
> @@ -2586,9 +2586,10 @@ static int ip_route_output_slow(struct net *net, struct rtable **rp,
> goto out;
>
> /* RACE: Check return value of inet_select_addr instead. */
> - if (__in_dev_get_rtnl(dev_out) == NULL) {
> + if (!(dev_out->flags & IFF_UP) || __in_dev_get_rtnl(dev_out) == NULL) {
> dev_put(dev_out);
> - goto out; /* Wrong error code */
> + err = -ENETUNREACH;
> + goto out;
> }
>
> if (ipv4_is_local_multicast(oldflp->fl4_dst) ||
>
>
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox