s390/qeth: updates 2020-03-06
Subsystems (5)
cover s390/qeth: updates 2020-03-06
Hi Dave,
please apply the following patch series for qeth to netdev's net-next tree.
Just a small update to take care of a regression wrt to IRQ handling in net-next, reported by Qian Cai. The fix needs some qdio layer changes, so you will find Vasily's Acked-by in that patch.
Thanks,
Julian
Julian Wiedmann (2):
s390/qdio: add tighter controls for IRQ polling s390/qeth: remove VNICC callback parameter struct
drivers/s390/cio/qdio_setup.c | 11 +++++++++-- drivers/s390/net/qeth_core_main.c | 5 ++--- drivers/s390/net/qeth_l2_main.c | 29 ++++------------------------- 3 files changed, 15 insertions(+), 30 deletions(-)
Signature · 1 line
2.17.1
PATCH 1/2 s390/qdio: add tighter controls for IRQ polling
Once the call to qdio_establish() has completed, qdio is free to deliver data IRQs to the device driver's IRQ poll handler.
For qeth (the only qdio driver that currently uses IRQ polling) this is problematic, since the IRQs can arrive before its NAPI instance is even registered. Calling napi_schedule() from qeth_qdio_start_poll() then crashes in various nasty ways.
Until recently qeth checked for IFF_UP to drop such early interrupts, but that's fragile as well since it doesn't enforce any ordering.
Fix this properly by bringing up the qdio device in IRQS_DISABLED mode, and have the driver explicitly opt-in to receive data IRQs. qeth does so from qeth_open(), which kick-starts a NAPI poll and then calls qdio_start_irq() from qeth_poll().
Also add a matching qdio_stop_irq() in qeth_stop() to switch the qdio dataplane back into a disabled state.
--- drivers/s390/cio/qdio_setup.c | 11 +++++++++-- drivers/s390/net/qeth_core_main.c | 5 ++--- 2 files changed, 11 insertions(+), 5 deletions(-)
drivers/s390/cio/qdio_setup.c +9 -2
@@ -224,8 +224,15 @@ static void setup_queues(struct qdio_irq *irq_ptr,setup_queues_misc (q, irq_ptr, qdio_init->input_handler, i); q->is_input_q =1 ; - q->u.in.queue_start_poll = qdio_init->queue_start_poll_array ? - qdio_init->queue_start_poll_array[i] :NULL ; +if (qdio_init->queue_start_poll_array && + qdio_init->queue_start_poll_array[i]) { + q->u.in.queue_start_poll = + qdio_init->queue_start_poll_array[i]; +set_bit (QDIO_QUEUE_IRQS_DISABLED , + &q->u.in.queue_irq_state); + }else { + q->u.in.queue_start_poll =NULL ; + }setup_storage_lists (q, irq_ptr, input_sbal_array, i); input_sbal_array +=QDIO_MAX_BUFFERS_PER_Q ;
drivers/s390/net/qeth_core_main.c +2 -3
@@ -6582,9 +6582,6 @@ int qeth_open(struct net_device *dev)QETH_CARD_TEXT (card,4 ,"qethopen" ); -if (qdio_stop_irq (CARD_DDEV (card),0 ) <0 ) -return -EIO ; - card->data.state =CH_STATE_UP ;netif_tx_start_all_queues (dev);
@@ -6634,6 +6631,8 @@ int qeth_stop(struct net_device *dev) }napi_disable (&card->napi); +qdio_stop_irq (CARD_DDEV (card),0 ); +return 0 ; }EXPORT_SYMBOL_GPL (qeth_stop);
Signature · 1 line
2.17.1
PATCH 2/2 s390/qeth: remove VNICC callback parameter struct
In reply to Julian WiedmannAfter recent cleanups this is just a complicated wrapper around an u32*.
--- drivers/s390/net/qeth_l2_main.c | 29 ++++------------------------- 1 file changed, 4 insertions(+), 25 deletions(-)
drivers/s390/net/qeth_l2_main.c +4 -25
@@ -1567,23 +1567,11 @@ static int qeth_l2_vnicc_makerc(struct qeth_card *card, u16 ipa_rc)return rc; } -/* generic VNICC request call back control */ -struct _qeth_l2_vnicc_request_cbctl { -struct { -union { -u32 *sup_cmds; -u32 *timeout; - }; - } result; -}; -/* generic VNICC request call back */ static int qeth_l2_vnicc_request_cb (struct qeth_card *card,struct qeth_reply *reply,unsigned long data) { -struct _qeth_l2_vnicc_request_cbctl *cbctl = - (struct _qeth_l2_vnicc_request_cbctl *) reply->param;struct qeth_ipa_cmd *cmd = (struct qeth_ipa_cmd *) data;struct qeth_ipacmd_vnicc *rep = &cmd->data.vnicc;u32 sub_cmd = cmd->data.vnicc.hdr.sub_command;
@@ -1596,9 +1584,9 @@ static int qeth_l2_vnicc_request_cb(struct qeth_card *card, card->options.vnicc.cur_chars = rep->vnicc_cmds.enabled;if (sub_cmd ==IPA_VNICC_QUERY_CMDS ) - *cbctl->result.sup_cmds = rep->data.query_cmds.sup_cmds; + *(u32 *)reply->param = rep->data.query_cmds.sup_cmds;else if (sub_cmd ==IPA_VNICC_GET_TIMEOUT ) - *cbctl->result.timeout = rep->data.getset_timeout.timeout; + *(u32 *)reply->param = rep->data.getset_timeout.timeout;return 0 ; }
@@ -1639,7 +1627,6 @@ static int qeth_l2_vnicc_query_chars(struct qeth_card *card)static int qeth_l2_vnicc_query_cmds (struct qeth_card *card,u32 vnic_char,u32 *sup_cmds) { -struct _qeth_l2_vnicc_request_cbctl cbctl;struct qeth_cmd_buffer *iob;QETH_CARD_TEXT (card,2 ,"vniccqcm" );
@@ -1650,10 +1637,7 @@ static int qeth_l2_vnicc_query_cmds(struct qeth_card *card, u32 vnic_char,__ipa_cmd (iob)->data.vnicc.data.query_cmds.vnic_char = vnic_char; -/* prepare callback control */ - cbctl.result.sup_cmds = sup_cmds; - -return qeth_send_ipa_cmd (card, iob, qeth_l2_vnicc_request_cb, &cbctl); +return qeth_send_ipa_cmd (card, iob, qeth_l2_vnicc_request_cb, sup_cmds); }/* VNICC enable/disable characteristic request */
@@ -1677,7 +1661,6 @@ static int qeth_l2_vnicc_getset_timeout(struct qeth_card *card, u32 vnicc, u32 cmd, u32 *timeout) { struct qeth_vnicc_getset_timeout *getset_timeout; -struct _qeth_l2_vnicc_request_cbctl cbctl;struct qeth_cmd_buffer *iob;QETH_CARD_TEXT (card,2 ,"vniccgst" );
@@ -1692,11 +1675,7 @@ static int qeth_l2_vnicc_getset_timeout(struct qeth_card *card, u32 vnicc,if (cmd ==IPA_VNICC_SET_TIMEOUT ) getset_timeout->timeout = *timeout; -/* prepare callback control */ -if (cmd ==IPA_VNICC_GET_TIMEOUT ) - cbctl.result.timeout = timeout; - -return qeth_send_ipa_cmd (card, iob, qeth_l2_vnicc_request_cb, &cbctl); +return qeth_send_ipa_cmd (card, iob, qeth_l2_vnicc_request_cb, timeout); }/* set current VNICC flag state; called from sysfs store function */
Signature · 1 line
2.17.1
Quoted 8 lines · On Fri, 6 Mar 2020 09:13:09 +0100 Julian Wiedmann wrote:
Hi Dave,
please apply the following patch series for qeth to netdev's net-next tree.
Just a small update to take care of a regression wrt to IRQ handling in net-next, reported by Qian Cai. The fix needs some qdio layer changes, so you will find Vasily's Acked-by in that patch.
David Miller
maintainerFrom: Julian Wiedmann <[email protected]>
Date: Fri, 6 Mar 2020 09:13:09 +0100
Quoted 6 lines
please apply the following patch series for qeth to netdev's net-next tree.
Just a small update to take care of a regression wrt to IRQ handling in net-next, reported by Qian Cai. The fix needs some qdio layer changes, so you will find Vasily's Acked-by in that patch.
Series applied, thank you.