netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Michael Chan <michael.chan@broadcom.com>
To: davem@davemloft.net
Cc: netdev@vger.kernel.org, edumazet@google.com, kuba@kernel.org,
	pabeni@redhat.com, andrew.gospodarek@broadcom.com,
	David Wei <dw@davidwei.uk>,
	Pavan Chebbi <pavan.chebbi@broadcom.com>,
	Ajit Khaparde <ajit.khaparde@broadcom.com>
Subject: [PATCH net-next 1/7] bnxt_en: Fix and simplify bnxt_get_avail_msix() calls
Date: Tue, 30 Apr 2024 15:44:32 -0700	[thread overview]
Message-ID: <20240430224438.91494-2-michael.chan@broadcom.com> (raw)
In-Reply-To: <20240430224438.91494-1-michael.chan@broadcom.com>

[-- Attachment #1: Type: text/plain, Size: 5036 bytes --]

David reported this issue of the driver not initializing on an older
chip not running in the new resource manager mode (!BNXT_NEW_RM()).

Sample dmesg:

bnxt_en 0000:02:00.0 (unnamed net_device) (uninitialized): Able to reserve only 0 out of 9 requested RX rings
bnxt_en 0000:02:00.0 (unnamed net_device) (uninitialized): Unable to reserve tx rings
bnxt_en 0000:02:00.0 (unnamed net_device) (uninitialized): 2nd rings reservation failed.
bnxt_en 0000:02:00.0 (unnamed net_device) (uninitialized): Not enough rings available.
bnxt_en 0000:02:00.0: probe with driver bnxt_en failed with error -12

This is a regression caused by a recent commit that adds a call to
bnxt_get_avail_msix() before MSIX is initialized:

bnxt_set_dflt_rings()
  __bnxt_reserve_rings()
    bnxt_get_avail_msix()

bnxt_get_avail_msix() returns a negative number if !BNXT_NEW_RM() and
when MSIX is not initialized.  This causes __bnxt_reserve_rings() to
fail in this call sequence and the driver aborts initialization.

Before this commit in 2022:

303432211324 ("bnxt_en: Remove runtime interrupt vector allocation")

MSIX allocation for RoCE was dynamic and bnxt_get_avail_msix() was
used to determine the available run-time MSIX available for RoCE.

Today, bnxt_get_avail_msix() is only used to reserve some available
MSIX ahead of time to be ready when the RoCE driver loads.  It
only needs to be called when BNXT_NEW_RM() is true because older
chips do not require reservations for MSIX.  Simplify
bnxt_get_avail_msix() to only consider the BNXT_NEW_RM() case and
only make this call if BNXT_NEW_RM() is true.

Also change bnxt_get_avail_msix() to static since it is only used
in bnxt.c.

Reported-by: David Wei <dw@davidwei.uk>
Fixes: d630624ebd70 ("bnxt_en: Utilize ulp client resources if RoCE is not registered")
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
Cc: David Wei <dw@davidwei.uk>
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 20 ++++++++------------
 drivers/net/ethernet/broadcom/bnxt/bnxt.h |  1 -
 2 files changed, 8 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index be96bb494ae6..0eb880766012 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -7459,6 +7459,8 @@ static bool bnxt_rings_ok(struct bnxt *bp, struct bnxt_hw_rings *hwr)
 	       hwr->stat && (hwr->cp_p5 || !(bp->flags & BNXT_FLAG_CHIP_P5_PLUS));
 }
 
+static int bnxt_get_avail_msix(struct bnxt *bp, int num);
+
 static int __bnxt_reserve_rings(struct bnxt *bp)
 {
 	struct bnxt_hw_rings hwr = {0};
@@ -7471,7 +7473,7 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
 	if (!bnxt_need_reserve_rings(bp))
 		return 0;
 
-	if (!bnxt_ulp_registered(bp->edev)) {
+	if (!bnxt_ulp_registered(bp->edev) && BNXT_NEW_RM(bp)) {
 		ulp_msix = bnxt_get_avail_msix(bp, bp->ulp_num_msix_want);
 		if (!ulp_msix)
 			bnxt_set_ulp_stat_ctxs(bp, 0);
@@ -10474,19 +10476,13 @@ unsigned int bnxt_get_avail_stat_ctxs_for_en(struct bnxt *bp)
 	return bnxt_get_max_func_stat_ctxs(bp) - bnxt_get_func_stat_ctxs(bp);
 }
 
-int bnxt_get_avail_msix(struct bnxt *bp, int num)
+/* Only called if BNXT_NEW_RM() is true to get the available MSIX to
+ * reserve ahead of time before RoCE is registered.
+ */
+static int bnxt_get_avail_msix(struct bnxt *bp, int num)
 {
-	int max_cp = bnxt_get_max_func_cp_rings(bp);
 	int max_irq = bnxt_get_max_func_irqs(bp);
 	int total_req = bp->cp_nr_rings + num;
-	int max_idx, avail_msix;
-
-	max_idx = bp->total_irqs;
-	if (!(bp->flags & BNXT_FLAG_CHIP_P5_PLUS))
-		max_idx = min_t(int, bp->total_irqs, max_cp);
-	avail_msix = max_idx - bp->cp_nr_rings;
-	if (!BNXT_NEW_RM(bp) || avail_msix >= num)
-		return avail_msix;
 
 	if (max_irq < total_req) {
 		num = max_irq - bp->cp_nr_rings;
@@ -10619,7 +10615,7 @@ int bnxt_reserve_rings(struct bnxt *bp, bool irq_re_init)
 	if (!bnxt_need_reserve_rings(bp))
 		return 0;
 
-	if (!bnxt_ulp_registered(bp->edev)) {
+	if (!bnxt_ulp_registered(bp->edev) && BNXT_NEW_RM(bp)) {
 		int ulp_msix = bnxt_get_avail_msix(bp, bp->ulp_num_msix_want);
 
 		if (ulp_msix > bp->ulp_num_msix_want)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index ad57ef051798..0c680032ab66 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -2736,7 +2736,6 @@ unsigned int bnxt_get_max_func_stat_ctxs(struct bnxt *bp);
 unsigned int bnxt_get_avail_stat_ctxs_for_en(struct bnxt *bp);
 unsigned int bnxt_get_max_func_cp_rings(struct bnxt *bp);
 unsigned int bnxt_get_avail_cp_rings_for_en(struct bnxt *bp);
-int bnxt_get_avail_msix(struct bnxt *bp, int num);
 int bnxt_reserve_rings(struct bnxt *bp, bool irq_re_init);
 void bnxt_tx_disable(struct bnxt *bp);
 void bnxt_tx_enable(struct bnxt *bp);
-- 
2.30.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]

  reply	other threads:[~2024-04-30 22:46 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-04-30 22:44 [PATCH net-next 0/7] bnxt_en: Updates for net-next Michael Chan
2024-04-30 22:44 ` Michael Chan [this message]
2024-04-30 23:19   ` [PATCH net-next 1/7] bnxt_en: Fix and simplify bnxt_get_avail_msix() calls Jakub Kicinski
2024-04-30 23:35     ` Michael Chan
2024-04-30 23:44       ` Jakub Kicinski
2024-04-30 22:44 ` [PATCH net-next 2/7] bnxt_en: share NQ ring sw_stats memory with subrings Michael Chan
2024-04-30 22:44 ` [PATCH net-next 3/7] bnxt_en: Don't support offline self test when RoCE driver is loaded Michael Chan
2024-04-30 22:44 ` [PATCH net-next 4/7] bnxt_en: Don't call ULP_STOP/ULP_START during L2 reset Michael Chan
2024-04-30 22:44 ` [PATCH net-next 5/7] bnxt_en: Add a mutex to synchronize ULP operations Michael Chan
2024-04-30 22:44 ` [PATCH net-next 6/7] bnxt_en: Optimize recovery path ULP locking in the driver Michael Chan
2024-04-30 22:44 ` [PATCH net-next 7/7] bnxt_en: Add VF PCI ID for 5760X (P7) chips Michael Chan

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20240430224438.91494-2-michael.chan@broadcom.com \
    --to=michael.chan@broadcom.com \
    --cc=ajit.khaparde@broadcom.com \
    --cc=andrew.gospodarek@broadcom.com \
    --cc=davem@davemloft.net \
    --cc=dw@davidwei.uk \
    --cc=edumazet@google.com \
    --cc=kuba@kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=pavan.chebbi@broadcom.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).