From: <Ajay.Kathat@microchip.com>
To: <linux-wireless@vger.kernel.org>
Cc: <devel@driverdev.osuosl.org>, <gregkh@linuxfoundation.org>,
<Adham.Abozaeid@microchip.com>, <johannes@sipsolutions.net>,
<Ajay.Kathat@microchip.com>
Subject: [PATCH 2/2] staging: wilc1000: avoid mutex unlock without lock in wilc_wlan_handle_txq()
Date: Thu, 23 Jan 2020 12:50:49 +0000 [thread overview]
Message-ID: <20200123182129.4053-2-ajay.kathat@microchip.com> (raw)
In-Reply-To: <20200123182129.4053-1-ajay.kathat@microchip.com>
From: Ajay Singh <ajay.kathat@microchip.com>
In wilc_wlan_handle_txq(), mutex unlock was called without acquiring
it. Also error code for full VMM condition was incorrect as discussed in
[1]. Now used a proper code to indicate VMM is full, for which transfer
to VMM is required again. 'wilc_wlan_handle_txq()' should be called
again if the VMM space was full earlier or otherwise based on
'txq_event' signal.
1. https://lore.kernel.org/driverdev-devel/20191113183322.a54mh2w6dulklgsd@kili.mountain/
Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
---
drivers/staging/wilc1000/netdev.c | 2 +-
drivers/staging/wilc1000/wlan.c | 15 ++++++++++-----
drivers/staging/wilc1000/wlan.h | 1 +
3 files changed, 12 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/wilc1000/netdev.c b/drivers/staging/wilc1000/netdev.c
index 0f48e74aa3ea..fce5bf2d82fa 100644
--- a/drivers/staging/wilc1000/netdev.c
+++ b/drivers/staging/wilc1000/netdev.c
@@ -174,7 +174,7 @@ static int wilc_txq_task(void *vp)
}
srcu_read_unlock(&wl->srcu, srcu_idx);
}
- } while (ret == -ENOBUFS && !wl->close);
+ } while (ret == WILC_VMM_ENTRY_FULL_RETRY && !wl->close);
}
return 0;
}
diff --git a/drivers/staging/wilc1000/wlan.c b/drivers/staging/wilc1000/wlan.c
index b904eda42806..601e4d1345d2 100644
--- a/drivers/staging/wilc1000/wlan.c
+++ b/drivers/staging/wilc1000/wlan.c
@@ -489,12 +489,12 @@ int wilc_wlan_handle_txq(struct wilc *wilc, u32 *txq_count)
struct wilc_vif *vif;
if (wilc->quit)
- goto out;
+ goto out_update_cnt;
mutex_lock(&wilc->txq_add_to_head_cs);
tqe = wilc_wlan_txq_get_first(wilc);
if (!tqe)
- goto out;
+ goto out_unlock;
dev = tqe->vif->ndev;
wilc_wlan_txq_filter_dup_tcp_ack(dev);
i = 0;
@@ -526,7 +526,7 @@ int wilc_wlan_handle_txq(struct wilc *wilc, u32 *txq_count)
}
if (i == 0)
- goto out;
+ goto out_unlock;
vmm_table[i] = 0x0;
acquire_bus(wilc, WILC_BUS_ACQUIRE_AND_WAKEUP);
@@ -595,7 +595,11 @@ int wilc_wlan_handle_txq(struct wilc *wilc, u32 *txq_count)
goto out_release_bus;
if (entries == 0) {
- ret = -ENOBUFS;
+ /*
+ * No VMM space available in firmware so retry to transmit
+ * the packet from tx queue.
+ */
+ ret = WILC_VMM_ENTRY_FULL_RETRY;
goto out_release_bus;
}
@@ -662,9 +666,10 @@ int wilc_wlan_handle_txq(struct wilc *wilc, u32 *txq_count)
out_release_bus:
release_bus(wilc, WILC_BUS_RELEASE_ALLOW_SLEEP);
-out:
+out_unlock:
mutex_unlock(&wilc->txq_add_to_head_cs);
+out_update_cnt:
*txq_count = wilc->txq_entries;
return ret;
}
diff --git a/drivers/staging/wilc1000/wlan.h b/drivers/staging/wilc1000/wlan.h
index 44ae6ed6882c..8c4634262adb 100644
--- a/drivers/staging/wilc1000/wlan.h
+++ b/drivers/staging/wilc1000/wlan.h
@@ -198,6 +198,7 @@
#define IS_MGMT_STATUS_SUCCES 0x040
#define WILC_WID_TYPE GENMASK(15, 12)
+#define WILC_VMM_ENTRY_FULL_RETRY 1
/********************************************
*
* Tx/Rx Queue Structure
--
2.24.0
next prev parent reply other threads:[~2020-01-23 12:50 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-01-23 12:50 [PATCH 1/2] staging: wilc1000: return zero on success and non-zero on function failure Ajay.Kathat
2020-01-23 12:50 ` Ajay.Kathat [this message]
2020-01-27 7:37 ` Dan Carpenter
2020-01-27 9:28 ` Ajay.Kathat
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=20200123182129.4053-2-ajay.kathat@microchip.com \
--to=ajay.kathat@microchip.com \
--cc=Adham.Abozaeid@microchip.com \
--cc=devel@driverdev.osuosl.org \
--cc=gregkh@linuxfoundation.org \
--cc=johannes@sipsolutions.net \
--cc=linux-wireless@vger.kernel.org \
/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).