From: Finn Thain <fthain@telegraphics.com.au>
To: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
Chris Zankel <chris@zankel.net>,
Laurent Vivier <laurent@vivier.eu>,
Geert Uytterhoeven <geert@linux-m68k.org>,
Eric Dumazet <eric.dumazet@gmail.com>,
Stephen Hemminger <stephen@networkplumber.org>,
netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [PATCH net v3 05/12] net/sonic: Fix receive buffer handling
Date: Thu, 23 Jan 2020 09:07:26 +1100 [thread overview]
Message-ID: <02f71a07f0194f3020f4ed8f8456e2c2c4c6a685.1579730846.git.fthain@telegraphics.com.au> (raw)
In-Reply-To: <cover.1579730846.git.fthain@telegraphics.com.au>
The SONIC can sometimes advance its rx buffer pointer (RRP register)
without advancing its rx descriptor pointer (CRDA register). As a result
the index of the current rx descriptor may not equal that of the current
rx buffer. The driver mistakenly assumes that they are always equal.
This assumption leads to incorrect packet lengths and possible packet
duplication. Avoid this by calling a new function to locate the buffer
corresponding to a given descriptor.
Fixes: efcce839360f ("[PATCH] macsonic/jazzsonic network drivers update")
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
Changed since v2:
- Removed 'inline' keyword.
- Removed blank line.
---
drivers/net/ethernet/natsemi/sonic.c | 35 ++++++++++++++++++++++++----
drivers/net/ethernet/natsemi/sonic.h | 5 ++--
2 files changed, 33 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/natsemi/sonic.c b/drivers/net/ethernet/natsemi/sonic.c
index b31c7af32dfa..c666bbf15116 100644
--- a/drivers/net/ethernet/natsemi/sonic.c
+++ b/drivers/net/ethernet/natsemi/sonic.c
@@ -413,6 +413,21 @@ static irqreturn_t sonic_interrupt(int irq, void *dev_id)
return IRQ_HANDLED;
}
+/* Return the array index corresponding to a given Receive Buffer pointer. */
+static int index_from_addr(struct sonic_local *lp, dma_addr_t addr,
+ unsigned int last)
+{
+ unsigned int i = last;
+
+ do {
+ i = (i + 1) & SONIC_RRS_MASK;
+ if (addr == lp->rx_laddr[i])
+ return i;
+ } while (i != last);
+
+ return -ENOENT;
+}
+
/*
* We have a good packet(s), pass it/them up the network stack.
*/
@@ -432,6 +447,16 @@ static void sonic_rx(struct net_device *dev)
status = sonic_rda_get(dev, entry, SONIC_RD_STATUS);
if (status & SONIC_RCR_PRX) {
+ u32 addr = (sonic_rda_get(dev, entry,
+ SONIC_RD_PKTPTR_H) << 16) |
+ sonic_rda_get(dev, entry, SONIC_RD_PKTPTR_L);
+ int i = index_from_addr(lp, addr, entry);
+
+ if (i < 0) {
+ WARN_ONCE(1, "failed to find buffer!\n");
+ break;
+ }
+
/* Malloc up new buffer. */
new_skb = netdev_alloc_skb(dev, SONIC_RBSIZE + 2);
if (new_skb == NULL) {
@@ -453,7 +478,7 @@ static void sonic_rx(struct net_device *dev)
/* now we have a new skb to replace it, pass the used one up the stack */
dma_unmap_single(lp->device, lp->rx_laddr[entry], SONIC_RBSIZE, DMA_FROM_DEVICE);
- used_skb = lp->rx_skb[entry];
+ used_skb = lp->rx_skb[i];
pkt_len = sonic_rda_get(dev, entry, SONIC_RD_PKTLEN);
skb_trim(used_skb, pkt_len);
used_skb->protocol = eth_type_trans(used_skb, dev);
@@ -462,13 +487,13 @@ static void sonic_rx(struct net_device *dev)
lp->stats.rx_bytes += pkt_len;
/* and insert the new skb */
- lp->rx_laddr[entry] = new_laddr;
- lp->rx_skb[entry] = new_skb;
+ lp->rx_laddr[i] = new_laddr;
+ lp->rx_skb[i] = new_skb;
bufadr_l = (unsigned long)new_laddr & 0xffff;
bufadr_h = (unsigned long)new_laddr >> 16;
- sonic_rra_put(dev, entry, SONIC_RR_BUFADR_L, bufadr_l);
- sonic_rra_put(dev, entry, SONIC_RR_BUFADR_H, bufadr_h);
+ sonic_rra_put(dev, i, SONIC_RR_BUFADR_L, bufadr_l);
+ sonic_rra_put(dev, i, SONIC_RR_BUFADR_H, bufadr_h);
} else {
/* This should only happen, if we enable accepting broken packets. */
}
diff --git a/drivers/net/ethernet/natsemi/sonic.h b/drivers/net/ethernet/natsemi/sonic.h
index 9e4ff8dd032d..e6d47e45c5c2 100644
--- a/drivers/net/ethernet/natsemi/sonic.h
+++ b/drivers/net/ethernet/natsemi/sonic.h
@@ -275,8 +275,9 @@
#define SONIC_NUM_RDS SONIC_NUM_RRS /* number of receive descriptors */
#define SONIC_NUM_TDS 16 /* number of transmit descriptors */
-#define SONIC_RDS_MASK (SONIC_NUM_RDS-1)
-#define SONIC_TDS_MASK (SONIC_NUM_TDS-1)
+#define SONIC_RRS_MASK (SONIC_NUM_RRS - 1)
+#define SONIC_RDS_MASK (SONIC_NUM_RDS - 1)
+#define SONIC_TDS_MASK (SONIC_NUM_TDS - 1)
#define SONIC_RBSIZE 1520 /* size of one resource buffer */
--
2.24.1
next prev parent reply other threads:[~2020-01-22 22:24 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-01-22 22:07 [PATCH net v3 00/12] Fixes for SONIC ethernet driver Finn Thain
2020-01-22 22:07 ` [PATCH net v3 10/12] net/sonic: Fix command register usage Finn Thain
2020-01-22 22:07 ` [PATCH net v3 04/12] net/sonic: Fix interface error stats collection Finn Thain
2020-01-22 22:07 ` [PATCH net v3 08/12] net/sonic: Fix receive buffer replenishment Finn Thain
2020-01-22 22:07 ` [PATCH net v3 02/12] net/sonic: Clear interrupt flags immediately Finn Thain
2020-01-22 22:07 ` [PATCH net v3 06/12] net/sonic: Avoid needless receive descriptor EOL flag updates Finn Thain
2020-01-22 22:07 ` [PATCH net v3 03/12] net/sonic: Use MMIO accessors Finn Thain
2020-01-22 22:07 ` [PATCH net v3 11/12] net/sonic: Fix CAM initialization Finn Thain
2020-01-22 22:07 ` [PATCH net v3 12/12] net/sonic: Prevent tx watchdog timeout Finn Thain
2020-01-22 22:07 ` [PATCH net v3 09/12] net/sonic: Quiesce SONIC before re-initializing descriptor memory Finn Thain
2020-01-22 22:07 ` [PATCH net v3 07/12] net/sonic: Improve receive descriptor status flag check Finn Thain
2020-01-22 22:07 ` [PATCH net v3 01/12] net/sonic: Add mutual exclusion for accessing shared state Finn Thain
2020-01-22 22:07 ` Finn Thain [this message]
2020-01-23 20:24 ` [PATCH net v3 00/12] Fixes for SONIC ethernet driver David Miller
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=02f71a07f0194f3020f4ed8f8456e2c2c4c6a685.1579730846.git.fthain@telegraphics.com.au \
--to=fthain@telegraphics.com.au \
--cc=chris@zankel.net \
--cc=davem@davemloft.net \
--cc=eric.dumazet@gmail.com \
--cc=geert@linux-m68k.org \
--cc=laurent@vivier.eu \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=stephen@networkplumber.org \
--cc=tsbogend@alpha.franken.de \
/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).