debian-packaging
/
linux
8
0
Fork 0
Code Releases Activity

Backport fix for TSO with mv643xx_eth driver 

Replacing previous workaround for #764162.

svn path=/dists/sid/linux/; revision=22061
This commit is contained in:
Ian Campbell 2014-11-09 10:54:17 +00:00
parent 534a88e05a
commit c58f0350f9
4 changed files with 78 additions and 43 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
debian/changelog vendored
View File

@ -7,6 +7,8 @@ linux (3.16.7-3) UNRELEASED; urgency=medium
[ Ian Campbell ]
* [xen] Backport various netback fixes (Closes: #767261).
* Backport fix for TSO with mv643xx_eth driver, replacing previous workaround
(#764162)
-- Ben Hutchings <ben@decadent.org.uk> Sun, 09 Nov 2014 10:13:09 +0000

42
debian/patches/bugfix/all/net-mv643xx-disable-tso-by-default.patch vendored
View File

@ -1,42 +0,0 @@
Subject: [1/1] net: mv643xx_eth: Make TSO disabled by default
From: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Date: Sat, 1 Nov 2014 12:30:20 -0300
Origin: http://patchwork.ozlabs.org/patch/405792/
Data corruption has been observed to be produced by TSO. For instance,
accessing files on a NFS-server with TSO enabled results in different data
transferred each time.
This has been observed only on Kirkwood platforms, i.e. with the mv643xx_eth
driver. Same tests on platforms using the mvneta ethernet driver have
passed without errors.
Make TSO disabled by default for now, until we can found a proper fix
for the regression.
Fixes: 3ae8f4e0b98 ('net: mv643xx_eth: Implement software TSO')
Reported-by: Slawomir Gajzner <slawomir.gajzner@gmail.com>
Reported-by: Julien D'Ascenzio <jdascenzio@yahoo.fr>
Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
---
drivers/net/ethernet/marvell/mv643xx_eth.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/marvell/mv643xx_eth.c b/drivers/net/ethernet/marvell/mv643xx_eth.c
index b151a94..8b72780 100644
--- a/drivers/net/ethernet/marvell/mv643xx_eth.c
+++ b/drivers/net/ethernet/marvell/mv643xx_eth.c
@@ -3110,11 +3110,11 @@ static int mv643xx_eth_probe(struct platform_device *pdev)
dev->watchdog_timeo = 2 * HZ;
dev->base_addr = 0;
- dev->features = NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_TSO;
+ dev->features = NETIF_F_SG | NETIF_F_IP_CSUM;
dev->vlan_features = dev->features;
dev->features |= NETIF_F_RXCSUM;
- dev->hw_features = dev->features;
+ dev->hw_features = dev->features | NETIF_F_TSO;
dev->priv_flags |= IFF_UNICAST_FLT;
dev->gso_max_segs = MV643XX_MAX_TSO_SEGS;

75
debian/patches/bugfix/all/net-mv643xx_eth-reclaim-TX-skbs-only-when-released-b.patch vendored Normal file
View File

@ -0,0 +1,75 @@
From 2c2a9cbd64387d6b70ac5db013e9bfe9412c7354 Mon Sep 17 00:00:00 2001
From: Karl Beldan <karl.beldan@rivierawaves.com>
Date: Wed, 5 Nov 2014 15:32:59 +0100
Subject: [PATCH] net: mv643xx_eth: reclaim TX skbs only when released by the
HW
Origin: https://git.kernel.org/cgit/linux/kernel/git/davem/net.git/commit/?id=2c2a9cbd64387d6b70ac5db013e9bfe9412c7354
ATM, txq_reclaim will dequeue and free an skb for each tx desc released
by the hw that has TX_LAST_DESC set. However, in case of TSO, each
hw desc embedding the last part of a segment has TX_LAST_DESC set,
losing the one-to-one 'last skb frag'/'TX_LAST_DESC set' correspondance,
which causes data corruption.
Fix this by checking TX_ENABLE_INTERRUPT instead of TX_LAST_DESC, and
warn when trying to dequeue from an empty txq (which can be symptomatic
of releasing skbs prematurely).
Fixes: 3ae8f4e0b98 ('net: mv643xx_eth: Implement software TSO')
Reported-by: Slawomir Gajzner <slawomir.gajzner@gmail.com>
Reported-by: Julien D'Ascenzio <jdascenzio@yahoo.fr>
Signed-off-by: Karl Beldan <karl.beldan@rivierawaves.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
drivers/net/ethernet/marvell/mv643xx_eth.c | 18 ++++++++++--------
1 file changed, 10 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/marvell/mv643xx_eth.c b/drivers/net/ethernet/marvell/mv643xx_eth.c
index b151a94..d44560d 100644
--- a/drivers/net/ethernet/marvell/mv643xx_eth.c
+++ b/drivers/net/ethernet/marvell/mv643xx_eth.c
@@ -1047,7 +1047,6 @@ static int txq_reclaim(struct tx_queue *txq, int budget, int force)
int tx_index;
struct tx_desc *desc;
u32 cmd_sts;
- struct sk_buff *skb;
tx_index = txq->tx_used_desc;
desc = &txq->tx_desc_area[tx_index];
@@ -1066,19 +1065,22 @@ static int txq_reclaim(struct tx_queue *txq, int budget, int force)
reclaimed++;
txq->tx_desc_count--;
- skb = NULL;
- if (cmd_sts & TX_LAST_DESC)
- skb = __skb_dequeue(&txq->tx_skb);
+ if (!IS_TSO_HEADER(txq, desc->buf_ptr))
+ dma_unmap_single(mp->dev->dev.parent, desc->buf_ptr,
+ desc->byte_cnt, DMA_TO_DEVICE);
+
+ if (cmd_sts & TX_ENABLE_INTERRUPT) {
+ struct sk_buff *skb = __skb_dequeue(&txq->tx_skb);
+
+ if (!WARN_ON(!skb))
+ dev_kfree_skb(skb);
+ }
if (cmd_sts & ERROR_SUMMARY) {
netdev_info(mp->dev, "tx error\n");
mp->dev->stats.tx_errors++;
}
- if (!IS_TSO_HEADER(txq, desc->buf_ptr))
- dma_unmap_single(mp->dev->dev.parent, desc->buf_ptr,
- desc->byte_cnt, DMA_TO_DEVICE);
- dev_kfree_skb(skb);
}
__netif_tx_unlock_bh(nq);
--
1.7.10.4

2
debian/patches/series vendored
View File

@ -145,7 +145,7 @@ bugfix/all/net-sctp-fix-skb_over_panic-when-receiving-malformed.patch
bugfix/all/net-sctp-fix-panic-on-duplicate-ASCONF-chunks.patch
bugfix/all/net-sctp-fix-remote-memory-pressure-from-excessive-q.patch
bugfix/all/mnt-Prevent-pivot_root-from-creating-a-loop-in-the-m.patch
bugfix/all/net-mv643xx-disable-tso-by-default.patch
bugfix/all/net-mv643xx_eth-reclaim-TX-skbs-only-when-released-b.patch
bugfix/all/xen-netback-Adding-debugfs-io_ring_qX-files.patch
bugfix/all/xen-netback-Using-a-new-state-bit-instead-of-carrier.patch
bugfix/all/xen-netback-Turn-off-the-carrier-if-the-guest-is-not.patch