2022-10-09 23:50:22

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 01/73] libbpf: Do not require executable permission for shared libraries

From: Hengqi Chen <[email protected]>

[ Upstream commit 9e32084ef1c33a87a736d6ce3fcb95b60dac9aa1 ]

Currently, resolve_full_path() requires executable permission for both
programs and shared libraries. This causes failures on distos like Debian
since the shared libraries are not installed executable and Linux is not
requiring shared libraries to have executable permissions. Let's remove
executable permission check for shared libraries.

Reported-by: Goro Fuji <[email protected]>
Signed-off-by: Hengqi Chen <[email protected]>
Signed-off-by: Andrii Nakryiko <[email protected]>
Link: https://lore.kernel.org/bpf/[email protected]
Signed-off-by: Sasha Levin <[email protected]>
---
tools/lib/bpf/libbpf.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 266357b1dca1..91bfe42e5cf4 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -11206,15 +11206,17 @@ static const char *arch_specific_lib_paths(void)
static int resolve_full_path(const char *file, char *result, size_t result_sz)
{
const char *search_paths[3] = {};
- int i;
+ int i, perm;

if (str_has_sfx(file, ".so") || strstr(file, ".so.")) {
search_paths[0] = getenv("LD_LIBRARY_PATH");
search_paths[1] = "/usr/lib64:/usr/lib";
search_paths[2] = arch_specific_lib_paths();
+ perm = R_OK;
} else {
search_paths[0] = getenv("PATH");
search_paths[1] = "/usr/bin:/usr/sbin";
+ perm = R_OK | X_OK;
}

for (i = 0; i < ARRAY_SIZE(search_paths); i++) {
@@ -11233,8 +11235,8 @@ static int resolve_full_path(const char *file, char *result, size_t result_sz)
if (!seg_len)
continue;
snprintf(result, result_sz, "%.*s/%s", seg_len, s, file);
- /* ensure it is an executable file/link */
- if (access(result, R_OK | X_OK) < 0)
+ /* ensure it has required permissions */
+ if (access(result, perm) < 0)
continue;
pr_debug("resolved '%s' to '%s'\n", file, result);
return 0;
--
2.35.1


2022-10-09 23:50:23

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 34/73] net: mscc: ocelot: report FIFO drop counters through stats->rx_dropped

From: Vladimir Oltean <[email protected]>

[ Upstream commit cc160fc29a264726b2bfbc2f551081430db3df03 ]

if_link.h says:

* @rx_dropped: Number of packets received but not processed,
* e.g. due to lack of resources or unsupported protocol.
* For hardware interfaces this counter may include packets discarded
* due to L2 address filtering but should not include packets dropped
* by the device due to buffer exhaustion which are counted separately in
* @rx_missed_errors (since procfs folds those two counters together).

Currently we report "stats->rx_dropped = dev->stats.rx_dropped", the
latter being incremented by various entities in the stack. This is not
wrong, but we'd like to move ocelot_get_stats64() in the common ocelot
switch lib which is independent of struct net_device.

To do that, report the hardware RX drop counters instead. These drops
are due to policer action, or due to no destinations. When we have no
memory in the queue system, report this through rx_missed_errors, as
instructed.

Signed-off-by: Vladimir Oltean <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/mscc/ocelot_net.c | 27 +++++++++++++++++++++++++-
1 file changed, 26 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mscc/ocelot_net.c b/drivers/net/ethernet/mscc/ocelot_net.c
index 330d30841cdc..d7956fd051e6 100644
--- a/drivers/net/ethernet/mscc/ocelot_net.c
+++ b/drivers/net/ethernet/mscc/ocelot_net.c
@@ -745,7 +745,32 @@ static void ocelot_get_stats64(struct net_device *dev,
s[OCELOT_STAT_RX_1024_1526] +
s[OCELOT_STAT_RX_1527_MAX];
stats->multicast = s[OCELOT_STAT_RX_MULTICAST];
- stats->rx_dropped = dev->stats.rx_dropped;
+ stats->rx_missed_errors = s[OCELOT_STAT_DROP_TAIL];
+ stats->rx_dropped = s[OCELOT_STAT_RX_RED_PRIO_0] +
+ s[OCELOT_STAT_RX_RED_PRIO_1] +
+ s[OCELOT_STAT_RX_RED_PRIO_2] +
+ s[OCELOT_STAT_RX_RED_PRIO_3] +
+ s[OCELOT_STAT_RX_RED_PRIO_4] +
+ s[OCELOT_STAT_RX_RED_PRIO_5] +
+ s[OCELOT_STAT_RX_RED_PRIO_6] +
+ s[OCELOT_STAT_RX_RED_PRIO_7] +
+ s[OCELOT_STAT_DROP_LOCAL] +
+ s[OCELOT_STAT_DROP_YELLOW_PRIO_0] +
+ s[OCELOT_STAT_DROP_YELLOW_PRIO_1] +
+ s[OCELOT_STAT_DROP_YELLOW_PRIO_2] +
+ s[OCELOT_STAT_DROP_YELLOW_PRIO_3] +
+ s[OCELOT_STAT_DROP_YELLOW_PRIO_4] +
+ s[OCELOT_STAT_DROP_YELLOW_PRIO_5] +
+ s[OCELOT_STAT_DROP_YELLOW_PRIO_6] +
+ s[OCELOT_STAT_DROP_YELLOW_PRIO_7] +
+ s[OCELOT_STAT_DROP_GREEN_PRIO_0] +
+ s[OCELOT_STAT_DROP_GREEN_PRIO_1] +
+ s[OCELOT_STAT_DROP_GREEN_PRIO_2] +
+ s[OCELOT_STAT_DROP_GREEN_PRIO_3] +
+ s[OCELOT_STAT_DROP_GREEN_PRIO_4] +
+ s[OCELOT_STAT_DROP_GREEN_PRIO_5] +
+ s[OCELOT_STAT_DROP_GREEN_PRIO_6] +
+ s[OCELOT_STAT_DROP_GREEN_PRIO_7];

/* Get Tx stats */
stats->tx_bytes = s[OCELOT_STAT_TX_OCTETS];
--
2.35.1

2022-10-09 23:50:23

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 70/73] i2c: designware-pci: Group AMD NAVI quirk parts together

From: Andy Shevchenko <[email protected]>

[ Upstream commit 65769162ae4b7f2d82e54998be446226b05fcd8f ]

The code is ogranized in a way that all related parts
to the certain platform quirk go together. This is not
the case for AMD NAVI. Shuffle code to make it happen.

While at it, drop the frequency definition and use
hard coded value as it's done for other platforms and
add a comment to the PCI ID list.

Signed-off-by: Andy Shevchenko <[email protected]>
Acked-by: Jarkko Nikula <[email protected]>
Signed-off-by: Wolfram Sang <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/i2c/busses/i2c-designware-pcidrv.c | 30 +++++++++++-----------
1 file changed, 15 insertions(+), 15 deletions(-)

diff --git a/drivers/i2c/busses/i2c-designware-pcidrv.c b/drivers/i2c/busses/i2c-designware-pcidrv.c
index 608e61209455..ca368482b246 100644
--- a/drivers/i2c/busses/i2c-designware-pcidrv.c
+++ b/drivers/i2c/busses/i2c-designware-pcidrv.c
@@ -27,7 +27,6 @@
#include "i2c-ccgx-ucsi.h"

#define DRIVER_NAME "i2c-designware-pci"
-#define AMD_CLK_RATE_HZ 100000

enum dw_pci_ctl_id_t {
medfield,
@@ -100,11 +99,6 @@ static u32 mfld_get_clk_rate_khz(struct dw_i2c_dev *dev)
return 25000;
}

-static u32 navi_amd_get_clk_rate_khz(struct dw_i2c_dev *dev)
-{
- return AMD_CLK_RATE_HZ;
-}
-
static int mfld_setup(struct pci_dev *pdev, struct dw_pci_controller *c)
{
struct dw_i2c_dev *dev = dev_get_drvdata(&pdev->dev);
@@ -126,15 +120,6 @@ static int mfld_setup(struct pci_dev *pdev, struct dw_pci_controller *c)
return -ENODEV;
}

-static int navi_amd_setup(struct pci_dev *pdev, struct dw_pci_controller *c)
-{
- struct dw_i2c_dev *dev = dev_get_drvdata(&pdev->dev);
-
- dev->flags |= MODEL_AMD_NAVI_GPU;
- dev->timings.bus_freq_hz = I2C_MAX_STANDARD_MODE_FREQ;
- return 0;
-}
-
static int mrfld_setup(struct pci_dev *pdev, struct dw_pci_controller *c)
{
/*
@@ -159,6 +144,20 @@ static u32 ehl_get_clk_rate_khz(struct dw_i2c_dev *dev)
return 100000;
}

+static u32 navi_amd_get_clk_rate_khz(struct dw_i2c_dev *dev)
+{
+ return 100000;
+}
+
+static int navi_amd_setup(struct pci_dev *pdev, struct dw_pci_controller *c)
+{
+ struct dw_i2c_dev *dev = dev_get_drvdata(&pdev->dev);
+
+ dev->flags |= MODEL_AMD_NAVI_GPU;
+ dev->timings.bus_freq_hz = I2C_MAX_STANDARD_MODE_FREQ;
+ return 0;
+}
+
static struct dw_pci_controller dw_pci_controllers[] = {
[medfield] = {
.bus_num = -1,
@@ -389,6 +388,7 @@ static const struct pci_device_id i2_designware_pci_ids[] = {
{ PCI_VDEVICE(INTEL, 0x4bbe), elkhartlake },
{ PCI_VDEVICE(INTEL, 0x4bbf), elkhartlake },
{ PCI_VDEVICE(INTEL, 0x4bc0), elkhartlake },
+ /* AMD NAVI */
{ PCI_VDEVICE(ATI, 0x7314), navi_amd },
{ PCI_VDEVICE(ATI, 0x73a4), navi_amd },
{ PCI_VDEVICE(ATI, 0x73e4), navi_amd },
--
2.35.1

2022-10-09 23:50:31

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 48/73] net: wwan: t7xx: Fix return type of t7xx_ccmni_start_xmit

From: Nathan Huckleberry <[email protected]>

[ Upstream commit 73c99e26036529e633a0f2d628ad7ddff6594668 ]

The ndo_start_xmit field in net_device_ops is expected to be of type
netdev_tx_t (*ndo_start_xmit)(struct sk_buff *skb, struct net_device *dev).

The mismatched return type breaks forward edge kCFI since the underlying
function definition does not match the function hook definition.

The return type of t7xx_ccmni_start_xmit should be changed from int to
netdev_tx_t.

Reported-by: Dan Carpenter <[email protected]>
Link: https://github.com/ClangBuiltLinux/linux/issues/1703
Cc: [email protected]
Signed-off-by: Nathan Huckleberry <[email protected]>
Acked-by: Sergey Ryazanov <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/wwan/t7xx/t7xx_netdev.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/wwan/t7xx/t7xx_netdev.c b/drivers/net/wwan/t7xx/t7xx_netdev.c
index c6b6547f2c6f..f71d3bc3b237 100644
--- a/drivers/net/wwan/t7xx/t7xx_netdev.c
+++ b/drivers/net/wwan/t7xx/t7xx_netdev.c
@@ -74,7 +74,7 @@ static int t7xx_ccmni_send_packet(struct t7xx_ccmni *ccmni, struct sk_buff *skb,
return 0;
}

-static int t7xx_ccmni_start_xmit(struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t t7xx_ccmni_start_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct t7xx_ccmni *ccmni = wwan_netdev_drvpriv(dev);
int skb_len = skb->len;
--
2.35.1

2022-10-09 23:50:32

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 71/73] net: sparx5: Fix return type of sparx5_port_xmit_impl

From: Nathan Huckleberry <[email protected]>

[ Upstream commit 73ea735073599430818e89b8901452287a15a718 ]

The ndo_start_xmit field in net_device_ops is expected to be of type
netdev_tx_t (*ndo_start_xmit)(struct sk_buff *skb, struct net_device *dev).

The mismatched return type breaks forward edge kCFI since the underlying
function definition does not match the function hook definition.

The return type of sparx5_port_xmit_impl should be changed from int to
netdev_tx_t.

Reported-by: Dan Carpenter <[email protected]>
Link: https://github.com/ClangBuiltLinux/linux/issues/1703
Cc: [email protected]
Signed-off-by: Nathan Huckleberry <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/microchip/sparx5/sparx5_main.h | 2 +-
drivers/net/ethernet/microchip/sparx5/sparx5_packet.c | 4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/microchip/sparx5/sparx5_main.h b/drivers/net/ethernet/microchip/sparx5/sparx5_main.h
index d071ac3b7106..705d8852078f 100644
--- a/drivers/net/ethernet/microchip/sparx5/sparx5_main.h
+++ b/drivers/net/ethernet/microchip/sparx5/sparx5_main.h
@@ -291,7 +291,7 @@ struct frame_info {
void sparx5_xtr_flush(struct sparx5 *sparx5, u8 grp);
void sparx5_ifh_parse(u32 *ifh, struct frame_info *info);
irqreturn_t sparx5_xtr_handler(int irq, void *_priv);
-int sparx5_port_xmit_impl(struct sk_buff *skb, struct net_device *dev);
+netdev_tx_t sparx5_port_xmit_impl(struct sk_buff *skb, struct net_device *dev);
int sparx5_manual_injection_mode(struct sparx5 *sparx5);
void sparx5_port_inj_timer_setup(struct sparx5_port *port);

diff --git a/drivers/net/ethernet/microchip/sparx5/sparx5_packet.c b/drivers/net/ethernet/microchip/sparx5/sparx5_packet.c
index 21844beba72d..83c16ca5b30f 100644
--- a/drivers/net/ethernet/microchip/sparx5/sparx5_packet.c
+++ b/drivers/net/ethernet/microchip/sparx5/sparx5_packet.c
@@ -222,13 +222,13 @@ static int sparx5_inject(struct sparx5 *sparx5,
return NETDEV_TX_OK;
}

-int sparx5_port_xmit_impl(struct sk_buff *skb, struct net_device *dev)
+netdev_tx_t sparx5_port_xmit_impl(struct sk_buff *skb, struct net_device *dev)
{
struct net_device_stats *stats = &dev->stats;
struct sparx5_port *port = netdev_priv(dev);
struct sparx5 *sparx5 = port->sparx5;
u32 ifh[IFH_LEN];
- int ret;
+ netdev_tx_t ret;

memset(ifh, 0, IFH_LEN * 4);
sparx5_set_port_ifh(ifh, port->portno);
--
2.35.1

2022-10-09 23:50:50

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 45/73] net: ethernet: litex: Fix return type of liteeth_start_xmit

From: Nathan Huckleberry <[email protected]>

[ Upstream commit 40662333dd7c64664247a6138bc33f3974e3a331 ]

The ndo_start_xmit field in net_device_ops is expected to be of type
netdev_tx_t (*ndo_start_xmit)(struct sk_buff *skb, struct net_device *dev).

The mismatched return type breaks forward edge kCFI since the underlying
function definition does not match the function hook definition.

The return type of liteeth_start_xmit should be changed from int to
netdev_tx_t.

Reported-by: Dan Carpenter <[email protected]>
Link: https://github.com/ClangBuiltLinux/linux/issues/1703
Cc: [email protected]
Signed-off-by: Nathan Huckleberry <[email protected]>
Reviewed-by: Nathan Chancellor <[email protected]>
Acked-by: Gabriel Somlo <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/litex/litex_liteeth.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/litex/litex_liteeth.c b/drivers/net/ethernet/litex/litex_liteeth.c
index fdd99f0de424..35f24e0f0934 100644
--- a/drivers/net/ethernet/litex/litex_liteeth.c
+++ b/drivers/net/ethernet/litex/litex_liteeth.c
@@ -152,7 +152,8 @@ static int liteeth_stop(struct net_device *netdev)
return 0;
}

-static int liteeth_start_xmit(struct sk_buff *skb, struct net_device *netdev)
+static netdev_tx_t liteeth_start_xmit(struct sk_buff *skb,
+ struct net_device *netdev)
{
struct liteeth *priv = netdev_priv(netdev);
void __iomem *txbuffer;
--
2.35.1

2022-10-09 23:50:56

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 28/73] net: lantiq_etop: Fix return type for implementation of ndo_start_xmit

From: GUO Zihua <[email protected]>

[ Upstream commit c8ef3c94bda0e21123202d057d4a299698fa0ed9 ]

Since Linux now supports CFI, it will be a good idea to fix mismatched
return type for implementation of hooks. Otherwise this might get
cought out by CFI and cause a panic.

ltq_etop_tx() would return either NETDEV_TX_BUSY or NETDEV_TX_OK, so
change the return type to netdev_tx_t directly.

Signed-off-by: GUO Zihua <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/lantiq_etop.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/lantiq_etop.c b/drivers/net/ethernet/lantiq_etop.c
index 7cedbe1fdfd7..59aab4086dcc 100644
--- a/drivers/net/ethernet/lantiq_etop.c
+++ b/drivers/net/ethernet/lantiq_etop.c
@@ -470,7 +470,7 @@ ltq_etop_stop(struct net_device *dev)
return 0;
}

-static int
+static netdev_tx_t
ltq_etop_tx(struct sk_buff *skb, struct net_device *dev)
{
int queue = skb_get_queue_mapping(skb);
--
2.35.1

2022-10-09 23:51:02

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 42/73] net: ax88796c: Fix return type of ax88796c_start_xmit

From: Nathan Huckleberry <[email protected]>

[ Upstream commit fcb7c210a24209ea8f6f32593580b57f52382ec2 ]

The ndo_start_xmit field in net_device_ops is expected to be of type
netdev_tx_t (*ndo_start_xmit)(struct sk_buff *skb, struct net_device *dev).

The mismatched return type breaks forward edge kCFI since the underlying
function definition does not match the function hook definition.

The return type of ax88796c_start_xmit should be changed from int to
netdev_tx_t.

Reported-by: Dan Carpenter <[email protected]>
Link: https://github.com/ClangBuiltLinux/linux/issues/1703
Cc: [email protected]
Signed-off-by: Nathan Huckleberry <[email protected]>
Acked-by: Lukasz Stelmach <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/asix/ax88796c_main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/asix/ax88796c_main.c b/drivers/net/ethernet/asix/ax88796c_main.c
index 6ba5b024a7be..f1d610efd69e 100644
--- a/drivers/net/ethernet/asix/ax88796c_main.c
+++ b/drivers/net/ethernet/asix/ax88796c_main.c
@@ -381,7 +381,7 @@ static int ax88796c_hard_xmit(struct ax88796c_device *ax_local)
return 1;
}

-static int
+static netdev_tx_t
ax88796c_start_xmit(struct sk_buff *skb, struct net_device *ndev)
{
struct ax88796c_device *ax_local = to_ax88796c_device(ndev);
--
2.35.1

2022-10-09 23:51:03

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 31/73] iavf: Fix race between iavf_close and iavf_reset_task

From: Michal Jaron <[email protected]>

[ Upstream commit 11c12adcbc1598d91e73ab6ddfa41d25a01478ed ]

During stress tests with adding VF to namespace and changing vf's
trust there was a race between iavf_reset_task and iavf_close.
Sometimes when IAVF_FLAG_AQ_DISABLE_QUEUES from iavf_close was sent
to PF after reset and before IAVF_AQ_GET_CONFIG was sent then PF
returns error IAVF_NOT_SUPPORTED to disable queues request and
following requests. There is need to get_config before other
aq_required will be send but iavf_close clears all flags, if
get_config was not sent before iavf_close, then it will not be send
at all.

In case when IAVF_FLAG_AQ_GET_OFFLOAD_VLAN_V2_CAPS was sent before
IAVF_FLAG_AQ_DISABLE_QUEUES then there was rtnl_lock deadlock
between iavf_close and iavf_adminq_task until iavf_close timeouts
and disable queues was sent after iavf_close ends.

There was also a problem with sending delete/add filters.
Sometimes when filters was not yet added to PF and in
iavf_close all filters was set to remove there might be a try
to remove nonexistent filters on PF.

Add aq_required_tmp to save aq_required flags and send them after
disable_queues will be handled. Clear flags given to iavf_down
different than IAVF_FLAG_AQ_GET_CONFIG as this flag is necessary
to sent other aq_required. Remove some flags that we don't
want to send as we are in iavf_close and we want to disable
interface. Remove filters which was not yet sent and send del
filters flags only when there are filters to remove.

Signed-off-by: Michal Jaron <[email protected]>
Signed-off-by: Mateusz Palczewski <[email protected]>
Tested-by: Konrad Jankowski <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/intel/iavf/iavf_main.c | 177 ++++++++++++++++----
1 file changed, 141 insertions(+), 36 deletions(-)

diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c
index 981c43b204ff..d3822c264642 100644
--- a/drivers/net/ethernet/intel/iavf/iavf_main.c
+++ b/drivers/net/ethernet/intel/iavf/iavf_main.c
@@ -1184,66 +1184,138 @@ static void iavf_up_complete(struct iavf_adapter *adapter)
}

/**
- * iavf_down - Shutdown the connection processing
+ * iavf_clear_mac_vlan_filters - Remove mac and vlan filters not sent to PF
+ * yet and mark other to be removed.
* @adapter: board private structure
- *
- * Expects to be called while holding the __IAVF_IN_CRITICAL_TASK bit lock.
**/
-void iavf_down(struct iavf_adapter *adapter)
+static void iavf_clear_mac_vlan_filters(struct iavf_adapter *adapter)
{
- struct net_device *netdev = adapter->netdev;
- struct iavf_vlan_filter *vlf;
- struct iavf_cloud_filter *cf;
- struct iavf_fdir_fltr *fdir;
- struct iavf_mac_filter *f;
- struct iavf_adv_rss *rss;
-
- if (adapter->state <= __IAVF_DOWN_PENDING)
- return;
-
- netif_carrier_off(netdev);
- netif_tx_disable(netdev);
- adapter->link_up = false;
- iavf_napi_disable_all(adapter);
- iavf_irq_disable(adapter);
+ struct iavf_vlan_filter *vlf, *vlftmp;
+ struct iavf_mac_filter *f, *ftmp;

spin_lock_bh(&adapter->mac_vlan_list_lock);
-
/* clear the sync flag on all filters */
__dev_uc_unsync(adapter->netdev, NULL);
__dev_mc_unsync(adapter->netdev, NULL);

/* remove all MAC filters */
- list_for_each_entry(f, &adapter->mac_filter_list, list) {
- f->remove = true;
+ list_for_each_entry_safe(f, ftmp, &adapter->mac_filter_list,
+ list) {
+ if (f->add) {
+ list_del(&f->list);
+ kfree(f);
+ } else {
+ f->remove = true;
+ }
}

/* remove all VLAN filters */
- list_for_each_entry(vlf, &adapter->vlan_filter_list, list) {
- vlf->remove = true;
+ list_for_each_entry_safe(vlf, vlftmp, &adapter->vlan_filter_list,
+ list) {
+ if (vlf->add) {
+ list_del(&vlf->list);
+ kfree(vlf);
+ } else {
+ vlf->remove = true;
+ }
}
-
spin_unlock_bh(&adapter->mac_vlan_list_lock);
+}
+
+/**
+ * iavf_clear_cloud_filters - Remove cloud filters not sent to PF yet and
+ * mark other to be removed.
+ * @adapter: board private structure
+ **/
+static void iavf_clear_cloud_filters(struct iavf_adapter *adapter)
+{
+ struct iavf_cloud_filter *cf, *cftmp;

/* remove all cloud filters */
spin_lock_bh(&adapter->cloud_filter_list_lock);
- list_for_each_entry(cf, &adapter->cloud_filter_list, list) {
- cf->del = true;
+ list_for_each_entry_safe(cf, cftmp, &adapter->cloud_filter_list,
+ list) {
+ if (cf->add) {
+ list_del(&cf->list);
+ kfree(cf);
+ adapter->num_cloud_filters--;
+ } else {
+ cf->del = true;
+ }
}
spin_unlock_bh(&adapter->cloud_filter_list_lock);
+}
+
+/**
+ * iavf_clear_fdir_filters - Remove fdir filters not sent to PF yet and mark
+ * other to be removed.
+ * @adapter: board private structure
+ **/
+static void iavf_clear_fdir_filters(struct iavf_adapter *adapter)
+{
+ struct iavf_fdir_fltr *fdir, *fdirtmp;

/* remove all Flow Director filters */
spin_lock_bh(&adapter->fdir_fltr_lock);
- list_for_each_entry(fdir, &adapter->fdir_list_head, list) {
- fdir->state = IAVF_FDIR_FLTR_DEL_REQUEST;
+ list_for_each_entry_safe(fdir, fdirtmp, &adapter->fdir_list_head,
+ list) {
+ if (fdir->state == IAVF_FDIR_FLTR_ADD_REQUEST) {
+ list_del(&fdir->list);
+ kfree(fdir);
+ adapter->fdir_active_fltr--;
+ } else {
+ fdir->state = IAVF_FDIR_FLTR_DEL_REQUEST;
+ }
}
spin_unlock_bh(&adapter->fdir_fltr_lock);
+}
+
+/**
+ * iavf_clear_adv_rss_conf - Remove adv rss conf not sent to PF yet and mark
+ * other to be removed.
+ * @adapter: board private structure
+ **/
+static void iavf_clear_adv_rss_conf(struct iavf_adapter *adapter)
+{
+ struct iavf_adv_rss *rss, *rsstmp;

/* remove all advance RSS configuration */
spin_lock_bh(&adapter->adv_rss_lock);
- list_for_each_entry(rss, &adapter->adv_rss_list_head, list)
- rss->state = IAVF_ADV_RSS_DEL_REQUEST;
+ list_for_each_entry_safe(rss, rsstmp, &adapter->adv_rss_list_head,
+ list) {
+ if (rss->state == IAVF_ADV_RSS_ADD_REQUEST) {
+ list_del(&rss->list);
+ kfree(rss);
+ } else {
+ rss->state = IAVF_ADV_RSS_DEL_REQUEST;
+ }
+ }
spin_unlock_bh(&adapter->adv_rss_lock);
+}
+
+/**
+ * iavf_down - Shutdown the connection processing
+ * @adapter: board private structure
+ *
+ * Expects to be called while holding the __IAVF_IN_CRITICAL_TASK bit lock.
+ **/
+void iavf_down(struct iavf_adapter *adapter)
+{
+ struct net_device *netdev = adapter->netdev;
+
+ if (adapter->state <= __IAVF_DOWN_PENDING)
+ return;
+
+ netif_carrier_off(netdev);
+ netif_tx_disable(netdev);
+ adapter->link_up = false;
+ iavf_napi_disable_all(adapter);
+ iavf_irq_disable(adapter);
+
+ iavf_clear_mac_vlan_filters(adapter);
+ iavf_clear_cloud_filters(adapter);
+ iavf_clear_fdir_filters(adapter);
+ iavf_clear_adv_rss_conf(adapter);

if (!(adapter->flags & IAVF_FLAG_PF_COMMS_FAILED)) {
/* cancel any current operation */
@@ -1252,11 +1324,16 @@ void iavf_down(struct iavf_adapter *adapter)
* here for this to complete. The watchdog is still running
* and it will take care of this.
*/
- adapter->aq_required = IAVF_FLAG_AQ_DEL_MAC_FILTER;
- adapter->aq_required |= IAVF_FLAG_AQ_DEL_VLAN_FILTER;
- adapter->aq_required |= IAVF_FLAG_AQ_DEL_CLOUD_FILTER;
- adapter->aq_required |= IAVF_FLAG_AQ_DEL_FDIR_FILTER;
- adapter->aq_required |= IAVF_FLAG_AQ_DEL_ADV_RSS_CFG;
+ if (!list_empty(&adapter->mac_filter_list))
+ adapter->aq_required |= IAVF_FLAG_AQ_DEL_MAC_FILTER;
+ if (!list_empty(&adapter->vlan_filter_list))
+ adapter->aq_required |= IAVF_FLAG_AQ_DEL_VLAN_FILTER;
+ if (!list_empty(&adapter->cloud_filter_list))
+ adapter->aq_required |= IAVF_FLAG_AQ_DEL_CLOUD_FILTER;
+ if (!list_empty(&adapter->fdir_list_head))
+ adapter->aq_required |= IAVF_FLAG_AQ_DEL_FDIR_FILTER;
+ if (!list_empty(&adapter->adv_rss_list_head))
+ adapter->aq_required |= IAVF_FLAG_AQ_DEL_ADV_RSS_CFG;
adapter->aq_required |= IAVF_FLAG_AQ_DISABLE_QUEUES;
}

@@ -4082,6 +4159,7 @@ static int iavf_open(struct net_device *netdev)
static int iavf_close(struct net_device *netdev)
{
struct iavf_adapter *adapter = netdev_priv(netdev);
+ u64 aq_to_restore;
int status;

mutex_lock(&adapter->crit_lock);
@@ -4094,6 +4172,29 @@ static int iavf_close(struct net_device *netdev)
set_bit(__IAVF_VSI_DOWN, adapter->vsi.state);
if (CLIENT_ENABLED(adapter))
adapter->flags |= IAVF_FLAG_CLIENT_NEEDS_CLOSE;
+ /* We cannot send IAVF_FLAG_AQ_GET_OFFLOAD_VLAN_V2_CAPS before
+ * IAVF_FLAG_AQ_DISABLE_QUEUES because in such case there is rtnl
+ * deadlock with adminq_task() until iavf_close timeouts. We must send
+ * IAVF_FLAG_AQ_GET_CONFIG before IAVF_FLAG_AQ_DISABLE_QUEUES to make
+ * disable queues possible for vf. Give only necessary flags to
+ * iavf_down and save other to set them right before iavf_close()
+ * returns, when IAVF_FLAG_AQ_DISABLE_QUEUES will be already sent and
+ * iavf will be in DOWN state.
+ */
+ aq_to_restore = adapter->aq_required;
+ adapter->aq_required &= IAVF_FLAG_AQ_GET_CONFIG;
+
+ /* Remove flags which we do not want to send after close or we want to
+ * send before disable queues.
+ */
+ aq_to_restore &= ~(IAVF_FLAG_AQ_GET_CONFIG |
+ IAVF_FLAG_AQ_ENABLE_QUEUES |
+ IAVF_FLAG_AQ_CONFIGURE_QUEUES |
+ IAVF_FLAG_AQ_ADD_VLAN_FILTER |
+ IAVF_FLAG_AQ_ADD_MAC_FILTER |
+ IAVF_FLAG_AQ_ADD_CLOUD_FILTER |
+ IAVF_FLAG_AQ_ADD_FDIR_FILTER |
+ IAVF_FLAG_AQ_ADD_ADV_RSS_CFG);

iavf_down(adapter);
iavf_change_state(adapter, __IAVF_DOWN_PENDING);
@@ -4117,6 +4218,10 @@ static int iavf_close(struct net_device *netdev)
msecs_to_jiffies(500));
if (!status)
netdev_warn(netdev, "Device resources not yet released\n");
+
+ mutex_lock(&adapter->crit_lock);
+ adapter->aq_required |= aq_to_restore;
+ mutex_unlock(&adapter->crit_lock);
return 0;
}

--
2.35.1

2022-10-09 23:51:11

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 12/73] tcp: annotate data-race around tcp_md5sig_pool_populated

From: Eric Dumazet <[email protected]>

[ Upstream commit aacd467c0a576e5e44d2de4205855dc0fe43f6fb ]

tcp_md5sig_pool_populated can be read while another thread
changes its value.

The race has no consequence because allocations
are protected with tcp_md5sig_mutex.

This patch adds READ_ONCE() and WRITE_ONCE() to document
the race and silence KCSAN.

Reported-by: Abhishek Shah <[email protected]>
Signed-off-by: Eric Dumazet <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
net/ipv4/tcp.c | 14 ++++++++++----
1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index ab03977b6578..6426d5e01cb6 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -4347,12 +4347,16 @@ static void __tcp_alloc_md5sig_pool(void)
* to memory. See smp_rmb() in tcp_get_md5sig_pool()
*/
smp_wmb();
- tcp_md5sig_pool_populated = true;
+ /* Paired with READ_ONCE() from tcp_alloc_md5sig_pool()
+ * and tcp_get_md5sig_pool().
+ */
+ WRITE_ONCE(tcp_md5sig_pool_populated, true);
}

bool tcp_alloc_md5sig_pool(void)
{
- if (unlikely(!tcp_md5sig_pool_populated)) {
+ /* Paired with WRITE_ONCE() from __tcp_alloc_md5sig_pool() */
+ if (unlikely(!READ_ONCE(tcp_md5sig_pool_populated))) {
mutex_lock(&tcp_md5sig_mutex);

if (!tcp_md5sig_pool_populated) {
@@ -4363,7 +4367,8 @@ bool tcp_alloc_md5sig_pool(void)

mutex_unlock(&tcp_md5sig_mutex);
}
- return tcp_md5sig_pool_populated;
+ /* Paired with WRITE_ONCE() from __tcp_alloc_md5sig_pool() */
+ return READ_ONCE(tcp_md5sig_pool_populated);
}
EXPORT_SYMBOL(tcp_alloc_md5sig_pool);

@@ -4379,7 +4384,8 @@ struct tcp_md5sig_pool *tcp_get_md5sig_pool(void)
{
local_bh_disable();

- if (tcp_md5sig_pool_populated) {
+ /* Paired with WRITE_ONCE() from __tcp_alloc_md5sig_pool() */
+ if (READ_ONCE(tcp_md5sig_pool_populated)) {
/* coupled with smp_wmb() in __tcp_alloc_md5sig_pool() */
smp_rmb();
return this_cpu_ptr(&tcp_md5sig_pool);
--
2.35.1

2022-10-09 23:51:37

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 20/73] net: axienet: Switch to 64-bit RX/TX statistics

From: Robert Hancock <[email protected]>

[ Upstream commit cb45a8bf4693965e89d115cd2c510f12bc127c37 ]

The RX and TX byte/packet statistics in this driver could be overflowed
relatively quickly on a 32-bit platform. Switch these stats to use the
u64_stats infrastructure to avoid this.

Signed-off-by: Robert Hancock <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/xilinx/xilinx_axienet.h | 12 ++++++
.../net/ethernet/xilinx/xilinx_axienet_main.c | 37 +++++++++++++++++--
2 files changed, 45 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/xilinx/xilinx_axienet.h b/drivers/net/ethernet/xilinx/xilinx_axienet.h
index f2e2261b4b7d..8ff4333de2ad 100644
--- a/drivers/net/ethernet/xilinx/xilinx_axienet.h
+++ b/drivers/net/ethernet/xilinx/xilinx_axienet.h
@@ -402,6 +402,9 @@ struct axidma_bd {
* @rx_bd_num: Size of RX buffer descriptor ring
* @rx_bd_ci: Stores the index of the Rx buffer descriptor in the ring being
* accessed currently.
+ * @rx_packets: RX packet count for statistics
+ * @rx_bytes: RX byte count for statistics
+ * @rx_stat_sync: Synchronization object for RX stats
* @napi_tx: NAPI TX control structure
* @tx_dma_cr: Nominal content of TX DMA control register
* @tx_bd_v: Virtual address of the TX buffer descriptor ring
@@ -411,6 +414,9 @@ struct axidma_bd {
* complete. Only updated at runtime by TX NAPI poll.
* @tx_bd_tail: Stores the index of the next Tx buffer descriptor in the ring
* to be populated.
+ * @tx_packets: TX packet count for statistics
+ * @tx_bytes: TX byte count for statistics
+ * @tx_stat_sync: Synchronization object for TX stats
* @dma_err_task: Work structure to process Axi DMA errors
* @tx_irq: Axidma TX IRQ number
* @rx_irq: Axidma RX IRQ number
@@ -458,6 +464,9 @@ struct axienet_local {
dma_addr_t rx_bd_p;
u32 rx_bd_num;
u32 rx_bd_ci;
+ u64_stats_t rx_packets;
+ u64_stats_t rx_bytes;
+ struct u64_stats_sync rx_stat_sync;

struct napi_struct napi_tx;
u32 tx_dma_cr;
@@ -466,6 +475,9 @@ struct axienet_local {
u32 tx_bd_num;
u32 tx_bd_ci;
u32 tx_bd_tail;
+ u64_stats_t tx_packets;
+ u64_stats_t tx_bytes;
+ struct u64_stats_sync tx_stat_sync;

struct work_struct dma_err_task;

diff --git a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
index 1760930ec0c4..9262988d26a3 100644
--- a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
+++ b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
@@ -752,8 +752,10 @@ static int axienet_tx_poll(struct napi_struct *napi, int budget)
if (lp->tx_bd_ci >= lp->tx_bd_num)
lp->tx_bd_ci %= lp->tx_bd_num;

- ndev->stats.tx_packets += packets;
- ndev->stats.tx_bytes += size;
+ u64_stats_update_begin(&lp->tx_stat_sync);
+ u64_stats_add(&lp->tx_packets, packets);
+ u64_stats_add(&lp->tx_bytes, size);
+ u64_stats_update_end(&lp->tx_stat_sync);

/* Matches barrier in axienet_start_xmit */
smp_mb();
@@ -984,8 +986,10 @@ static int axienet_rx_poll(struct napi_struct *napi, int budget)
cur_p = &lp->rx_bd_v[lp->rx_bd_ci];
}

- lp->ndev->stats.rx_packets += packets;
- lp->ndev->stats.rx_bytes += size;
+ u64_stats_update_begin(&lp->rx_stat_sync);
+ u64_stats_add(&lp->rx_packets, packets);
+ u64_stats_add(&lp->rx_bytes, size);
+ u64_stats_update_end(&lp->rx_stat_sync);

if (tail_p)
axienet_dma_out_addr(lp, XAXIDMA_RX_TDESC_OFFSET, tail_p);
@@ -1292,10 +1296,32 @@ static int axienet_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
return phylink_mii_ioctl(lp->phylink, rq, cmd);
}

+static void
+axienet_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats)
+{
+ struct axienet_local *lp = netdev_priv(dev);
+ unsigned int start;
+
+ netdev_stats_to_stats64(stats, &dev->stats);
+
+ do {
+ start = u64_stats_fetch_begin_irq(&lp->rx_stat_sync);
+ stats->rx_packets = u64_stats_read(&lp->rx_packets);
+ stats->rx_bytes = u64_stats_read(&lp->rx_bytes);
+ } while (u64_stats_fetch_retry_irq(&lp->rx_stat_sync, start));
+
+ do {
+ start = u64_stats_fetch_begin_irq(&lp->tx_stat_sync);
+ stats->tx_packets = u64_stats_read(&lp->tx_packets);
+ stats->tx_bytes = u64_stats_read(&lp->tx_bytes);
+ } while (u64_stats_fetch_retry_irq(&lp->tx_stat_sync, start));
+}
+
static const struct net_device_ops axienet_netdev_ops = {
.ndo_open = axienet_open,
.ndo_stop = axienet_stop,
.ndo_start_xmit = axienet_start_xmit,
+ .ndo_get_stats64 = axienet_get_stats64,
.ndo_change_mtu = axienet_change_mtu,
.ndo_set_mac_address = netdev_set_mac_address,
.ndo_validate_addr = eth_validate_addr,
@@ -1850,6 +1876,9 @@ static int axienet_probe(struct platform_device *pdev)
lp->rx_bd_num = RX_BD_NUM_DEFAULT;
lp->tx_bd_num = TX_BD_NUM_DEFAULT;

+ u64_stats_init(&lp->rx_stat_sync);
+ u64_stats_init(&lp->tx_stat_sync);
+
netif_napi_add(ndev, &lp->napi_rx, axienet_rx_poll, NAPI_POLL_WEIGHT);
netif_napi_add(ndev, &lp->napi_tx, axienet_tx_poll, NAPI_POLL_WEIGHT);

--
2.35.1

2022-10-10 00:00:04

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 36/73] regulator: core: Prevent integer underflow

From: Patrick Rudolph <[email protected]>

[ Upstream commit 8d8e16592022c9650df8aedfe6552ed478d7135b ]

By using a ratio of delay to poll_enabled_time that is not integer
time_remaining underflows and does not exit the loop as expected.
As delay could be derived from DT and poll_enabled_time is defined
in the driver this can easily happen.

Use a signed iterator to make sure that the loop exits once
the remaining time is negative.

Signed-off-by: Patrick Rudolph <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/regulator/core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index a9daaf4d5aaa..9567d2fc3f00 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -2680,7 +2680,7 @@ static int _regulator_do_enable(struct regulator_dev *rdev)
* return -ETIMEDOUT.
*/
if (rdev->desc->poll_enabled_time) {
- unsigned int time_remaining = delay;
+ int time_remaining = delay;

while (time_remaining > 0) {
_regulator_delay_helper(rdev->desc->poll_enabled_time);
--
2.35.1

2022-10-10 00:02:27

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 29/73] netlink: Bounds-check struct nlmsgerr creation

From: Kees Cook <[email protected]>

[ Upstream commit 710d21fdff9a98d621cd4e64167f3ef8af4e2fd1 ]

In preparation for FORTIFY_SOURCE doing bounds-check on memcpy(),
switch from __nlmsg_put to nlmsg_put(), and explain the bounds check
for dealing with the memcpy() across a composite flexible array struct.
Avoids this future run-time warning:

memcpy: detected field-spanning write (size 32) of single field "&errmsg->msg" at net/netlink/af_netlink.c:2447 (size 16)

Cc: Jakub Kicinski <[email protected]>
Cc: Pablo Neira Ayuso <[email protected]>
Cc: Jozsef Kadlecsik <[email protected]>
Cc: Florian Westphal <[email protected]>
Cc: "David S. Miller" <[email protected]>
Cc: Eric Dumazet <[email protected]>
Cc: Paolo Abeni <[email protected]>
Cc: syzbot <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Kees Cook <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: David S. Miller <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
net/netfilter/ipset/ip_set_core.c | 8 +++++---
net/netlink/af_netlink.c | 8 +++++---
2 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/net/netfilter/ipset/ip_set_core.c b/net/netfilter/ipset/ip_set_core.c
index 16ae92054baa..6b31746f9be3 100644
--- a/net/netfilter/ipset/ip_set_core.c
+++ b/net/netfilter/ipset/ip_set_core.c
@@ -1719,11 +1719,13 @@ call_ad(struct net *net, struct sock *ctnl, struct sk_buff *skb,
skb2 = nlmsg_new(payload, GFP_KERNEL);
if (!skb2)
return -ENOMEM;
- rep = __nlmsg_put(skb2, NETLINK_CB(skb).portid,
- nlh->nlmsg_seq, NLMSG_ERROR, payload, 0);
+ rep = nlmsg_put(skb2, NETLINK_CB(skb).portid,
+ nlh->nlmsg_seq, NLMSG_ERROR, payload, 0);
errmsg = nlmsg_data(rep);
errmsg->error = ret;
- memcpy(&errmsg->msg, nlh, nlh->nlmsg_len);
+ unsafe_memcpy(&errmsg->msg, nlh, nlh->nlmsg_len,
+ /* Bounds checked by the skb layer. */);
+
cmdattr = (void *)&errmsg->msg + min_len;

ret = nla_parse(cda, IPSET_ATTR_CMD_MAX, cmdattr,
diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c
index 0cd91f813a3b..d8d3ed2096a3 100644
--- a/net/netlink/af_netlink.c
+++ b/net/netlink/af_netlink.c
@@ -2440,11 +2440,13 @@ void netlink_ack(struct sk_buff *in_skb, struct nlmsghdr *nlh, int err,
return;
}

- rep = __nlmsg_put(skb, NETLINK_CB(in_skb).portid, nlh->nlmsg_seq,
- NLMSG_ERROR, payload, flags);
+ rep = nlmsg_put(skb, NETLINK_CB(in_skb).portid, nlh->nlmsg_seq,
+ NLMSG_ERROR, payload, flags);
errmsg = nlmsg_data(rep);
errmsg->error = err;
- memcpy(&errmsg->msg, nlh, payload > sizeof(*errmsg) ? nlh->nlmsg_len : sizeof(*nlh));
+ unsafe_memcpy(&errmsg->msg, nlh, payload > sizeof(*errmsg)
+ ? nlh->nlmsg_len : sizeof(*nlh),
+ /* Bounds checked by the skb layer. */);

if (nlk_has_extack && extack) {
if (extack->_msg) {
--
2.35.1

2022-10-10 00:02:33

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 07/73] net: ethernet: ti: davinci_mdio: Add workaround for errata i2329

From: Ravi Gunasekaran <[email protected]>

[ Upstream commit d04807b80691c6041ca8e3dcf1870d1bf1082c22 ]

On the CPSW and ICSS peripherals, there is a possibility that the MDIO
interface returns corrupt data on MDIO reads or writes incorrect data
on MDIO writes. There is also a possibility for the MDIO interface to
become unavailable until the next peripheral reset.

The workaround is to configure the MDIO in manual mode and disable the
MDIO state machine and emulate the MDIO protocol by reading and writing
appropriate fields in MDIO_MANUAL_IF_REG register of the MDIO controller
to manipulate the MDIO clock and data pins.

More details about the errata i2329 and the workaround is available in:
https://www.ti.com/lit/er/sprz487a/sprz487a.pdf

Add implementation to disable MDIO state machine, configure MDIO in manual
mode and achieve MDIO read and writes via MDIO Bitbanging

Signed-off-by: Ravi Gunasekaran <[email protected]>
Reported-by: kernel test robot <[email protected]>
Reviewed-by: Andrew Lunn <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/ti/davinci_mdio.c | 242 +++++++++++++++++++++++--
1 file changed, 231 insertions(+), 11 deletions(-)

diff --git a/drivers/net/ethernet/ti/davinci_mdio.c b/drivers/net/ethernet/ti/davinci_mdio.c
index ea3772618043..946b9753ccfb 100644
--- a/drivers/net/ethernet/ti/davinci_mdio.c
+++ b/drivers/net/ethernet/ti/davinci_mdio.c
@@ -26,6 +26,8 @@
#include <linux/of_device.h>
#include <linux/of_mdio.h>
#include <linux/pinctrl/consumer.h>
+#include <linux/mdio-bitbang.h>
+#include <linux/sys_soc.h>

/*
* This timeout definition is a worst-case ultra defensive measure against
@@ -41,6 +43,7 @@

struct davinci_mdio_of_param {
int autosuspend_delay_ms;
+ bool manual_mode;
};

struct davinci_mdio_regs {
@@ -49,6 +52,15 @@ struct davinci_mdio_regs {
#define CONTROL_IDLE BIT(31)
#define CONTROL_ENABLE BIT(30)
#define CONTROL_MAX_DIV (0xffff)
+#define CONTROL_CLKDIV GENMASK(15, 0)
+
+#define MDIO_MAN_MDCLK_O BIT(2)
+#define MDIO_MAN_OE BIT(1)
+#define MDIO_MAN_PIN BIT(0)
+#define MDIO_MANUALMODE BIT(31)
+
+#define MDIO_PIN 0
+

u32 alive;
u32 link;
@@ -59,7 +71,9 @@ struct davinci_mdio_regs {
u32 userintmasked;
u32 userintmaskset;
u32 userintmaskclr;
- u32 __reserved_1[20];
+ u32 manualif;
+ u32 poll;
+ u32 __reserved_1[18];

struct {
u32 access;
@@ -79,6 +93,7 @@ static const struct mdio_platform_data default_pdata = {

struct davinci_mdio_data {
struct mdio_platform_data pdata;
+ struct mdiobb_ctrl bb_ctrl;
struct davinci_mdio_regs __iomem *regs;
struct clk *clk;
struct device *dev;
@@ -90,6 +105,7 @@ struct davinci_mdio_data {
*/
bool skip_scan;
u32 clk_div;
+ bool manual_mode;
};

static void davinci_mdio_init_clk(struct davinci_mdio_data *data)
@@ -128,9 +144,122 @@ static void davinci_mdio_enable(struct davinci_mdio_data *data)
writel(data->clk_div | CONTROL_ENABLE, &data->regs->control);
}

-static int davinci_mdio_reset(struct mii_bus *bus)
+static void davinci_mdio_disable(struct davinci_mdio_data *data)
+{
+ u32 reg;
+
+ /* Disable MDIO state machine */
+ reg = readl(&data->regs->control);
+
+ reg &= ~CONTROL_CLKDIV;
+ reg |= data->clk_div;
+
+ reg &= ~CONTROL_ENABLE;
+ writel(reg, &data->regs->control);
+}
+
+static void davinci_mdio_enable_manual_mode(struct davinci_mdio_data *data)
+{
+ u32 reg;
+ /* set manual mode */
+ reg = readl(&data->regs->poll);
+ reg |= MDIO_MANUALMODE;
+ writel(reg, &data->regs->poll);
+}
+
+static void davinci_set_mdc(struct mdiobb_ctrl *ctrl, int level)
+{
+ struct davinci_mdio_data *data;
+ u32 reg;
+
+ data = container_of(ctrl, struct davinci_mdio_data, bb_ctrl);
+ reg = readl(&data->regs->manualif);
+
+ if (level)
+ reg |= MDIO_MAN_MDCLK_O;
+ else
+ reg &= ~MDIO_MAN_MDCLK_O;
+
+ writel(reg, &data->regs->manualif);
+}
+
+static void davinci_set_mdio_dir(struct mdiobb_ctrl *ctrl, int output)
+{
+ struct davinci_mdio_data *data;
+ u32 reg;
+
+ data = container_of(ctrl, struct davinci_mdio_data, bb_ctrl);
+ reg = readl(&data->regs->manualif);
+
+ if (output)
+ reg |= MDIO_MAN_OE;
+ else
+ reg &= ~MDIO_MAN_OE;
+
+ writel(reg, &data->regs->manualif);
+}
+
+static void davinci_set_mdio_data(struct mdiobb_ctrl *ctrl, int value)
+{
+ struct davinci_mdio_data *data;
+ u32 reg;
+
+ data = container_of(ctrl, struct davinci_mdio_data, bb_ctrl);
+ reg = readl(&data->regs->manualif);
+
+ if (value)
+ reg |= MDIO_MAN_PIN;
+ else
+ reg &= ~MDIO_MAN_PIN;
+
+ writel(reg, &data->regs->manualif);
+}
+
+static int davinci_get_mdio_data(struct mdiobb_ctrl *ctrl)
+{
+ struct davinci_mdio_data *data;
+ unsigned long reg;
+
+ data = container_of(ctrl, struct davinci_mdio_data, bb_ctrl);
+ reg = readl(&data->regs->manualif);
+ return test_bit(MDIO_PIN, &reg);
+}
+
+static int davinci_mdiobb_read(struct mii_bus *bus, int phy, int reg)
+{
+ int ret;
+
+ ret = pm_runtime_resume_and_get(bus->parent);
+ if (ret < 0)
+ return ret;
+
+ ret = mdiobb_read(bus, phy, reg);
+
+ pm_runtime_mark_last_busy(bus->parent);
+ pm_runtime_put_autosuspend(bus->parent);
+
+ return ret;
+}
+
+static int davinci_mdiobb_write(struct mii_bus *bus, int phy, int reg,
+ u16 val)
+{
+ int ret;
+
+ ret = pm_runtime_resume_and_get(bus->parent);
+ if (ret < 0)
+ return ret;
+
+ ret = mdiobb_write(bus, phy, reg, val);
+
+ pm_runtime_mark_last_busy(bus->parent);
+ pm_runtime_put_autosuspend(bus->parent);
+
+ return ret;
+}
+
+static int davinci_mdio_common_reset(struct davinci_mdio_data *data)
{
- struct davinci_mdio_data *data = bus->priv;
u32 phy_mask, ver;
int ret;

@@ -138,6 +267,11 @@ static int davinci_mdio_reset(struct mii_bus *bus)
if (ret < 0)
return ret;

+ if (data->manual_mode) {
+ davinci_mdio_disable(data);
+ davinci_mdio_enable_manual_mode(data);
+ }
+
/* wait for scan logic to settle */
msleep(PHY_MAX_ADDR * data->access_time);

@@ -171,6 +305,23 @@ static int davinci_mdio_reset(struct mii_bus *bus)
return 0;
}

+static int davinci_mdio_reset(struct mii_bus *bus)
+{
+ struct davinci_mdio_data *data = bus->priv;
+
+ return davinci_mdio_common_reset(data);
+}
+
+static int davinci_mdiobb_reset(struct mii_bus *bus)
+{
+ struct mdiobb_ctrl *ctrl = bus->priv;
+ struct davinci_mdio_data *data;
+
+ data = container_of(ctrl, struct davinci_mdio_data, bb_ctrl);
+
+ return davinci_mdio_common_reset(data);
+}
+
/* wait until hardware is ready for another user access */
static inline int wait_for_user_access(struct davinci_mdio_data *data)
{
@@ -318,6 +469,28 @@ static int davinci_mdio_probe_dt(struct mdio_platform_data *data,
return 0;
}

+struct k3_mdio_soc_data {
+ bool manual_mode;
+};
+
+static const struct k3_mdio_soc_data am65_mdio_soc_data = {
+ .manual_mode = true,
+};
+
+static const struct soc_device_attribute k3_mdio_socinfo[] = {
+ { .family = "AM62X", .revision = "SR1.0", .data = &am65_mdio_soc_data },
+ { .family = "AM64X", .revision = "SR1.0", .data = &am65_mdio_soc_data },
+ { .family = "AM64X", .revision = "SR2.0", .data = &am65_mdio_soc_data },
+ { .family = "AM65X", .revision = "SR1.0", .data = &am65_mdio_soc_data },
+ { .family = "AM65X", .revision = "SR2.0", .data = &am65_mdio_soc_data },
+ { .family = "J7200", .revision = "SR1.0", .data = &am65_mdio_soc_data },
+ { .family = "J7200", .revision = "SR2.0", .data = &am65_mdio_soc_data },
+ { .family = "J721E", .revision = "SR1.0", .data = &am65_mdio_soc_data },
+ { .family = "J721E", .revision = "SR2.0", .data = &am65_mdio_soc_data },
+ { .family = "J721S2", .revision = "SR1.0", .data = &am65_mdio_soc_data},
+ { /* sentinel */ },
+};
+
#if IS_ENABLED(CONFIG_OF)
static const struct davinci_mdio_of_param of_cpsw_mdio_data = {
.autosuspend_delay_ms = 100,
@@ -331,6 +504,14 @@ static const struct of_device_id davinci_mdio_of_mtable[] = {
MODULE_DEVICE_TABLE(of, davinci_mdio_of_mtable);
#endif

+static const struct mdiobb_ops davinci_mdiobb_ops = {
+ .owner = THIS_MODULE,
+ .set_mdc = davinci_set_mdc,
+ .set_mdio_dir = davinci_set_mdio_dir,
+ .set_mdio_data = davinci_set_mdio_data,
+ .get_mdio_data = davinci_get_mdio_data,
+};
+
static int davinci_mdio_probe(struct platform_device *pdev)
{
struct mdio_platform_data *pdata = dev_get_platdata(&pdev->dev);
@@ -345,7 +526,26 @@ static int davinci_mdio_probe(struct platform_device *pdev)
if (!data)
return -ENOMEM;

- data->bus = devm_mdiobus_alloc(dev);
+ data->manual_mode = false;
+ data->bb_ctrl.ops = &davinci_mdiobb_ops;
+
+ if (IS_ENABLED(CONFIG_OF) && dev->of_node) {
+ const struct soc_device_attribute *soc_match_data;
+
+ soc_match_data = soc_device_match(k3_mdio_socinfo);
+ if (soc_match_data && soc_match_data->data) {
+ const struct k3_mdio_soc_data *socdata =
+ soc_match_data->data;
+
+ data->manual_mode = socdata->manual_mode;
+ }
+ }
+
+ if (data->manual_mode)
+ data->bus = alloc_mdio_bitbang(&data->bb_ctrl);
+ else
+ data->bus = devm_mdiobus_alloc(dev);
+
if (!data->bus) {
dev_err(dev, "failed to alloc mii bus\n");
return -ENOMEM;
@@ -371,11 +571,20 @@ static int davinci_mdio_probe(struct platform_device *pdev)
}

data->bus->name = dev_name(dev);
- data->bus->read = davinci_mdio_read;
- data->bus->write = davinci_mdio_write;
- data->bus->reset = davinci_mdio_reset;
+
+ if (data->manual_mode) {
+ data->bus->read = davinci_mdiobb_read;
+ data->bus->write = davinci_mdiobb_write;
+ data->bus->reset = davinci_mdiobb_reset;
+
+ dev_info(dev, "Configuring MDIO in manual mode\n");
+ } else {
+ data->bus->read = davinci_mdio_read;
+ data->bus->write = davinci_mdio_write;
+ data->bus->reset = davinci_mdio_reset;
+ data->bus->priv = data;
+ }
data->bus->parent = dev;
- data->bus->priv = data;

data->clk = devm_clk_get(dev, "fck");
if (IS_ERR(data->clk)) {
@@ -433,9 +642,13 @@ static int davinci_mdio_remove(struct platform_device *pdev)
{
struct davinci_mdio_data *data = platform_get_drvdata(pdev);

- if (data->bus)
+ if (data->bus) {
mdiobus_unregister(data->bus);

+ if (data->manual_mode)
+ free_mdio_bitbang(data->bus);
+ }
+
pm_runtime_dont_use_autosuspend(&pdev->dev);
pm_runtime_disable(&pdev->dev);

@@ -452,7 +665,9 @@ static int davinci_mdio_runtime_suspend(struct device *dev)
ctrl = readl(&data->regs->control);
ctrl &= ~CONTROL_ENABLE;
writel(ctrl, &data->regs->control);
- wait_for_idle(data);
+
+ if (!data->manual_mode)
+ wait_for_idle(data);

return 0;
}
@@ -461,7 +676,12 @@ static int davinci_mdio_runtime_resume(struct device *dev)
{
struct davinci_mdio_data *data = dev_get_drvdata(dev);

- davinci_mdio_enable(data);
+ if (data->manual_mode) {
+ davinci_mdio_disable(data);
+ davinci_mdio_enable_manual_mode(data);
+ } else {
+ davinci_mdio_enable(data);
+ }
return 0;
}
#endif
--
2.35.1

2022-10-10 00:04:52

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 27/73] net: sunplus: Fix return type for implementation of ndo_start_xmit

From: GUO Zihua <[email protected]>

[ Upstream commit 7b620e156097028e4c9b6481a84ec1e1e72877ca ]

Since Linux now supports CFI, it will be a good idea to fix mismatched
return type for implementation of hooks. Otherwise this might get
cought out by CFI and cause a panic.

spl2sw_ethernet_start_xmit() would return either NETDEV_TX_BUSY or
NETDEV_TX_OK, so change the return type to netdev_tx_t directly.

Signed-off-by: GUO Zihua <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/sunplus/spl2sw_driver.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/sunplus/spl2sw_driver.c b/drivers/net/ethernet/sunplus/spl2sw_driver.c
index 3773ce5e12cc..dbbb3727040f 100644
--- a/drivers/net/ethernet/sunplus/spl2sw_driver.c
+++ b/drivers/net/ethernet/sunplus/spl2sw_driver.c
@@ -62,7 +62,8 @@ static int spl2sw_ethernet_stop(struct net_device *ndev)
return 0;
}

-static int spl2sw_ethernet_start_xmit(struct sk_buff *skb, struct net_device *ndev)
+static netdev_tx_t spl2sw_ethernet_start_xmit(struct sk_buff *skb,
+ struct net_device *ndev)
{
struct spl2sw_mac *mac = netdev_priv(ndev);
struct spl2sw_common *comm = mac->comm;
--
2.35.1

2022-10-10 00:05:23

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 51/73] net: sfp: move Alcatel Lucent 3FE46541AA fixup

From: "Russell King (Oracle)" <[email protected]>

[ Upstream commit 275416754e9a262c97a1ad6f806a4bc6e0464aa2 ]

Add a new fixup mechanism to the SFP quirks, and use it for this
module.

Signed-off-by: Russell King (Oracle) <[email protected]>
Signed-off-by: Jakub Kicinski <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/phy/sfp.c | 14 +++++++++-----
drivers/net/phy/sfp.h | 1 +
2 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/drivers/net/phy/sfp.c b/drivers/net/phy/sfp.c
index 1b35cf0ac02a..38ce6c579ee3 100644
--- a/drivers/net/phy/sfp.c
+++ b/drivers/net/phy/sfp.c
@@ -311,6 +311,11 @@ static const struct of_device_id sfp_of_match[] = {
};
MODULE_DEVICE_TABLE(of, sfp_of_match);

+static void sfp_fixup_long_startup(struct sfp *sfp)
+{
+ sfp->module_t_start_up = T_START_UP_BAD_GPON;
+}
+
static void sfp_quirk_2500basex(const struct sfp_eeprom_id *id,
unsigned long *modes)
{
@@ -341,6 +346,7 @@ static const struct sfp_quirk sfp_quirks[] = {
.vendor = "ALCATELLUCENT",
.part = "3FE46541AA",
.modes = sfp_quirk_2500basex,
+ .fixup = sfp_fixup_long_startup,
}, {
// Huawei MA5671A can operate at 2500base-X, but report 1.2GBd
// NRZ in their EEPROM
@@ -2052,11 +2058,7 @@ static int sfp_sm_mod_probe(struct sfp *sfp, bool report)
if (sfp->gpio[GPIO_LOS])
sfp->state_hw_mask |= SFP_F_LOS;

- if (!memcmp(id.base.vendor_name, "ALCATELLUCENT ", 16) &&
- !memcmp(id.base.vendor_pn, "3FE46541AA ", 16))
- sfp->module_t_start_up = T_START_UP_BAD_GPON;
- else
- sfp->module_t_start_up = T_START_UP;
+ sfp->module_t_start_up = T_START_UP;

if (!memcmp(id.base.vendor_name, "HUAWEI ", 16) &&
!memcmp(id.base.vendor_pn, "MA5671A ", 16))
@@ -2065,6 +2067,8 @@ static int sfp_sm_mod_probe(struct sfp *sfp, bool report)
sfp->tx_fault_ignore = false;

sfp->quirk = sfp_lookup_quirk(&id);
+ if (sfp->quirk && sfp->quirk->fixup)
+ sfp->quirk->fixup(sfp);

return 0;
}
diff --git a/drivers/net/phy/sfp.h b/drivers/net/phy/sfp.h
index 03f1d47fe6ca..7ad06deae76c 100644
--- a/drivers/net/phy/sfp.h
+++ b/drivers/net/phy/sfp.h
@@ -10,6 +10,7 @@ struct sfp_quirk {
const char *vendor;
const char *part;
void (*modes)(const struct sfp_eeprom_id *id, unsigned long *modes);
+ void (*fixup)(struct sfp *sfp);
};

struct sfp_socket_ops {
--
2.35.1

2022-10-10 00:05:36

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 13/73] micrel: ksz8851: fixes struct pointer issue

From: Jerry Ray <[email protected]>

[ Upstream commit fef5de753ff01887cfa50990532c3890fccb9338 ]

Issue found during code review. This bug has no impact as long as the
ks8851_net structure is the first element of the ks8851_net_spi structure.
As long as the offset to the ks8851_net struct is zero, the container_of()
macro is subtracting 0 and therefore no damage done. But if the
ks8851_net_spi struct is ever modified such that the ks8851_net struct
within it is no longer the first element of the struct, then the bug would
manifest itself and cause problems.

struct ks8851_net is contained within ks8851_net_spi.
ks is contained within kss.
kss is the priv_data of the netdev structure.

Signed-off-by: Jerry Ray <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/micrel/ks8851_spi.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/micrel/ks8851_spi.c b/drivers/net/ethernet/micrel/ks8851_spi.c
index 82d55fc27edc..70bc7253454f 100644
--- a/drivers/net/ethernet/micrel/ks8851_spi.c
+++ b/drivers/net/ethernet/micrel/ks8851_spi.c
@@ -413,7 +413,8 @@ static int ks8851_probe_spi(struct spi_device *spi)

spi->bits_per_word = 8;

- ks = netdev_priv(netdev);
+ kss = netdev_priv(netdev);
+ ks = &kss->ks8851;

ks->lock = ks8851_lock_spi;
ks->unlock = ks8851_unlock_spi;
@@ -433,8 +434,6 @@ static int ks8851_probe_spi(struct spi_device *spi)
IRQ_RXPSI) /* RX process stop */
ks->rc_ier = STD_IRQ;

- kss = to_ks8851_spi(ks);
-
kss->spidev = spi;
mutex_init(&kss->lock);
INIT_WORK(&kss->tx_work, ks8851_tx_work);
--
2.35.1

2022-10-10 00:06:11

by Sasha Levin

[permalink] [raw]
Subject: [PATCH AUTOSEL 5.19 11/73] net: mscc: ocelot: adjust forwarding domain for CPU ports in a LAG

From: Vladimir Oltean <[email protected]>

[ Upstream commit 291ac1517af58670740528466ccebe3caefb9093 ]

Currently when we have 2 CPU ports configured for DSA tag_8021q mode and
we put them in a LAG, a PGID dump looks like this:

PGID_SRC[0] = ports 4,
PGID_SRC[1] = ports 4,
PGID_SRC[2] = ports 4,
PGID_SRC[3] = ports 4,
PGID_SRC[4] = ports 0, 1, 2, 3, 4, 5,
PGID_SRC[5] = no ports

(ports 0-3 are user ports, ports 4 and 5 are CPU ports)

There are 2 problems with the configuration above:

- user ports should enable forwarding towards both CPU ports, not just 4,
and the aggregation PGIDs should prune one CPU port or the other from
the destination port mask, based on a hash computed from packet headers.

- CPU ports should not be allowed to forward towards themselves and also
not towards other ports in the same LAG as themselves

The first problem requires fixing up the PGID_SRC of user ports, when
ocelot_port_assigned_dsa_8021q_cpu_mask() is called. We need to say that
when a user port is assigned to a tag_8021q CPU port and that port is in
a LAG, it should forward towards all ports in that LAG.

The second problem requires fixing up the PGID_SRC of port 4, to remove
ports 4 and 5 (in a LAG) from the allowed destinations.

After this change, the PGID source masks look as follows:

PGID_SRC[0] = ports 4, 5,
PGID_SRC[1] = ports 4, 5,
PGID_SRC[2] = ports 4, 5,
PGID_SRC[3] = ports 4, 5,
PGID_SRC[4] = ports 0, 1, 2, 3,
PGID_SRC[5] = no ports

Note that PGID_SRC[5] still looks weird (it should say "0, 1, 2, 3" just
like PGID_SRC[4] does), but I've tested forwarding through this CPU port
and it doesn't seem like anything is affected (it appears that PGID_SRC[4]
is being looked up on forwarding from the CPU, since both ports 4 and 5
have logical port ID 4). The reason why it looks weird is because
we've never called ocelot_port_assign_dsa_8021q_cpu() for any user port
towards port 5 (all user ports are assigned to port 4 which is in a LAG
with 5).

Since things aren't broken, I'm willing to leave it like that for now
and just document the oddity.

Signed-off-by: Vladimir Oltean <[email protected]>
Signed-off-by: Paolo Abeni <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
drivers/net/ethernet/mscc/ocelot.c | 19 +++++++++++++++++++
1 file changed, 19 insertions(+)

diff --git a/drivers/net/ethernet/mscc/ocelot.c b/drivers/net/ethernet/mscc/ocelot.c
index 68991b021c56..f6ec299a2b5a 100644
--- a/drivers/net/ethernet/mscc/ocelot.c
+++ b/drivers/net/ethernet/mscc/ocelot.c
@@ -2065,6 +2065,16 @@ static int ocelot_bond_get_id(struct ocelot *ocelot, struct net_device *bond)
return __ffs(bond_mask);
}

+/* Returns the mask of user ports assigned to this DSA tag_8021q CPU port.
+ * Note that when CPU ports are in a LAG, the user ports are assigned to the
+ * 'primary' CPU port, the one whose physical port number gives the logical
+ * port number of the LAG.
+ *
+ * We leave PGID_SRC poorly configured for the 'secondary' CPU port in the LAG
+ * (to which no user port is assigned), but it appears that forwarding from
+ * this secondary CPU port looks at the PGID_SRC associated with the logical
+ * port ID that it's assigned to, which *is* configured properly.
+ */
static u32 ocelot_dsa_8021q_cpu_assigned_ports(struct ocelot *ocelot,
struct ocelot_port *cpu)
{
@@ -2081,9 +2091,15 @@ static u32 ocelot_dsa_8021q_cpu_assigned_ports(struct ocelot *ocelot,
mask |= BIT(port);
}

+ if (cpu->bond)
+ mask &= ~ocelot_get_bond_mask(ocelot, cpu->bond);
+
return mask;
}

+/* Returns the DSA tag_8021q CPU port that the given port is assigned to,
+ * or the bit mask of CPU ports if said CPU port is in a LAG.
+ */
u32 ocelot_port_assigned_dsa_8021q_cpu_mask(struct ocelot *ocelot, int port)
{
struct ocelot_port *ocelot_port = ocelot->ports[port];
@@ -2092,6 +2108,9 @@ u32 ocelot_port_assigned_dsa_8021q_cpu_mask(struct ocelot *ocelot, int port)
if (!cpu_port)
return 0;

+ if (cpu_port->bond)
+ return ocelot_get_bond_mask(ocelot, cpu_port->bond);
+
return BIT(cpu_port->index);
}
EXPORT_SYMBOL_GPL(ocelot_port_assigned_dsa_8021q_cpu_mask);
--
2.35.1

2022-10-10 12:19:59

by Vladimir Oltean

[permalink] [raw]
Subject: Re: [PATCH AUTOSEL 5.19 34/73] net: mscc: ocelot: report FIFO drop counters through stats->rx_dropped

On Sun, Oct 09, 2022 at 06:14:12PM -0400, Sasha Levin wrote:
> From: Vladimir Oltean <[email protected]>
>
> [ Upstream commit cc160fc29a264726b2bfbc2f551081430db3df03 ]
>
> if_link.h says:
>
> * @rx_dropped: Number of packets received but not processed,
> * e.g. due to lack of resources or unsupported protocol.
> * For hardware interfaces this counter may include packets discarded
> * due to L2 address filtering but should not include packets dropped
> * by the device due to buffer exhaustion which are counted separately in
> * @rx_missed_errors (since procfs folds those two counters together).
>
> Currently we report "stats->rx_dropped = dev->stats.rx_dropped", the
> latter being incremented by various entities in the stack. This is not
> wrong, but we'd like to move ocelot_get_stats64() in the common ocelot
> switch lib which is independent of struct net_device.
>
> To do that, report the hardware RX drop counters instead. These drops
> are due to policer action, or due to no destinations. When we have no
> memory in the queue system, report this through rx_missed_errors, as
> instructed.
>
> Signed-off-by: Vladimir Oltean <[email protected]>
> Signed-off-by: David S. Miller <[email protected]>
> Signed-off-by: Sasha Levin <[email protected]>
> ---

Not needed for stable kernels, please drop, thanks.

2022-10-10 12:21:52

by Vladimir Oltean

[permalink] [raw]
Subject: Re: [PATCH AUTOSEL 5.19 11/73] net: mscc: ocelot: adjust forwarding domain for CPU ports in a LAG

On Sun, Oct 09, 2022 at 06:13:49PM -0400, Sasha Levin wrote:
> From: Vladimir Oltean <[email protected]>
>
> [ Upstream commit 291ac1517af58670740528466ccebe3caefb9093 ]
>
> Currently when we have 2 CPU ports configured for DSA tag_8021q mode and
> we put them in a LAG, a PGID dump looks like this:
>
> PGID_SRC[0] = ports 4,
> PGID_SRC[1] = ports 4,
> PGID_SRC[2] = ports 4,
> PGID_SRC[3] = ports 4,
> PGID_SRC[4] = ports 0, 1, 2, 3, 4, 5,
> PGID_SRC[5] = no ports
>
> (ports 0-3 are user ports, ports 4 and 5 are CPU ports)
>
> There are 2 problems with the configuration above:
>
> - user ports should enable forwarding towards both CPU ports, not just 4,
> and the aggregation PGIDs should prune one CPU port or the other from
> the destination port mask, based on a hash computed from packet headers.
>
> - CPU ports should not be allowed to forward towards themselves and also
> not towards other ports in the same LAG as themselves
>
> The first problem requires fixing up the PGID_SRC of user ports, when
> ocelot_port_assigned_dsa_8021q_cpu_mask() is called. We need to say that
> when a user port is assigned to a tag_8021q CPU port and that port is in
> a LAG, it should forward towards all ports in that LAG.
>
> The second problem requires fixing up the PGID_SRC of port 4, to remove
> ports 4 and 5 (in a LAG) from the allowed destinations.
>
> After this change, the PGID source masks look as follows:
>
> PGID_SRC[0] = ports 4, 5,
> PGID_SRC[1] = ports 4, 5,
> PGID_SRC[2] = ports 4, 5,
> PGID_SRC[3] = ports 4, 5,
> PGID_SRC[4] = ports 0, 1, 2, 3,
> PGID_SRC[5] = no ports
>
> Note that PGID_SRC[5] still looks weird (it should say "0, 1, 2, 3" just
> like PGID_SRC[4] does), but I've tested forwarding through this CPU port
> and it doesn't seem like anything is affected (it appears that PGID_SRC[4]
> is being looked up on forwarding from the CPU, since both ports 4 and 5
> have logical port ID 4). The reason why it looks weird is because
> we've never called ocelot_port_assign_dsa_8021q_cpu() for any user port
> towards port 5 (all user ports are assigned to port 4 which is in a LAG
> with 5).
>
> Since things aren't broken, I'm willing to leave it like that for now
> and just document the oddity.
>
> Signed-off-by: Vladimir Oltean <[email protected]>
> Signed-off-by: Paolo Abeni <[email protected]>
> Signed-off-by: Sasha Levin <[email protected]>
> ---

Not needed for stable kernels, please drop, thanks.