* Re: [virtio-dev] Re: [Qemu-devel] [PATCH] qemu: Introduce VIRTIO_NET_F_STANDBY feature bit to virtio_net
From: Cornelia Huck @ 2018-06-20 9:53 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: Alexander Duyck, virtio-dev, Jiri Pirko, konrad.wilk,
Jakub Kicinski, Samudrala, Sridhar, qemu-devel, virtualization,
Siwei Liu, Venu Busireddy, Netdev, boris.ostrovsky, aaron.f.brown,
Joao Martins
In-Reply-To: <20180619233001-mutt-send-email-mst@kernel.org>
On Tue, 19 Jun 2018 23:32:06 +0300
"Michael S. Tsirkin" <mst@redhat.com> wrote:
> On Tue, Jun 19, 2018 at 12:54:53PM +0200, Cornelia Huck wrote:
> > Sorry about dragging mainframes into this, but this will only work for
> > homogenous device coupling, not for heterogenous. Consider my vfio-pci
> > + virtio-net-ccw example again: The guest cannot find out that the two
> > belong together by checking some group ID, it has to either use the MAC
> > or some needs-to-be-architectured property.
> >
> > Alternatively, we could propose that mechanism as pci-only, which means
> > we can rely on mechanisms that won't necessarily work on non-pci
> > transports. (FWIW, I don't see a use case for using vfio-ccw to pass
> > through a network card anytime in the near future, due to the nature of
> > network cards currently in use on s390.)
>
> That's what it boils down to, yes. If there's need to have this for
> non-pci devices, then we should put it in config space.
> Cornelia, what do you think?
>
I think the only really useful config on s390 is the vfio-pci network
card coupled with a virtio-net-ccw device: Using an s390 network card
via vfio-ccw is out due to the nature of the s390 network cards, and
virtio-ccw is the default transport (virtio-pci is not supported on any
enterprise distro AFAIK).
For this, having a uuid in the config space could work (vfio-pci
devices have a config space by virtue of being pci devices, and
virtio-net-ccw devices have a config space by virtue of being virtio
devices -- ccw devices usually don't have that concept).
^ permalink raw reply
* Re: [PATCH][v2] xfrm: replace NR_CPU with nr_cpu_ids
From: Florian Westphal @ 2018-06-20 9:55 UTC (permalink / raw)
To: Steffen Klassert; +Cc: Florian Westphal, Li RongQing, netdev
In-Reply-To: <20180619121755.3weaw5lsp4sho23r@gauss3.secunet.de>
Steffen Klassert <steffen.klassert@secunet.com> wrote:
> On Tue, Jun 19, 2018 at 09:53:49AM +0200, Florian Westphal wrote:
> > Li RongQing <lirongqing@baidu.com> wrote:
> > > The default NR_CPUS can be very large, but actual possible nr_cpu_ids
> > > usually is very small. For some x86 distribution, the NR_CPUS is 8192
> > > and nr_cpu_ids is 4, so replace NR_CPU to save some memory
> >
> > Steffen,
> >
> > I will soon submit a patch to remove the percpu cache; removal
> > improved performance for at least one user (and by quite a sizeable
> > amount).
> >
> > Would you consider such removal for ipsec or ipsec-next?
>
> I think this removel would better fit to ipsec-next.
Agree, it slows things down further for me in my tests.
Problem is that I get quite good re-use of pcpu cache due to
unidirectional flows and only one tunnel.
I suspect that even with tunnel the removal is a win in practice
though, netperf is quite artifical, so I rather trust Kristians results
(real world) than my own.
> considered to apply it to ipsec-next. If you plan
> to remove it, I'll wait for that.
I'll submit once net-next opens.
^ permalink raw reply
* [PATCH] ceph: use ktime_get_real_seconds()
From: Allen Pais @ 2018-06-20 10:00 UTC (permalink / raw)
To: netdev; +Cc: davem, linux-kernel, ceph-devel, y2038, Allen Pais
Use ktime_get_real_seconds() as get_seconds() is deprecated.
Signed-off-by: Allen Pais <allen.lkml@gmail.com>
---
net/ceph/auth_x.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/net/ceph/auth_x.c b/net/ceph/auth_x.c
index 2f4a1ba..99561c2 100644
--- a/net/ceph/auth_x.c
+++ b/net/ceph/auth_x.c
@@ -154,7 +154,7 @@ static int process_one_ticket(struct ceph_auth_client *ac,
void **ptp;
struct ceph_crypto_key new_session_key = { 0 };
struct ceph_buffer *new_ticket_blob;
- unsigned long new_expires, new_renew_after;
+ u32 new_expires, new_renew_after;
u64 new_secret_id;
int ret;
@@ -191,9 +191,9 @@ static int process_one_ticket(struct ceph_auth_client *ac,
ceph_decode_timespec(&validity, dp);
dp += sizeof(struct ceph_timespec);
- new_expires = get_seconds() + validity.tv_sec;
+ new_expires = (u32)ktime_get_real_seconds() + validity.tv_sec;
new_renew_after = new_expires - (validity.tv_sec / 4);
- dout(" expires=%lu renew_after=%lu\n", new_expires,
+ dout(" expires=%u renew_after=%u\n", new_expires,
new_renew_after);
/* ticket blob for service */
@@ -385,13 +385,13 @@ static bool need_key(struct ceph_x_ticket_handler *th)
if (!th->have_key)
return true;
- return get_seconds() >= th->renew_after;
+ return ktime_get_real_seconds() >= th->renew_after;
}
static bool have_key(struct ceph_x_ticket_handler *th)
{
if (th->have_key) {
- if (get_seconds() >= th->expires)
+ if (ktime_get_real_seconds() >= th->expires)
th->have_key = false;
}
--
2.7.4
^ permalink raw reply related
* Re: Grant
From: Maratovich M. Fridman @ 2018-06-20 7:34 UTC (permalink / raw)
--
I Mikhail Fridman. has selected you specially as one of my beneficiaries
for my Charitable Donation, Just as I have declared on May 23, 2016 to give
my fortune as charity.
Check the link below for confirmation:
http://www.ibtimes.co.uk/russias-second-wealthiest-man-mikhail-fridman-plans-leaving-14-2bn-fortune-charity-1561604
Reply as soon as possible with further directives.
Best Regards,
Mikhail Fridman.
^ permalink raw reply
* ISDN: use irqsave() in URB completion + usb_fill_int_urb
From: Sebastian Andrzej Siewior @ 2018-06-20 10:40 UTC (permalink / raw)
To: netdev; +Cc: Karsten Keil, linux-usb, tglx
This series is mostly about using _irqsave() primitives in the
completion callback in order to get rid of local_irq_save() in
__usb_hcd_giveback_urb(). While at it, I also tried to move drivers to
use usb_fill_int_urb() otherwise it is hard find users of a certain API.
Sebastian
^ permalink raw reply
* [PATCH 1/4] isdn: gigaset: use usb_fill_int_urb()
From: Sebastian Andrzej Siewior @ 2018-06-20 10:40 UTC (permalink / raw)
To: netdev
Cc: Karsten Keil, linux-usb, tglx, Sebastian Andrzej Siewior,
Paul Bolle, gigaset307x-common
In-Reply-To: <20180620104028.18283-1-bigeasy@linutronix.de>
Using usb_fill_int_urb() helps to find code which initializes an
URB. A grep for members of the struct (like ->complete) reveal lots
of other things, too.
Cc: Paul Bolle <pebolle@tiscali.nl>
Cc: Karsten Keil <isdn@linux-pingi.de>
Cc: gigaset307x-common@lists.sourceforge.net
Cc: netdev@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
drivers/isdn/gigaset/bas-gigaset.c | 26 ++++++++++++--------------
1 file changed, 12 insertions(+), 14 deletions(-)
diff --git a/drivers/isdn/gigaset/bas-gigaset.c b/drivers/isdn/gigaset/bas-gigaset.c
index 20d0a080a2b0..40c141163f0f 100644
--- a/drivers/isdn/gigaset/bas-gigaset.c
+++ b/drivers/isdn/gigaset/bas-gigaset.c
@@ -972,16 +972,14 @@ static int starturbs(struct bc_state *bcs)
rc = -EFAULT;
goto error;
}
+ usb_fill_int_urb(urb, bcs->cs->hw.bas->udev,
+ usb_rcvisocpipe(urb->dev, 3 + 2 * bcs->channel),
+ ubc->isoinbuf + k * BAS_INBUFSIZE,
+ BAS_INBUFSIZE, read_iso_callback, bcs,
+ BAS_FRAMETIME);
- urb->dev = bcs->cs->hw.bas->udev;
- urb->pipe = usb_rcvisocpipe(urb->dev, 3 + 2 * bcs->channel);
urb->transfer_flags = URB_ISO_ASAP;
- urb->transfer_buffer = ubc->isoinbuf + k * BAS_INBUFSIZE;
- urb->transfer_buffer_length = BAS_INBUFSIZE;
urb->number_of_packets = BAS_NUMFRAMES;
- urb->interval = BAS_FRAMETIME;
- urb->complete = read_iso_callback;
- urb->context = bcs;
for (j = 0; j < BAS_NUMFRAMES; j++) {
urb->iso_frame_desc[j].offset = j * BAS_MAXFRAME;
urb->iso_frame_desc[j].length = BAS_MAXFRAME;
@@ -1005,15 +1003,15 @@ static int starturbs(struct bc_state *bcs)
rc = -EFAULT;
goto error;
}
- urb->dev = bcs->cs->hw.bas->udev;
- urb->pipe = usb_sndisocpipe(urb->dev, 4 + 2 * bcs->channel);
+ usb_fill_int_urb(urb, bcs->cs->hw.bas->udev,
+ usb_sndisocpipe(urb->dev, 4 + 2 * bcs->channel),
+ ubc->isooutbuf->data,
+ sizeof(ubc->isooutbuf->data),
+ write_iso_callback, &ubc->isoouturbs[k],
+ BAS_FRAMETIME);
+
urb->transfer_flags = URB_ISO_ASAP;
- urb->transfer_buffer = ubc->isooutbuf->data;
- urb->transfer_buffer_length = sizeof(ubc->isooutbuf->data);
urb->number_of_packets = BAS_NUMFRAMES;
- urb->interval = BAS_FRAMETIME;
- urb->complete = write_iso_callback;
- urb->context = &ubc->isoouturbs[k];
for (j = 0; j < BAS_NUMFRAMES; ++j) {
urb->iso_frame_desc[j].offset = BAS_OUTBUFSIZE;
urb->iso_frame_desc[j].length = BAS_NORMFRAME;
--
2.17.1
^ permalink raw reply related
* [PATCH 2/4] isdn: hisax: hfc_usb: use usb_fill_int_urb()
From: Sebastian Andrzej Siewior @ 2018-06-20 10:40 UTC (permalink / raw)
To: netdev; +Cc: Karsten Keil, linux-usb, tglx, Sebastian Andrzej Siewior
In-Reply-To: <20180620104028.18283-1-bigeasy@linutronix.de>
Using usb_fill_int_urb() helps to find code which initializes an
URB. A grep for members of the struct (like ->complete) reveal lots
of other things, too.
The `interval' parameter is now set differently on HS and SS. The
argument is fed from bInterval so it should be the right thing to do.
Cc: Karsten Keil <isdn@linux-pingi.de>
Cc: netdev@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
drivers/isdn/hisax/hfc_usb.c | 10 +++-------
1 file changed, 3 insertions(+), 7 deletions(-)
diff --git a/drivers/isdn/hisax/hfc_usb.c b/drivers/isdn/hisax/hfc_usb.c
index 97ecb3073045..1d4cd01d4685 100644
--- a/drivers/isdn/hisax/hfc_usb.c
+++ b/drivers/isdn/hisax/hfc_usb.c
@@ -432,16 +432,12 @@ fill_isoc_urb(struct urb *urb, struct usb_device *dev, unsigned int pipe,
{
int k;
- urb->dev = dev;
- urb->pipe = pipe;
- urb->complete = complete;
+ usb_fill_int_urb(urb, dev, pipe, buf, packet_size * num_packets,
+ complete, context, interval);
+
urb->number_of_packets = num_packets;
- urb->transfer_buffer_length = packet_size * num_packets;
- urb->context = context;
- urb->transfer_buffer = buf;
urb->transfer_flags = URB_ISO_ASAP;
urb->actual_length = 0;
- urb->interval = interval;
for (k = 0; k < num_packets; k++) {
urb->iso_frame_desc[k].offset = packet_size * k;
urb->iso_frame_desc[k].length = packet_size;
--
2.17.1
^ permalink raw reply related
* [PATCH 3/4] isdn: hisax: st5481_usb: use usb_fill_int_urb()
From: Sebastian Andrzej Siewior @ 2018-06-20 10:40 UTC (permalink / raw)
To: netdev; +Cc: Karsten Keil, linux-usb, tglx, Sebastian Andrzej Siewior
In-Reply-To: <20180620104028.18283-1-bigeasy@linutronix.de>
Using usb_fill_int_urb() helps to find code which initializes an
URB. A grep for members of the struct (like ->complete) reveal lots
of other things, too.
Cc: Karsten Keil <isdn@linux-pingi.de>
Cc: netdev@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
drivers/isdn/hisax/st5481_usb.c | 11 +++--------
1 file changed, 3 insertions(+), 8 deletions(-)
diff --git a/drivers/isdn/hisax/st5481_usb.c b/drivers/isdn/hisax/st5481_usb.c
index 1cb9930d5e24..f207fda691c7 100644
--- a/drivers/isdn/hisax/st5481_usb.c
+++ b/drivers/isdn/hisax/st5481_usb.c
@@ -408,15 +408,10 @@ fill_isoc_urb(struct urb *urb, struct usb_device *dev,
{
int k;
- urb->dev = dev;
- urb->pipe = pipe;
- urb->interval = 1;
- urb->transfer_buffer = buf;
+ usb_fill_int_urb(urb, dev, pipe, buf, num_packets * packet_size,
+ complete, context, 1);
+
urb->number_of_packets = num_packets;
- urb->transfer_buffer_length = num_packets * packet_size;
- urb->actual_length = 0;
- urb->complete = complete;
- urb->context = context;
urb->transfer_flags = URB_ISO_ASAP;
for (k = 0; k < num_packets; k++) {
urb->iso_frame_desc[k].offset = packet_size * k;
--
2.17.1
^ permalink raw reply related
* [PATCH 4/4] isdn: mISDN: use irqsave() in USB's complete callback
From: Sebastian Andrzej Siewior @ 2018-06-20 10:40 UTC (permalink / raw)
To: netdev; +Cc: Karsten Keil, linux-usb, tglx, Sebastian Andrzej Siewior
In-Reply-To: <20180620104028.18283-1-bigeasy@linutronix.de>
The USB completion callback does not disable interrupts while acquiring
the ->lock. We want to remove the local_irq_disable() invocation from
__usb_hcd_giveback_urb() and therefore it is required for the callback
handler to disable the interrupts while acquiring the lock.
The callback may be invoked either in IRQ or BH context depending on the
USB host controller.
Use the _irqsave() variant of the locking primitives.
Cc: Karsten Keil <isdn@linux-pingi.de>
Cc: netdev@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
drivers/isdn/hardware/mISDN/hfcsusb.c | 36 +++++++++++++++------------
1 file changed, 20 insertions(+), 16 deletions(-)
diff --git a/drivers/isdn/hardware/mISDN/hfcsusb.c b/drivers/isdn/hardware/mISDN/hfcsusb.c
index 17cc879ad2bb..6d05946b445e 100644
--- a/drivers/isdn/hardware/mISDN/hfcsusb.c
+++ b/drivers/isdn/hardware/mISDN/hfcsusb.c
@@ -819,6 +819,7 @@ hfcsusb_rx_frame(struct usb_fifo *fifo, __u8 *data, unsigned int len,
int fifon = fifo->fifonum;
int i;
int hdlc = 0;
+ unsigned long flags;
if (debug & DBG_HFC_CALL_TRACE)
printk(KERN_DEBUG "%s: %s: fifo(%i) len(%i) "
@@ -835,7 +836,7 @@ hfcsusb_rx_frame(struct usb_fifo *fifo, __u8 *data, unsigned int len,
return;
}
- spin_lock(&hw->lock);
+ spin_lock_irqsave(&hw->lock, flags);
if (fifo->dch) {
rx_skb = fifo->dch->rx_skb;
maxlen = fifo->dch->maxlen;
@@ -844,7 +845,7 @@ hfcsusb_rx_frame(struct usb_fifo *fifo, __u8 *data, unsigned int len,
if (fifo->bch) {
if (test_bit(FLG_RX_OFF, &fifo->bch->Flags)) {
fifo->bch->dropcnt += len;
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
return;
}
maxlen = bchannel_get_rxbuf(fifo->bch, len);
@@ -854,7 +855,7 @@ hfcsusb_rx_frame(struct usb_fifo *fifo, __u8 *data, unsigned int len,
skb_trim(rx_skb, 0);
pr_warning("%s.B%d: No bufferspace for %d bytes\n",
hw->name, fifo->bch->nr, len);
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
return;
}
maxlen = fifo->bch->maxlen;
@@ -878,7 +879,7 @@ hfcsusb_rx_frame(struct usb_fifo *fifo, __u8 *data, unsigned int len,
} else {
printk(KERN_DEBUG "%s: %s: No mem for rx_skb\n",
hw->name, __func__);
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
return;
}
}
@@ -888,7 +889,7 @@ hfcsusb_rx_frame(struct usb_fifo *fifo, __u8 *data, unsigned int len,
"for fifo(%d) HFCUSB_D_RX\n",
hw->name, __func__, fifon);
skb_trim(rx_skb, 0);
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
return;
}
}
@@ -942,7 +943,7 @@ hfcsusb_rx_frame(struct usb_fifo *fifo, __u8 *data, unsigned int len,
/* deliver transparent data to layer2 */
recv_Bchannel(fifo->bch, MISDN_ID_ANY, false);
}
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
}
static void
@@ -979,18 +980,19 @@ rx_iso_complete(struct urb *urb)
__u8 *buf;
static __u8 eof[8];
__u8 s0_state;
+ unsigned long flags;
fifon = fifo->fifonum;
status = urb->status;
- spin_lock(&hw->lock);
+ spin_lock_irqsave(&hw->lock, flags);
if (fifo->stop_gracefull) {
fifo->stop_gracefull = 0;
fifo->active = 0;
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
return;
}
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
/*
* ISO transfer only partially completed,
@@ -1096,15 +1098,16 @@ rx_int_complete(struct urb *urb)
struct usb_fifo *fifo = (struct usb_fifo *) urb->context;
struct hfcsusb *hw = fifo->hw;
static __u8 eof[8];
+ unsigned long flags;
- spin_lock(&hw->lock);
+ spin_lock_irqsave(&hw->lock, flags);
if (fifo->stop_gracefull) {
fifo->stop_gracefull = 0;
fifo->active = 0;
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
return;
}
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
fifon = fifo->fifonum;
if ((!fifo->active) || (urb->status)) {
@@ -1172,12 +1175,13 @@ tx_iso_complete(struct urb *urb)
int *tx_idx;
int frame_complete, fifon, status, fillempty = 0;
__u8 threshbit, *p;
+ unsigned long flags;
- spin_lock(&hw->lock);
+ spin_lock_irqsave(&hw->lock, flags);
if (fifo->stop_gracefull) {
fifo->stop_gracefull = 0;
fifo->active = 0;
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
return;
}
@@ -1195,7 +1199,7 @@ tx_iso_complete(struct urb *urb)
} else {
printk(KERN_DEBUG "%s: %s: neither BCH nor DCH\n",
hw->name, __func__);
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
return;
}
@@ -1375,7 +1379,7 @@ tx_iso_complete(struct urb *urb)
hw->name, __func__,
symbolic(urb_errlist, status), status, fifon);
}
- spin_unlock(&hw->lock);
+ spin_unlock_irqrestore(&hw->lock, flags);
}
/*
--
2.17.1
^ permalink raw reply related
* [PATCH v4 3/5] batman-adv: use BIT_ULL for NL80211_STA_INFO_* attribute types
From: Omer Efrat @ 2018-06-20 10:46 UTC (permalink / raw)
To: linux-wireless-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
b.a.t.m.a.n-ZwoEplunGu2X36UT3dwllkB+6BGkLq7r
Cc: Omer Efrat
The BIT macro uses unsigned long which some architectures handle as 32 bit
and therefore might cause macro's shift to overflow when used on a value
equals or larger than 32 (NL80211_STA_INFO_RX_DURATION and afterwards).
Since 'filled' member in station_info changed to u64, BIT_ULL macro
should be used with all NL80211_STA_INFO_* attribute types instead of BIT
to prevent future possible bugs when one will use BIT macro for higher
attributes by mistake.
This commit cleans up all usages of BIT macro with the above field
in batman-adv by changing it to BIT_ULL instead.
Signed-off-by: Omer Efrat <omer.efrat-CtGflUZwD1xBDgjK7y7TUQ@public.gmane.org>
---
net/batman-adv/bat_v_elp.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/batman-adv/bat_v_elp.c b/net/batman-adv/bat_v_elp.c
index 71c20c1..2f0de55 100644
--- a/net/batman-adv/bat_v_elp.c
+++ b/net/batman-adv/bat_v_elp.c
@@ -114,7 +114,8 @@ static u32 batadv_v_elp_get_throughput(struct batadv_hardif_neigh_node *neigh)
}
if (ret)
goto default_throughput;
- if (!(sinfo.filled & BIT(NL80211_STA_INFO_EXPECTED_THROUGHPUT)))
+ if (!(sinfo.filled &
+ BIT_ULL(NL80211_STA_INFO_EXPECTED_THROUGHPUT)))
goto default_throughput;
return sinfo.expected_throughput / 100;
--
2.7.4
^ permalink raw reply related
* Re: [PATCH net] net: sungem: fix rx checksum support
From: Mathieu Malaterre @ 2018-06-20 11:22 UTC (permalink / raw)
To: David S. Miller; +Cc: Eric Dumazet, netdev, Meelis Roos, schwab, eric.dumazet
In-Reply-To: <20180620.143050.313454768369559179.davem@davemloft.net>
On Wed, Jun 20, 2018 at 7:31 AM David Miller <davem@davemloft.net> wrote:
>
> From: Eric Dumazet <edumazet@google.com>
> Date: Tue, 19 Jun 2018 19:18:50 -0700
>
> > After commit 88078d98d1bb ("net: pskb_trim_rcsum() and CHECKSUM_COMPLETE
> > are friends"), sungem owners reported the infamous "eth0: hw csum failure"
> > message.
> >
> > CHECKSUM_COMPLETE has in fact never worked for this driver, but this
> > was masked by the fact that upper stacks had to strip the FCS, and
> > therefore skb->ip_summed was set back to CHECKSUM_NONE before
> > my recent change.
> >
> > Driver configures a number of bytes to skip when the chip computes
> > the checksum, and for some reason only half of the Ethernet header
> > was skipped.
> >
> > Then a second problem is that we should strip the FCS by default,
> > unless the driver is updated to eventually support NETIF_F_RXFCS in
> > the future.
> >
> > Finally, a driver should check if NETIF_F_RXCSUM feature is enabled
> > or not, so that the admin can turn off rx checksum if wanted.
> >
> > Many thanks to Andreas Schwab and Mathieu Malaterre for their
> > help in debugging this issue.
> >
> > Signed-off-by: Eric Dumazet <edumazet@google.com>
> > Reported-by: Meelis Roos <mroos@linux.ee>
> > Reported-by: Mathieu Malaterre <malat@debian.org>
> > Reported-by: Andreas Schwab <schwab@linux-m68k.org>
> > Tested-by: Andreas Schwab <schwab@linux-m68k.org>
>
> Applied and queued up for -stable, thanks Eric.
Thanks for the stable tag.
IMHO the commit message should have also reference commit 7ce5a27f2ef8
("Revert "net: Handle CHECKSUM_COMPLETE more adequately in
pskb_trim_rcsum().""). Which means that commit 88078d98d1bb ("net:
pskb_trim_rcsum() and CHECKSUM_COMPLETE are friends") should be ready
for stable also, since it seems that the only driver responsible for
the revert was sungem.
my 2cts
^ permalink raw reply
* Re: [PATCH net] net: mscc: fix the injection header
From: Alexandre Belloni @ 2018-06-20 11:29 UTC (permalink / raw)
To: Antoine Tenart
Cc: davem, f.fainelli, andrew, netdev, linux-kernel, thomas.petazzoni,
quentin.schulz, allan.nielsen
In-Reply-To: <20180620085046.2377-1-antoine.tenart@bootlin.com>
On 20/06/2018 10:50:46+0200, Antoine Ténart wrote:
> When injecting frames in the Ocelot switch driver an injection header
> (IFH) should be used to configure various parameters related to a given
> frame, such as the port onto which the frame should be departed or its
> vlan id. Other parameters in the switch configuration can led to an
> injected frame being sent without an IFH but this led to various issues
> as the per-frame parameters are then not used. This is especially true
> when using multiple ports for injection.
>
> The IFH was injected with the wrong endianness which led to the switch
> not taking it into account as the IFH_INJ_BYPASS bit was then unset.
> (The bit tells the switch to use the IFH over its internal
> configuration). This patch fixes it.
>
> In addition to the endianness fix, the IFH is also fixed. As it was
> (unwillingly) unused, some of its fields were not configured the right
> way.
>
> Fixes: a556c76adc05 ("net: mscc: Add initial Ocelot switch support")
> Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
Reviewed-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
> ---
> drivers/net/ethernet/mscc/ocelot.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/net/ethernet/mscc/ocelot.c b/drivers/net/ethernet/mscc/ocelot.c
> index fb2c8f8071e6..52c57e0ff617 100644
> --- a/drivers/net/ethernet/mscc/ocelot.c
> +++ b/drivers/net/ethernet/mscc/ocelot.c
> @@ -344,10 +344,9 @@ static int ocelot_port_stop(struct net_device *dev)
> static int ocelot_gen_ifh(u32 *ifh, struct frame_info *info)
> {
> ifh[0] = IFH_INJ_BYPASS;
> - ifh[1] = (0xff00 & info->port) >> 8;
> + ifh[1] = (0xf00 & info->port) >> 8;
> ifh[2] = (0xff & info->port) << 24;
> - ifh[3] = IFH_INJ_POP_CNT_DISABLE | (info->cpuq << 20) |
> - (info->tag_type << 16) | info->vid;
> + ifh[3] = (info->tag_type << 16) | info->vid;
>
> return 0;
> }
> @@ -370,11 +369,12 @@ static int ocelot_port_xmit(struct sk_buff *skb, struct net_device *dev)
> QS_INJ_CTRL_SOF, QS_INJ_CTRL, grp);
>
> info.port = BIT(port->chip_port);
> - info.cpuq = 0xff;
> + info.tag_type = IFH_TAG_TYPE_C;
> + info.vid = skb_vlan_tag_get(skb);
> ocelot_gen_ifh(ifh, &info);
>
> for (i = 0; i < IFH_LEN; i++)
> - ocelot_write_rix(ocelot, ifh[i], QS_INJ_WR, grp);
> + ocelot_write_rix(ocelot, cpu_to_be32(ifh[i]), QS_INJ_WR, grp);
>
> count = (skb->len + 3) / 4;
> last = skb->len % 4;
> --
> 2.17.1
>
--
Alexandre Belloni, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com
^ permalink raw reply
* Re: [PATCH v4 3/5] batman-adv: use BIT_ULL for NL80211_STA_INFO_* attribute types
From: Sven Eckelmann @ 2018-06-20 11:34 UTC (permalink / raw)
To: b.a.t.m.a.n-ZwoEplunGu2X36UT3dwllkB+6BGkLq7r
Cc: netdev-u79uwXL29TY76Z2rM5mHXA,
linux-wireless-u79uwXL29TY76Z2rM5mHXA, Omer Efrat
In-Reply-To: <1529491616-23222-1-git-send-email-omer.efrat-CtGflUZwD1xBDgjK7y7TUQ@public.gmane.org>
[-- Attachment #1: Type: text/plain, Size: 438 bytes --]
On Mittwoch, 20. Juni 2018 13:46:56 CEST Omer Efrat wrote:
> - if (!(sinfo.filled & BIT(NL80211_STA_INFO_EXPECTED_THROUGHPUT)))
> + if (!(sinfo.filled &
> + BIT_ULL(NL80211_STA_INFO_EXPECTED_THROUGHPUT)))
> goto default_throughput;
Please fix the alignment:
* Found wrong alignment at net/batman-adv/bat_v_elp.c:118, was 27 but expected 23
Kind regards,
Sven
[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply
* [PATCH] net: macb: Fix ptp time adjustment for large negative delta
From: Harini Katakam @ 2018-06-20 11:34 UTC (permalink / raw)
To: nicolas.ferre, davem
Cc: netdev, linux-kernel, michal.simek, harinikatakamlinux,
harini.katakam
When delta passed to gem_ptp_adjtime is negative, the sign is
maintained in the ns_to_timespec64 conversion. Hence timespec_add
should be used directly. timespec_sub will just subtract the negative
value thus increasing the time difference.
Signed-off-by: Harini Katakam <harini.katakam@xilinx.com>
---
drivers/net/ethernet/cadence/macb_ptp.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/cadence/macb_ptp.c b/drivers/net/ethernet/cadence/macb_ptp.c
index 2220c77..6788351 100644
--- a/drivers/net/ethernet/cadence/macb_ptp.c
+++ b/drivers/net/ethernet/cadence/macb_ptp.c
@@ -170,10 +170,7 @@ static int gem_ptp_adjtime(struct ptp_clock_info *ptp, s64 delta)
if (delta > TSU_NSEC_MAX_VAL) {
gem_tsu_get_time(&bp->ptp_clock_info, &now);
- if (sign)
- now = timespec64_sub(now, then);
- else
- now = timespec64_add(now, then);
+ now = timespec64_add(now, then);
gem_tsu_set_time(&bp->ptp_clock_info,
(const struct timespec64 *)&now);
--
2.7.4
^ permalink raw reply related
* Re: [PATCH v4 net-next] net:sched: add action inheritdsfield to skbedit
From: Jamal Hadi Salim @ 2018-06-20 11:53 UTC (permalink / raw)
To: Michel Machado, Fu, Qiaobin, davem@davemloft.net
Cc: netdev@vger.kernel.org, Marcelo Ricardo Leitner,
xiyou.wangcong@gmail.com
In-Reply-To: <785f294c-080a-bfb4-5c7f-3ab48056649f@digirati.com.br>
On 19/06/18 08:39 AM, Michel Machado wrote:
> Notice that, different from skbmod, there's no field parm->flags in
> skbedit. Skbedit infers the flags in d->flags from the presence of the
> parameters of each of its actions. But SKBEDIT_F_INHERITDSFIELD has no
> parameter and adding field parm->flags breaks backward compatibility
> with user space as pointed out by Marcelo Ricardo Leitner. Our solution
> was to add TCA_SKBEDIT_FLAGS, so SKBEDIT_F_INHERITDSFIELD and future
> flag-only actions can be added.
Ok, that makes sense - thanks. I am not so sure about using
64 bits (32 bits would have been fine to match the size of
the kernel flags), but other than that LGTM.
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
cheers,
jamal
^ permalink raw reply
* [PATCH] r8169: Fix netpoll oops
From: Ville Syrjala @ 2018-06-20 12:01 UTC (permalink / raw)
To: netdev
Cc: Realtek linux nic maintainers, Heiner Kallweit, David S . Miller,
Ville Syrjälä
From: Ville Syrjälä <ville.syrjala@linux.intel.com>
Pass the correct thing to rtl8169_interrupt() from netpoll.
Cc: Realtek linux nic maintainers <nic_swsd@realtek.com>
Cc: netdev@vger.kernel.org
Cc: Heiner Kallweit <hkallweit1@gmail.com>
Cc: David S. Miller <davem@davemloft.net>
Fixes: ebcd5daa7ffd ("r8169: change interrupt handler argument type")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
drivers/net/ethernet/realtek/r8169.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 75dfac0248f4..f4cae2be0fda 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -7148,7 +7148,7 @@ static void rtl8169_netpoll(struct net_device *dev)
{
struct rtl8169_private *tp = netdev_priv(dev);
- rtl8169_interrupt(pci_irq_vector(tp->pci_dev, 0), dev);
+ rtl8169_interrupt(pci_irq_vector(tp->pci_dev, 0), tp);
}
#endif
--
2.16.4
^ permalink raw reply related
* Re: [PATCH net 1/5] net sched actions: fix coding style in pedit action
From: Davide Caratti @ 2018-06-20 12:29 UTC (permalink / raw)
To: Roman Mashak, davem; +Cc: netdev, kernel, jhs, xiyou.wangcong, jiri
In-Reply-To: <1529427368-17129-2-git-send-email-mrv@mojatatu.com>
On Tue, 2018-06-19 at 12:56 -0400, Roman Mashak wrote:
> Fix coding style issues in tc pedit action detected by the
> checkpatch script.
>
> Signed-off-by: Roman Mashak <mrv@mojatatu.com>
...
> ---
> @@ -316,16 +318,15 @@ static int tcf_pedit(struct sk_buff *skb, const struct tc_action *a,
> hoffset + tkey->at);
> goto bad;
> }
> - d = skb_header_pointer(skb, hoffset + tkey->at, 1,
> - &_d);
> + d = skb_header_pointer(skb, hoffset + tkey->at,
> + 1, &_d);
> if (!d)
> goto bad;
> offset += (*d & tkey->offmask) >> tkey->shift;
> }
hello Roman,
nit: while we are here, what about changing the declaration of _d and *d
to u8, so that the bitwise operation is done on unsigned?
BTW: the patch (and the series) looks ok, but I guess it will better
target net-next when the branch reopens
thanks!
--
davide
^ permalink raw reply
* Re: [PATCH] ceph: use ktime_get_real_seconds()
From: Ilya Dryomov @ 2018-06-20 12:33 UTC (permalink / raw)
To: allen.lkml; +Cc: netdev, David S. Miller, linux-kernel, Ceph Development, y2038
In-Reply-To: <1529488801-22093-1-git-send-email-allen.lkml@gmail.com>
On Wed, Jun 20, 2018 at 12:00 PM Allen Pais <allen.lkml@gmail.com> wrote:
>
> Use ktime_get_real_seconds() as get_seconds() is deprecated.
>
> Signed-off-by: Allen Pais <allen.lkml@gmail.com>
> ---
> net/ceph/auth_x.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/net/ceph/auth_x.c b/net/ceph/auth_x.c
> index 2f4a1ba..99561c2 100644
> --- a/net/ceph/auth_x.c
> +++ b/net/ceph/auth_x.c
> @@ -154,7 +154,7 @@ static int process_one_ticket(struct ceph_auth_client *ac,
> void **ptp;
> struct ceph_crypto_key new_session_key = { 0 };
> struct ceph_buffer *new_ticket_blob;
> - unsigned long new_expires, new_renew_after;
> + u32 new_expires, new_renew_after;
> u64 new_secret_id;
> int ret;
>
> @@ -191,9 +191,9 @@ static int process_one_ticket(struct ceph_auth_client *ac,
>
> ceph_decode_timespec(&validity, dp);
> dp += sizeof(struct ceph_timespec);
> - new_expires = get_seconds() + validity.tv_sec;
> + new_expires = (u32)ktime_get_real_seconds() + validity.tv_sec;
Why the change to u32 and this cast? If the type has to change,
wouldn't time64_t make more sense? ktime_get_real_seconds() returns
time64_t, after all.
Thanks,
Ilya
^ permalink raw reply
* [GIT] Networking
From: David Miller @ 2018-06-20 12:35 UTC (permalink / raw)
To: torvalds; +Cc: akpm, netdev, linux-kernel
1) Fix crash on bpf_prog_load() errors, from Daniel Borkmann.
2) Fix ATM VCC memory accounting, from David Woodhouse.
3) fib6_info objects need RCU freeing, from Eric Dumazet.
4) Fix SO_BINDTODEVICE handling for TCP sockets, from David Ahern.
5) Fix clobbered error code in enic_open() failure path, from
Govindarajulu Varadarajan.
6) Propagate dev_get_valid_name() error returns properly, from
Li RongQing.
7) Fix suspend/resume in davinci_emac driver, from Bartosz
Golaszewski.
8) Various act_ife fixes (recursive locking, IDR leaks, etc.)
from Davide Caratti.
9) Fix buggy checksum handling in sungem driver, from Eric
Dumazet.
Please pull, thanks a lot!
The following changes since commit 35773c93817c5f2df264d013978e7551056a063a:
Merge branch 'afs-proc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs (2018-06-16 16:32:04 +0900)
are available in the Git repository at:
gitolite@ra.kernel.org:/pub/scm/linux/kernel/git/davem/net.git
for you to fetch changes up to 9887cba19978a5f288100ef90a37684cc8d5e0a6:
ip: limit use of gso_size to udp (2018-06-20 14:41:04 +0900)
----------------------------------------------------------------
Alexei Starovoitov (1):
Merge branch 'bpf-fixes'
Anders Roxell (1):
selftests: bpf: config: add config fragments
Arnd Bergmann (1):
ptp: replace getnstimeofday64() with ktime_get_real_ts64()
Bartosz Golaszewski (1):
net: ethernet: fix suspend/resume in davinci_emac
Bhadram Varka (1):
stmmac: fix DMA channel hang in half-duplex mode
Björn Töpel (1):
xsk: re-add queue id check for XDP_SKB path
Daniel Borkmann (4):
Merge branch 'bpf-misc-fixes'
bpf: fix panic in prog load calls cleanup
bpf: reject any prog that failed read-only lock
bpf, xdp, i40e: fix i40e_build_skb skb reserve and truesize
Daniel Lezcano (1):
net/usb/drivers: Remove useless hrtimer_active check
David Ahern (1):
net/tcp: Fix socket lookups with SO_BINDTODEVICE
David S. Miller (4):
Merge git://git.kernel.org/.../bpf/bpf
bluetooth: hci_nokia: Don't include linux/unaligned/le_struct.h directly.
Merge branch 'qed-fixes'
Merge branch 'NCSI-silence-warning-messages'
David Woodhouse (1):
atm: Preserve value of skb->truesize when accounting to vcc
Davide Caratti (2):
net/sched: act_ife: fix recursive lock and idr leak
net/sched: act_ife: preserve the action control in case of error
Dinh Nguyen (1):
net: stmmac: socfpga: add additional ocp reset line for Stratix10
Eric Dumazet (2):
net/ipv6: respect rcu grace period before freeing fib6_info
net: sungem: fix rx checksum support
Govindarajulu Varadarajan (2):
enic: initialize enic->rfs_h.lock in enic_probe
enic: do not overwrite error code
Jakub Kicinski (2):
tools: bpftool: improve accuracy of load time
selftests/bpf: test offloads even with BPF programs present
Jian Wang (1):
bpf, selftest: check tunnel type more accurately
Joel Stanley (4):
net/ncsi: Silence debug messages
net/ncsi: Drop no more channels message
net/ncsi: Use netdev_dbg for debug messages
MAINTAINERS: Add Sam as the maintainer for NCSI
Konstantin Khlebnikov (1):
net_sched: blackhole: tell upper qdisc about dropped packets
Li RongQing (1):
net: propagate dev_get_valid_name return code
Liran Alon (1):
net: net_failover: fix typo in net_failover_slave_register()
Matteo Croce (2):
bpfilter: fix build error
bpfilter: ignore binary files
Stefan Agner (1):
net: hamradio: use eth_broadcast_addr
Sudarsana Reddy Kalluru (3):
qed: Fix possible memory leak in Rx error path handling.
qed: Add sanity check for SIMD fastpath handler.
qed: Do not advertise DCBX_LLD_MANAGED capability.
Toshiaki Makita (1):
xdp: Fix handling of devmap in generic XDP
Willem de Bruijn (1):
ip: limit use of gso_size to udp
William Tu (1):
bpf, selftests: delete xfrm tunnel when test exits.
Xin Long (1):
ipvlan: use ETH_MAX_MTU as max mtu
Yonghong Song (1):
tools/bpftool: fix a bug in bpftool perf
MAINTAINERS | 5 +++++
drivers/bluetooth/hci_nokia.c | 2 +-
drivers/net/ethernet/cisco/enic/enic_clsf.c | 3 +--
drivers/net/ethernet/cisco/enic/enic_main.c | 12 ++++++------
drivers/net/ethernet/faraday/ftgmac100.c | 4 ++--
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 7 +++----
drivers/net/ethernet/qlogic/qed/qed_dcbx.c | 11 ++++-------
drivers/net/ethernet/qlogic/qed/qed_ll2.c | 11 +++++++++--
drivers/net/ethernet/qlogic/qed/qed_main.c | 12 ++++++++++--
drivers/net/ethernet/stmicro/stmmac/Kconfig | 2 +-
drivers/net/ethernet/stmicro/stmmac/dwmac-socfpga.c | 18 ++++++++++++++----
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 10 ++++++++++
drivers/net/ethernet/sun/sungem.c | 22 ++++++++++++----------
drivers/net/ethernet/ti/davinci_emac.c | 15 +++++++++++++--
drivers/net/hamradio/bpqether.c | 8 ++------
drivers/net/ipvlan/ipvlan_main.c | 1 +
drivers/net/net_failover.c | 2 +-
drivers/net/usb/cdc_ncm.c | 3 +--
drivers/ptp/ptp_chardev.c | 4 ++--
drivers/ptp/ptp_qoriq.c | 2 +-
include/linux/atmdev.h | 15 +++++++++++++++
include/linux/bpf.h | 12 ++++++++++++
include/linux/filter.h | 79 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++----------------------
include/net/ip6_fib.h | 5 +++--
kernel/bpf/core.c | 69 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-------
kernel/bpf/devmap.c | 14 ++++++++++++++
kernel/bpf/syscall.c | 12 +++---------
net/atm/br2684.c | 3 +--
net/atm/clip.c | 3 +--
net/atm/common.c | 3 +--
net/atm/lec.c | 3 +--
net/atm/mpc.c | 3 +--
net/atm/pppoatm.c | 3 +--
net/atm/raw.c | 4 ++--
net/bpfilter/.gitignore | 1 +
net/bpfilter/Makefile | 6 ++++--
net/core/dev.c | 4 ++--
net/core/filter.c | 21 ++++-----------------
net/ipv4/inet_hashtables.c | 4 ++--
net/ipv4/ip_output.c | 3 ++-
net/ipv6/inet6_hashtables.c | 4 ++--
net/ipv6/ip6_fib.c | 5 +++--
net/ipv6/ip6_output.c | 3 ++-
net/ncsi/ncsi-aen.c | 10 +++++-----
net/ncsi/ncsi-manage.c | 49 ++++++++++++++++++++++---------------------------
net/sched/act_ife.c | 12 +++++-------
net/sched/sch_blackhole.c | 2 +-
net/xdp/xsk.c | 3 +++
tools/bpf/bpftool/perf.c | 5 +++--
tools/bpf/bpftool/prog.c | 4 +++-
tools/testing/selftests/bpf/config | 10 ++++++++++
tools/testing/selftests/bpf/test_offload.py | 12 ++++++++++--
tools/testing/selftests/bpf/test_tunnel.sh | 26 ++++++++++++++------------
53 files changed, 371 insertions(+), 195 deletions(-)
create mode 100644 net/bpfilter/.gitignore
^ permalink raw reply
* Re: [PATCH] bpfilter: fix build error
From: Stefano Brivio @ 2018-06-20 12:39 UTC (permalink / raw)
To: Matteo Croce; +Cc: netdev, Alexei Starovoitov
In-Reply-To: <20180619151620.1912-1-mcroce@redhat.com>
On Tue, 19 Jun 2018 17:16:20 +0200
Matteo Croce <mcroce@redhat.com> wrote:
> bpfilter Makefile assumes that the system locale is en_US, and the
> parsing of objdump output fails.
> Set LC_ALL=C and, while at it, rewrite the objdump parsing so it spawns
> only 2 processes instead of 7.
>
> Fixes: d2ba09c17a064 ("net: add skeleton of bpfilter kernel module")
> Signed-off-by: Matteo Croce <mcroce@redhat.com>
> ---
> net/bpfilter/Makefile | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/net/bpfilter/Makefile b/net/bpfilter/Makefile
> index e0bbe7583e58..dd86b022eff0 100644
> --- a/net/bpfilter/Makefile
> +++ b/net/bpfilter/Makefile
> @@ -21,8 +21,10 @@ endif
> # which bpfilter_kern.c passes further into umh blob loader at run-time
> quiet_cmd_copy_umh = GEN $@
> cmd_copy_umh = echo ':' > $(obj)/.bpfilter_umh.o.cmd; \
> - $(OBJCOPY) -I binary -O `$(OBJDUMP) -f $<|grep format|cut -d' ' -f8` \
> - -B `$(OBJDUMP) -f $<|grep architecture|cut -d, -f1|cut -d' ' -f2` \
> + $(OBJCOPY) -I binary \
> + `LC_ALL=C objdump -f net/bpfilter/bpfilter_umh \
Why do you use objdump instead of $(OBJDUMP) now? I guess this might
cause issues if you're cross-compiling.
--
Stefano
^ permalink raw reply
* Re: [PATCH v4 net-next] net:sched: add action inheritdsfield to skbedit
From: Michel Machado @ 2018-06-20 12:42 UTC (permalink / raw)
To: Jamal Hadi Salim, Fu, Qiaobin, davem@davemloft.net
Cc: netdev@vger.kernel.org, Marcelo Ricardo Leitner,
xiyou.wangcong@gmail.com
In-Reply-To: <e3b8ffa4-bad5-0e35-39e8-84836db7972e@mojatatu.com>
On 06/20/2018 07:53 AM, Jamal Hadi Salim wrote:
> On 19/06/18 08:39 AM, Michel Machado wrote:
>
>> Notice that, different from skbmod, there's no field parm->flags in
>> skbedit. Skbedit infers the flags in d->flags from the presence of the
>> parameters of each of its actions. But SKBEDIT_F_INHERITDSFIELD has no
>> parameter and adding field parm->flags breaks backward compatibility
>> with user space as pointed out by Marcelo Ricardo Leitner. Our
>> solution was to add TCA_SKBEDIT_FLAGS, so SKBEDIT_F_INHERITDSFIELD and
>> future flag-only actions can be added.
>
> Ok, that makes sense - thanks. I am not so sure about using
> 64 bits (32 bits would have been fine to match the size of
> the kernel flags), but other than that LGTM.
The choice for the u64 is meant to keep the interface between kernel
and user space the same for hopefully a longer time than it would be
with a u32. Changing from u32 to u64 in the kernel, when the need
arrives, won't impact applications. This interface choice was motivated
by the backward compatibility issue mentioned above.
Thank you for the review, Jamal.
[ ]'s
Michel Machado
^ permalink raw reply
* Re: [PATCH net] net: sungem: fix rx checksum support
From: Eric Dumazet @ 2018-06-20 12:49 UTC (permalink / raw)
To: Mathieu Malaterre, David S. Miller
Cc: Eric Dumazet, netdev, Meelis Roos, schwab, eric.dumazet
In-Reply-To: <CA+7wUswWV7FJ_n0Dd-p_2aK7m3xcUL0CJsokhCZH0J+scvJO3g@mail.gmail.com>
On 06/20/2018 04:22 AM, Mathieu Malaterre wrote:
> Thanks for the stable tag.
>
> IMHO the commit message should have also reference commit 7ce5a27f2ef8
> ("Revert "net: Handle CHECKSUM_COMPLETE more adequately in
> pskb_trim_rcsum().""). Which means that commit 88078d98d1bb ("net:
> pskb_trim_rcsum() and CHECKSUM_COMPLETE are friends") should be ready
> for stable also, since it seems that the only driver responsible for
> the revert was sungem.
>
> my 2cts
>
I disagree.
commit 88078d98d1bb is a performance improvement, and not needed for stable.
We might also discover another buggy driver later.
Thanks.
^ permalink raw reply
* Re: [RFC v2, net-next, PATCH 4/4] net/cpsw_switchdev: add switchdev mode of operation on cpsw driver
From: Ivan Vecera @ 2018-06-20 12:53 UTC (permalink / raw)
To: Jiri Pirko, Grygorii Strashko
Cc: Ilias Apalodimas, netdev, ivan.khoronzhuk, nsekhar, andrew,
f.fainelli, francois.ozog, yogeshs, spatton, Jose.Abreu
In-Reply-To: <20180620070808.GA2119@nanopsycho>
On 20.6.2018 09:08, Jiri Pirko wrote:
> Tue, Jun 19, 2018 at 01:19:00AM CEST, grygorii.strashko@ti.com wrote:
>>
>>
>> On 06/14/2018 06:43 AM, Ilias Apalodimas wrote:
>>> On Thu, Jun 14, 2018 at 01:39:58PM +0200, Jiri Pirko wrote:
>>>> Thu, Jun 14, 2018 at 01:34:04PM CEST, ilias.apalodimas@linaro.org wrote:
>>>>> On Thu, Jun 14, 2018 at 01:30:28PM +0200, Jiri Pirko wrote:
>>>>>> Thu, Jun 14, 2018 at 01:11:30PM CEST, ilias.apalodimas@linaro.org wrote:
>>>>>>
>>>>>> [...]
>>>>>>
>>>>>>> @@ -2711,6 +2789,10 @@ static int cpsw_probe_dt(struct cpsw_platform_data *data,
>>>>>>> if (of_property_read_bool(node, "dual_emac"))
>>>>>>> data->switch_mode = CPSW_DUAL_EMAC;
>>>>>>>
>>>>>>> + /* switchdev overrides DTS */
>>>>>>> + if (IS_ENABLED(CONFIG_TI_CPSW_SWITCHDEV))
>>>>>>> + data->switch_mode = CPSW_SWITCHDEV;
>>>>>>
>>>>>> So you force CPSW_SWITCHDEV mode if the CONFIG_TI_CPSW_SWITCHDEV is
>>>>>> enabled. That does not sound right. I think that user should tell what
>>>>>> mode does he want regardless what the kernel config is.
>>>>> We discussed this during the V1 of the RFC. Yes it doesn't seem good, but the
>>>>> device currently configures the modes using DTS (which is not correct). I choose
>>>>> the .config due to that. I can't think of anything better, but i am open to
>>>>> suggestions
>>>>
>>>> Agreed that DTS does fit as well. I think that this might be a job for
>>>> devlink parameters (patchset is going to be sent upstream next week).
>>>> You do have 1 bus address for the whole device (both ports), right?
>>>>
>>> Yes devlink sounds reasonable. I thyink there's only one bus for it, but then
>>> again i am far from an expert on the hardware interrnals. Grygorii can correct
>>> me if i am wrong.
>>
>> Devlink and NFS boot are not compatible as per my understanding, so ..
>
> ? Why do you say so?
Why aren't they compatible?? You can have devlink binary in initramfs and
configure the ASIC and ports via devlink batch file - prior mounting NFS root.
>>
>> Again, current driver, as is, supports NFS boot in all modes
>> (how good is the cur driver is question which out of scope of this discussion).
>>
>> And we discussed this in RFC v1 - driver mode selection will be changed
>> if we will proceed and it will be new driver for proper switch support.
>>
>> Not sure about about Devlink (need to study it and we never got any requests from end
>> users for this as I know), and I'd like to note (again) that this is embedded
>> (industrial/automotive etc), so everything preferred to be simple, fast and
>> preferably configured statically (in most of the cases end users what boot time
>> configuration).
>
> You need to study it indeed.
>
>>
>> --
>> regards,
>> -grygorii
^ permalink raw reply
* Re: [RFC v2, net-next, PATCH 4/4] net/cpsw_switchdev: add switchdev mode of operation on cpsw driver
From: Ivan Vecera @ 2018-06-20 12:56 UTC (permalink / raw)
To: Ilias Apalodimas, Andrew Lunn
Cc: netdev, grygorii.strashko, ivan.khoronzhuk, nsekhar, jiri,
f.fainelli, francois.ozog, yogeshs, spatton, Jose.Abreu
In-Reply-To: <20180618201940.GA5890@apalos>
On 18.6.2018 22:19, Ilias Apalodimas wrote:
> Jiri proposed using devlink, which makes sense, but i am not sure it's
> applicable on this patchset. This will change the driver completely and will
> totally break backwards compatibility.
Another good reason for a new driver.
I.
^ permalink raw reply
* Re: [RFC v2, net-next, PATCH 4/4] net/cpsw_switchdev: add switchdev mode of operation on cpsw driver
From: Ilias Apalodimas @ 2018-06-20 12:59 UTC (permalink / raw)
To: Ivan Vecera
Cc: Jiri Pirko, Grygorii Strashko, netdev, ivan.khoronzhuk, nsekhar,
andrew, f.fainelli, francois.ozog, yogeshs, spatton, Jose.Abreu
In-Reply-To: <77627ca3-3e24-ea0b-7ba6-55f1f2a1114e@redhat.com>
On Wed, Jun 20, 2018 at 02:53:47PM +0200, Ivan Vecera wrote:
> On 20.6.2018 09:08, Jiri Pirko wrote:
> > Tue, Jun 19, 2018 at 01:19:00AM CEST, grygorii.strashko@ti.com wrote:
> >>
> >>
> >> On 06/14/2018 06:43 AM, Ilias Apalodimas wrote:
> >>> On Thu, Jun 14, 2018 at 01:39:58PM +0200, Jiri Pirko wrote:
> >>>> Thu, Jun 14, 2018 at 01:34:04PM CEST, ilias.apalodimas@linaro.org wrote:
> >>>>> On Thu, Jun 14, 2018 at 01:30:28PM +0200, Jiri Pirko wrote:
> >>>>>> Thu, Jun 14, 2018 at 01:11:30PM CEST, ilias.apalodimas@linaro.org wrote:
> >>>>>>
> >>>>>> [...]
> >>>>>>
> >>>>>>> @@ -2711,6 +2789,10 @@ static int cpsw_probe_dt(struct cpsw_platform_data *data,
> >>>>>>> if (of_property_read_bool(node, "dual_emac"))
> >>>>>>> data->switch_mode = CPSW_DUAL_EMAC;
> >>>>>>>
> >>>>>>> + /* switchdev overrides DTS */
> >>>>>>> + if (IS_ENABLED(CONFIG_TI_CPSW_SWITCHDEV))
> >>>>>>> + data->switch_mode = CPSW_SWITCHDEV;
> >>>>>>
> >>>>>> So you force CPSW_SWITCHDEV mode if the CONFIG_TI_CPSW_SWITCHDEV is
> >>>>>> enabled. That does not sound right. I think that user should tell what
> >>>>>> mode does he want regardless what the kernel config is.
> >>>>> We discussed this during the V1 of the RFC. Yes it doesn't seem good, but the
> >>>>> device currently configures the modes using DTS (which is not correct). I choose
> >>>>> the .config due to that. I can't think of anything better, but i am open to
> >>>>> suggestions
> >>>>
> >>>> Agreed that DTS does fit as well. I think that this might be a job for
> >>>> devlink parameters (patchset is going to be sent upstream next week).
> >>>> You do have 1 bus address for the whole device (both ports), right?
> >>>>
> >>> Yes devlink sounds reasonable. I thyink there's only one bus for it, but then
> >>> again i am far from an expert on the hardware interrnals. Grygorii can correct
> >>> me if i am wrong.
> >>
> >> Devlink and NFS boot are not compatible as per my understanding, so ..
> >
> > ? Why do you say so?
>
> Why aren't they compatible?? You can have devlink binary in initramfs and
> configure the ASIC and ports via devlink batch file - prior mounting NFS root.
This is doable but, are we taking into consideration device reconfiguration
(which was the reason for creating the minimal filesystem) or are we just
talking about device booting/initial config?
>
> >>
> >> Again, current driver, as is, supports NFS boot in all modes
> >> (how good is the cur driver is question which out of scope of this discussion).
> >>
> >> And we discussed this in RFC v1 - driver mode selection will be changed
> >> if we will proceed and it will be new driver for proper switch support.
> >>
> >> Not sure about about Devlink (need to study it and we never got any requests from end
> >> users for this as I know), and I'd like to note (again) that this is embedded
> >> (industrial/automotive etc), so everything preferred to be simple, fast and
> >> preferably configured statically (in most of the cases end users what boot time
> >> configuration).
> >
> > You need to study it indeed.
> >
> >>
> >> --
> >> regards,
> >> -grygorii
>
Regards
Ilias
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox