| Message ID | 20240918193452.417115-1-shenwei.wang@nxp.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [v2,net] net: stmmac: dwmac4: extend timeout for VLAN Tag register busy bit check | expand |
On Wed, Sep 18, 2024 at 02:34:52PM -0500, Shenwei Wang wrote: > Increase the timeout for checking the busy bit of the VLAN Tag register > from 10µs to 500ms. This change is necessary to accommodate scenarios > where Energy Efficient Ethernet (EEE) is enabled. > > Overnight testing revealed that when EEE is active, the busy bit can > remain set for up to approximately 300ms. The new 500ms timeout provides > a safety margin. Do you know what EEE has to do with VLAN filtering? Could there be other registers which suffer from the same problem? Andrew
> -----Original Message----- > From: Andrew Lunn <andrew@lunn.ch> > Sent: Wednesday, September 18, 2024 6:58 PM > To: Shenwei Wang <shenwei.wang@nxp.com> > Cc: David S. Miller <davem@davemloft.net>; Eric Dumazet > <edumazet@google.com>; Jakub Kicinski <kuba@kernel.org>; Paolo Abeni > <pabeni@redhat.com>; Maxime Coquelin <mcoquelin.stm32@gmail.com>; > horms@kernel.org; Alexandre Torgue <alexandre.torgue@foss.st.com>; Jose > Abreu <joabreu@synopsys.com>; Ong Boon Leong <boon.leong.ong@intel.com>; > Wong Vee Khee <vee.khee.wong@intel.com>; Chuah Kim Tatt > <kim.tatt.chuah@intel.com>; netdev@vger.kernel.org; linux-stm32@st-md- > mailman.stormreply.com; linux-arm-kernel@lists.infradead.org; > imx@lists.linux.dev; dl-linux-imx <linux-imx@nxp.com> > Subject: [EXT] Re: [PATCH v2 net] net: stmmac: dwmac4: extend timeout for > VLAN Tag register busy bit check > > Overnight testing revealed that when EEE is active, the busy bit can > > remain set for up to approximately 300ms. The new 500ms timeout > > provides a safety margin. > > Do you know what EEE has to do with VLAN filtering? > The exact design details are not available to me, but my understanding is that the Busy Bit is synchronized to the RX clock supplied by the PHY. When EEE is active and the PHY enters LPI state, the RX clock is gated, preventing updates to the Busy Bit. The question is the significant delay observed, especially considering that the PHY transitions between active and LPI states multiple times during this period. There should have a lot of chances to update the Busy Bit sooner when it is in the active state. > Could there be other registers which suffer from the same problem? > So far I think it only impact the VLAN status register because those bits are driven by another clock instead of CSR clock. Based on current observations, it appears that this issue primarily affects the VLAN status register. The reason for this is that the bits in the VLAN status register are driven by a clock source distinct from the CSR clock. Regards, Shenwei > Andrew
> > Could there be other registers which suffer from the same problem? > > > > So far I think it only impact the VLAN status register because those bits are driven by another clock instead of CSR clock. > Based on current observations, it appears that this issue primarily affects the VLAN status register. The reason for this > is that the bits in the VLAN status register are driven by a clock source distinct from the CSR clock. Thanks for the explanation. Reviewed-by: Andrew Lunn <andrew@lunn.ch> Andrew
On Wed, Sep 18, 2024 at 02:34:52PM -0500, Shenwei Wang wrote: > Increase the timeout for checking the busy bit of the VLAN Tag register > from 10µs to 500ms. This change is necessary to accommodate scenarios > where Energy Efficient Ethernet (EEE) is enabled. > > Overnight testing revealed that when EEE is active, the busy bit can > remain set for up to approximately 300ms. The new 500ms timeout provides > a safety margin. > > Fixes: ed64639bc1e0 ("net: stmmac: Add support for VLAN Rx filtering") > Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com> > --- > Changes in v2: > - replace the udelay with readl_poll_timeout per Simon's review. > > --- > drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c | 14 +++++++------- > 1 file changed, 7 insertions(+), 7 deletions(-) > > diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c b/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c > index a1858f083eef..a0cfa2eaebb4 100644 > --- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c > +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c > @@ -14,6 +14,7 @@ > #include <linux/slab.h> > #include <linux/ethtool.h> > #include <linux/io.h> > +#include <linux/iopoll.h> > #include "stmmac.h" > #include "stmmac_pcs.h" > #include "dwmac4.h" > @@ -471,7 +472,7 @@ static int dwmac4_write_vlan_filter(struct net_device *dev, > u8 index, u32 data) > { > void __iomem *ioaddr = (void __iomem *)dev->base_addr; > - int i, timeout = 10; > + int ret, timeout = 500000; //500ms > u32 val; > > if (index >= hw->num_vlan) > @@ -487,12 +488,11 @@ static int dwmac4_write_vlan_filter(struct net_device *dev, > > writel(val, ioaddr + GMAC_VLAN_TAG); > > - for (i = 0; i < timeout; i++) { > - val = readl(ioaddr + GMAC_VLAN_TAG); > - if (!(val & GMAC_VLAN_TAG_CTRL_OB)) > - return 0; > - udelay(1); > - } > + ret = readl_poll_timeout(ioaddr + GMAC_VLAN_TAG, val, > + !(val & GMAC_VLAN_TAG_CTRL_OB), > + 1000, timeout); > + if (!ret) > + return 0; > > netdev_err(dev, "Timeout accessing MAC_VLAN_Tag_Filter\n"); 1. Just drop the timeout local variable and use the 500000 literal directly. There is no point in such parametrization especially you have already added the delay as is. 2. A more traditional, readable and maintainable pattern is the error-check statement after the call. So seeing you are changing this part of the method anyway, let's convert it to: + ret = readl_poll_timeout(ioaddr + GMAC_VLAN_TAG, val, + !(val & GMAC_VLAN_TAG_CTRL_OB), + 1000, timeout); + if (ret) { + netdev_err(dev, "Timeout accessing MAC_VLAN_Tag_Filter\n"); + return ret; + } + + return 0; -Serge(y) > > -- > 2.34.1 > >
> -----Original Message----- > From: Serge Semin <fancer.lancer@gmail.com> > Sent: Sunday, September 22, 2024 5:18 PM > To: Shenwei Wang <shenwei.wang@nxp.com> > Cc: David S. Miller <davem@davemloft.net>; Eric Dumazet > <edumazet@google.com>; Jakub Kicinski <kuba@kernel.org>; Paolo Abeni > <pabeni@redhat.com>; Maxime Coquelin <mcoquelin.stm32@gmail.com>; > horms@kernel.org; Alexandre Torgue <alexandre.torgue@foss.st.com>; Jose > Abreu <joabreu@synopsys.com>; Ong Boon Leong <boon.leong.ong@intel.com>; > Wong Vee Khee <vee.khee.wong@intel.com>; Chuah Kim Tatt > <kim.tatt.chuah@intel.com>; netdev@vger.kernel.org; linux-stm32@st-md- > mailman.stormreply.com; linux-arm-kernel@lists.infradead.org; > imx@lists.linux.dev; dl-linux-imx <linux-imx@nxp.com> > Subject: [EXT] Re: [PATCH v2 net] net: stmmac: dwmac4: extend timeout for > VLAN Tag register busy bit check > > > > - for (i = 0; i < timeout; i++) { > > - val = readl(ioaddr + GMAC_VLAN_TAG); > > - if (!(val & GMAC_VLAN_TAG_CTRL_OB)) > > - return 0; > > - udelay(1); > > - } > > > + ret = readl_poll_timeout(ioaddr + GMAC_VLAN_TAG, val, > > + !(val & GMAC_VLAN_TAG_CTRL_OB), > > + 1000, timeout); > > + if (!ret) > > + return 0; > > > > netdev_err(dev, "Timeout accessing MAC_VLAN_Tag_Filter\n"); > > 1. Just drop the timeout local variable and use the 500000 literal directly. There is > no point in such parametrization especially you have already added the delay as > is. > > 2. A more traditional, readable and maintainable pattern is the error-check > statement after the call. So seeing you are changing this part of the method > anyway, let's convert it to: > > + ret = readl_poll_timeout(ioaddr + GMAC_VLAN_TAG, val, > + !(val & GMAC_VLAN_TAG_CTRL_OB), > + 1000, timeout); > + if (ret) { > + netdev_err(dev, "Timeout accessing MAC_VLAN_Tag_Filter\n"); > + return ret; > + } > + > + return 0; > Hi Serge, Good suggestions! Will send out a new version. Thanks, Shenwei > -Serge(y) > > > > > -- > > 2.34.1 > > > >
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c b/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c index a1858f083eef..a0cfa2eaebb4 100644 --- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c @@ -14,6 +14,7 @@ #include <linux/slab.h> #include <linux/ethtool.h> #include <linux/io.h> +#include <linux/iopoll.h> #include "stmmac.h" #include "stmmac_pcs.h" #include "dwmac4.h" @@ -471,7 +472,7 @@ static int dwmac4_write_vlan_filter(struct net_device *dev, u8 index, u32 data) { void __iomem *ioaddr = (void __iomem *)dev->base_addr; - int i, timeout = 10; + int ret, timeout = 500000; //500ms u32 val; if (index >= hw->num_vlan) @@ -487,12 +488,11 @@ static int dwmac4_write_vlan_filter(struct net_device *dev, writel(val, ioaddr + GMAC_VLAN_TAG); - for (i = 0; i < timeout; i++) { - val = readl(ioaddr + GMAC_VLAN_TAG); - if (!(val & GMAC_VLAN_TAG_CTRL_OB)) - return 0; - udelay(1); - } + ret = readl_poll_timeout(ioaddr + GMAC_VLAN_TAG, val, + !(val & GMAC_VLAN_TAG_CTRL_OB), + 1000, timeout); + if (!ret) + return 0; netdev_err(dev, "Timeout accessing MAC_VLAN_Tag_Filter\n");
Increase the timeout for checking the busy bit of the VLAN Tag register from 10µs to 500ms. This change is necessary to accommodate scenarios where Energy Efficient Ethernet (EEE) is enabled. Overnight testing revealed that when EEE is active, the busy bit can remain set for up to approximately 300ms. The new 500ms timeout provides a safety margin. Fixes: ed64639bc1e0 ("net: stmmac: Add support for VLAN Rx filtering") Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com> --- Changes in v2: - replace the udelay with readl_poll_timeout per Simon's review. --- drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) -- 2.34.1