| Message ID | 20211008002225.2426-2-ansuelsmth@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | Multiple improvement for qca8337 switch | expand |
On Fri, 8 Oct 2021 02:22:11 +0200 Ansuel Smith wrote: > From Documentation phy resume triggers phy reset and restart > auto-negotiation. Add a dedicated function to wait reset to finish as > it was notice a regression where port sometime are not reliable after a > suspend/resume session. The reset wait logic is copied from phy_poll_reset. > Add dedicated suspend function to use genphy_suspend only with QCA8337 > phy and set only additional debug settings for QCA8327. With more test > it was reported that QCA8327 doesn't proprely support this mode and > using this cause the unreliability of the switch ports, especially the > malfunction of the port0. > > Fixes: 52a6cdbe43a3 ("net: phy: at803x: add resume/suspend function to qca83xx phy") Strange, checkpatch catches the wrong hash being used, but the verify_fixes script doesn't. Did you mean: Fixes: 15b9df4ece17 ("net: phy: at803x: add resume/suspend function to qca83xx phy") Or is 52a6cdbe43a3 the correct commit hash? Same question for patch 2. The fixes have to be a _separate_ series.
On Thu, Oct 07, 2021 at 07:23:04PM -0700, Jakub Kicinski wrote: > On Fri, 8 Oct 2021 02:22:11 +0200 Ansuel Smith wrote: > > From Documentation phy resume triggers phy reset and restart > > auto-negotiation. Add a dedicated function to wait reset to finish as > > it was notice a regression where port sometime are not reliable after a > > suspend/resume session. The reset wait logic is copied from phy_poll_reset. > > Add dedicated suspend function to use genphy_suspend only with QCA8337 > > phy and set only additional debug settings for QCA8327. With more test > > it was reported that QCA8327 doesn't proprely support this mode and > > using this cause the unreliability of the switch ports, especially the > > malfunction of the port0. > > > > Fixes: 52a6cdbe43a3 ("net: phy: at803x: add resume/suspend function to qca83xx phy") > > Strange, checkpatch catches the wrong hash being used, but the > verify_fixes script doesn't. Did you mean: > > Fixes: 15b9df4ece17 ("net: phy: at803x: add resume/suspend function to qca83xx phy") > > Or is 52a6cdbe43a3 the correct commit hash? Same question for patch 2. > > > The fixes have to be a _separate_ series. Hi, this series contains changes that depends on the fixes. (the 4th patch that rename the define is based on this 2 patch) How to handle that? I know it was wrong to put net and net-next patch in the same series but I don't know how to handle this strange situation. Any hint about that? About the wrong hash, yes I wrongly took the hash from my local branch.
On Fri, 8 Oct 2021 10:45:51 +0200 Ansuel Smith wrote: > On Thu, Oct 07, 2021 at 07:23:04PM -0700, Jakub Kicinski wrote: > > On Fri, 8 Oct 2021 02:22:11 +0200 Ansuel Smith wrote: > > > From Documentation phy resume triggers phy reset and restart > > > auto-negotiation. Add a dedicated function to wait reset to finish as > > > it was notice a regression where port sometime are not reliable after a > > > suspend/resume session. The reset wait logic is copied from phy_poll_reset. > > > Add dedicated suspend function to use genphy_suspend only with QCA8337 > > > phy and set only additional debug settings for QCA8327. With more test > > > it was reported that QCA8327 doesn't proprely support this mode and > > > using this cause the unreliability of the switch ports, especially the > > > malfunction of the port0. > > > > > > Fixes: 52a6cdbe43a3 ("net: phy: at803x: add resume/suspend function to qca83xx phy") > > > > Strange, checkpatch catches the wrong hash being used, but the > > verify_fixes script doesn't. Did you mean: > > > > Fixes: 15b9df4ece17 ("net: phy: at803x: add resume/suspend function to qca83xx phy") > > > > Or is 52a6cdbe43a3 the correct commit hash? Same question for patch 2. > > > > > > The fixes have to be a _separate_ series. > > this series contains changes that depends on the fixes. (the 4th patch > that rename the define is based on this 2 patch) How to handle that? > I know it was wrong to put net and net-next patch in the same series but > I don't know how to handle this strange situation. Any hint about that? If there is a functional dependency you'll need to send the net changes first and then wait until the trees are merged before sending net-next changes. Merge usually happens Thursday afternoon (pacific). You can post the net-next changes as RFC before the trees get merged to make sure they are reviewed and ready to go in. > About the wrong hash, yes I wrongly took the hash from my local branch. Indeed, looks like our checker got broken hence my confusion.
diff --git a/drivers/net/phy/at803x.c b/drivers/net/phy/at803x.c index 3feee4d59030..c6c87b82c95c 100644 --- a/drivers/net/phy/at803x.c +++ b/drivers/net/phy/at803x.c @@ -92,9 +92,14 @@ #define AT803X_DEBUG_REG_5 0x05 #define AT803X_DEBUG_TX_CLK_DLY_EN BIT(8) +#define AT803X_DEBUG_REG_HIB_CTRL 0x0b +#define AT803X_DEBUG_HIB_CTRL_SEL_RST_80U BIT(10) +#define AT803X_DEBUG_HIB_CTRL_EN_ANY_CHANGE BIT(13) + #define AT803X_DEBUG_REG_3C 0x3C #define AT803X_DEBUG_REG_3D 0x3D +#define AT803X_DEBUG_GATE_CLK_IN1000 BIT(6) #define AT803X_DEBUG_REG_1F 0x1F #define AT803X_DEBUG_PLL_ON BIT(2) @@ -1312,6 +1317,58 @@ static int qca83xx_config_init(struct phy_device *phydev) return 0; } +static int qca83xx_resume(struct phy_device *phydev) +{ + int ret, val; + + /* Skip reset if not suspended */ + if (!phydev->suspended) + return 0; + + /* Reinit the port, reset values set by suspend */ + qca83xx_config_init(phydev); + + /* Reset the port on port resume */ + phy_set_bits(phydev, MII_BMCR, BMCR_RESET | BMCR_ANENABLE); + + /* On resume from suspend the switch execute a reset and + * restart auto-negotiation. Wait for reset to complete. + */ + ret = phy_read_poll_timeout(phydev, MII_BMCR, val, !(val & BMCR_RESET), + 50000, 600000, true); + if (ret) + return ret; + + msleep(1); + + return 0; +} + +static int qca83xx_suspend(struct phy_device *phydev) +{ + u16 mask = 0; + + /* Only QCA8337 support actual suspend. + * QCA8327 cause port unreliability when phy suspend + * is set. + */ + if (phydev->drv->phy_id == QCA8337_PHY_ID) { + genphy_suspend(phydev); + } else { + mask |= ~(BMCR_SPEED1000 | BMCR_FULLDPLX); + phy_modify(phydev, MII_BMCR, mask, 0); + } + + at803x_debug_reg_mask(phydev, AT803X_DEBUG_REG_3D, + AT803X_DEBUG_GATE_CLK_IN1000, 0); + + at803x_debug_reg_mask(phydev, AT803X_DEBUG_REG_HIB_CTRL, + AT803X_DEBUG_HIB_CTRL_EN_ANY_CHANGE | + AT803X_DEBUG_HIB_CTRL_SEL_RST_80U, 0); + + return 0; +} + static struct phy_driver at803x_driver[] = { { /* Qualcomm Atheros AR8035 */ @@ -1421,8 +1478,8 @@ static struct phy_driver at803x_driver[] = { .get_sset_count = at803x_get_sset_count, .get_strings = at803x_get_strings, .get_stats = at803x_get_stats, - .suspend = genphy_suspend, - .resume = genphy_resume, + .suspend = qca83xx_suspend, + .resume = qca83xx_resume, }, { /* QCA8327-A from switch QCA8327-AL1A */ .phy_id = QCA8327_A_PHY_ID, @@ -1436,8 +1493,8 @@ static struct phy_driver at803x_driver[] = { .get_sset_count = at803x_get_sset_count, .get_strings = at803x_get_strings, .get_stats = at803x_get_stats, - .suspend = genphy_suspend, - .resume = genphy_resume, + .suspend = qca83xx_suspend, + .resume = qca83xx_resume, }, { /* QCA8327-B from switch QCA8327-BL1A */ .phy_id = QCA8327_B_PHY_ID, @@ -1451,8 +1508,8 @@ static struct phy_driver at803x_driver[] = { .get_sset_count = at803x_get_sset_count, .get_strings = at803x_get_strings, .get_stats = at803x_get_stats, - .suspend = genphy_suspend, - .resume = genphy_resume, + .suspend = qca83xx_suspend, + .resume = qca83xx_resume, }, }; module_phy_driver(at803x_driver);
From Documentation phy resume triggers phy reset and restart auto-negotiation. Add a dedicated function to wait reset to finish as it was notice a regression where port sometime are not reliable after a suspend/resume session. The reset wait logic is copied from phy_poll_reset. Add dedicated suspend function to use genphy_suspend only with QCA8337 phy and set only additional debug settings for QCA8327. With more test it was reported that QCA8327 doesn't proprely support this mode and using this cause the unreliability of the switch ports, especially the malfunction of the port0. Fixes: 52a6cdbe43a3 ("net: phy: at803x: add resume/suspend function to qca83xx phy") Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com> --- drivers/net/phy/at803x.c | 69 ++++++++++++++++++++++++++++++++++++---- 1 file changed, 63 insertions(+), 6 deletions(-)