| Message ID | YC+OpFGsDPXPnXM5@mwanda (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 4e9d9d1f4880ad358a8e5eb6ac4c811fd76dd617 |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [v2,net-next] net: phy: icplus: call phy_restore_page() when phy_select_page() fails | expand |
| Context | Check | Description |
|---|---|---|
| netdev/cover_letter | success | Link |
| netdev/fixes_present | success | Link |
| netdev/patch_count | success | Link |
| netdev/tree_selection | success | Clearly marked for net-next |
| netdev/subject_prefix | success | Link |
| netdev/cc_maintainers | success | CCed 7 of 7 maintainers |
| netdev/source_inline | success | Was 0 now: 0 |
| netdev/verify_signedoff | success | Link |
| netdev/module_param | success | Was 0 now: 0 |
| netdev/build_32bit | success | Errors and warnings before: 0 this patch: 0 |
| netdev/kdoc | success | Errors and warnings before: 0 this patch: 0 |
| netdev/verify_fixes | success | Link |
| netdev/checkpatch | success | total: 0 errors, 0 warnings, 0 checks, 33 lines checked |
| netdev/build_allmodconfig_warn | success | Errors and warnings before: 0 this patch: 0 |
| netdev/header_inline | success | Link |
| netdev/stable | success | Stable not CCed |
Am 2021-02-19 11:10, schrieb Dan Carpenter: > The comments to phy_select_page() say that "phy_restore_page() must > always be called after this, irrespective of success or failure of this > call." If we don't call phy_restore_page() then we are still holding > the phy_lock_mdio_bus() so it eventually leads to a dead lock. > > Fixes: 32ab60e53920 ("net: phy: icplus: add MDI/MDIX support for > IP101A/G") > Fixes: f9bc51e6cce2 ("net: phy: icplus: fix paged register access") > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by: Michael Walle <michael@walle.cc> I assume, this has to go through "net" if the merge window is closed, no? -michael > --- > v2: fix a couple other instances I missed in v1 > > drivers/net/phy/icplus.c | 9 +++++---- > 1 file changed, 5 insertions(+), 4 deletions(-) > > diff --git a/drivers/net/phy/icplus.c b/drivers/net/phy/icplus.c > index 3e431737c1ba..a00a667454a9 100644 > --- a/drivers/net/phy/icplus.c > +++ b/drivers/net/phy/icplus.c > @@ -239,7 +239,7 @@ static int ip101a_g_config_intr_pin(struct > phy_device *phydev) > > oldpage = phy_select_page(phydev, IP101G_DEFAULT_PAGE); > if (oldpage < 0) > - return oldpage; > + goto out; > > /* configure the RXER/INTR_32 pin of the 32-pin IP101GR if needed: */ > switch (priv->sel_intr32) { > @@ -314,7 +314,7 @@ static int ip101a_g_read_status(struct phy_device > *phydev) > > oldpage = phy_select_page(phydev, IP101G_DEFAULT_PAGE); > if (oldpage < 0) > - return oldpage; > + goto out; > > ret = __phy_read(phydev, IP10XX_SPEC_CTRL_STATUS); > if (ret < 0) > @@ -349,7 +349,8 @@ static int ip101a_g_read_status(struct phy_device > *phydev) > static int ip101a_g_config_mdix(struct phy_device *phydev) > { > u16 ctrl = 0, ctrl2 = 0; > - int oldpage, ret; > + int oldpage; > + int ret = 0; > > switch (phydev->mdix_ctrl) { > case ETH_TP_MDI: > @@ -367,7 +368,7 @@ static int ip101a_g_config_mdix(struct phy_device > *phydev) > > oldpage = phy_select_page(phydev, IP101G_DEFAULT_PAGE); > if (oldpage < 0) > - return oldpage; > + goto out; > > ret = __phy_modify(phydev, IP10XX_SPEC_CTRL_STATUS, > IP101A_G_AUTO_MDIX_DIS, ctrl);
On Fri, Feb 19, 2021 at 11:46:23AM +0100, Michael Walle wrote: > Am 2021-02-19 11:10, schrieb Dan Carpenter: > > The comments to phy_select_page() say that "phy_restore_page() must > > always be called after this, irrespective of success or failure of this > > call." If we don't call phy_restore_page() then we are still holding > > the phy_lock_mdio_bus() so it eventually leads to a dead lock. > > > > Fixes: 32ab60e53920 ("net: phy: icplus: add MDI/MDIX support for > > IP101A/G") > > Fixes: f9bc51e6cce2 ("net: phy: icplus: fix paged register access") > > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> > > Reviewed-by: Michael Walle <michael@walle.cc> > > I assume, this has to go through "net" if the merge window is closed, no? It applies to net-next and not to net. I expect that all of net-next will be merged into net soon enough, but as of a couple hours ago it only applied to net-next. regards, dan carpenter
On Fri, Feb 19, 2021 at 01:10:44PM +0300, Dan Carpenter wrote: > The comments to phy_select_page() say that "phy_restore_page() must > always be called after this, irrespective of success or failure of this > call." If we don't call phy_restore_page() then we are still holding > the phy_lock_mdio_bus() so it eventually leads to a dead lock. > > Fixes: 32ab60e53920 ("net: phy: icplus: add MDI/MDIX support for IP101A/G") > Fixes: f9bc51e6cce2 ("net: phy: icplus: fix paged register access") > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by: Russell King <rmk+kernel@armlinux.org.uk> Thanks.
Hello: This patch was applied to netdev/net.git (refs/heads/master): On Fri, 19 Feb 2021 13:10:44 +0300 you wrote: > The comments to phy_select_page() say that "phy_restore_page() must > always be called after this, irrespective of success or failure of this > call." If we don't call phy_restore_page() then we are still holding > the phy_lock_mdio_bus() so it eventually leads to a dead lock. > > Fixes: 32ab60e53920 ("net: phy: icplus: add MDI/MDIX support for IP101A/G") > Fixes: f9bc51e6cce2 ("net: phy: icplus: fix paged register access") > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> > > [...] Here is the summary with links: - [v2,net-next] net: phy: icplus: call phy_restore_page() when phy_select_page() fails https://git.kernel.org/netdev/net/c/4e9d9d1f4880 You are awesome, thank you! -- Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html
diff --git a/drivers/net/phy/icplus.c b/drivers/net/phy/icplus.c index 3e431737c1ba..a00a667454a9 100644 --- a/drivers/net/phy/icplus.c +++ b/drivers/net/phy/icplus.c @@ -239,7 +239,7 @@ static int ip101a_g_config_intr_pin(struct phy_device *phydev) oldpage = phy_select_page(phydev, IP101G_DEFAULT_PAGE); if (oldpage < 0) - return oldpage; + goto out; /* configure the RXER/INTR_32 pin of the 32-pin IP101GR if needed: */ switch (priv->sel_intr32) { @@ -314,7 +314,7 @@ static int ip101a_g_read_status(struct phy_device *phydev) oldpage = phy_select_page(phydev, IP101G_DEFAULT_PAGE); if (oldpage < 0) - return oldpage; + goto out; ret = __phy_read(phydev, IP10XX_SPEC_CTRL_STATUS); if (ret < 0) @@ -349,7 +349,8 @@ static int ip101a_g_read_status(struct phy_device *phydev) static int ip101a_g_config_mdix(struct phy_device *phydev) { u16 ctrl = 0, ctrl2 = 0; - int oldpage, ret; + int oldpage; + int ret = 0; switch (phydev->mdix_ctrl) { case ETH_TP_MDI: @@ -367,7 +368,7 @@ static int ip101a_g_config_mdix(struct phy_device *phydev) oldpage = phy_select_page(phydev, IP101G_DEFAULT_PAGE); if (oldpage < 0) - return oldpage; + goto out; ret = __phy_modify(phydev, IP10XX_SPEC_CTRL_STATUS, IP101A_G_AUTO_MDIX_DIS, ctrl);
The comments to phy_select_page() say that "phy_restore_page() must always be called after this, irrespective of success or failure of this call." If we don't call phy_restore_page() then we are still holding the phy_lock_mdio_bus() so it eventually leads to a dead lock. Fixes: 32ab60e53920 ("net: phy: icplus: add MDI/MDIX support for IP101A/G") Fixes: f9bc51e6cce2 ("net: phy: icplus: fix paged register access") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> --- v2: fix a couple other instances I missed in v1 drivers/net/phy/icplus.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)