| Message ID | 20231117130836.2.I79c8a6c8cafd89979af5407d77a6eda589833dca@changeid (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [1/2] r8152: Hold the rtnl_lock for all of reset | expand |
On Fri, Nov 17, 2023 at 1:10 PM Douglas Anderson <dianders@chromium.org> wrote: > > Previous commits added checks for RTL8152_INACCESSIBLE in the loops in > the driver. There are still a few more that keep tripping the driver > up in error cases and make things take longer than they should. Add > those in. > > Signed-off-by: Douglas Anderson <dianders@chromium.org> Reviewed-by: Grant Grundler <grundler@chromium.org> I've checked all the return paths and believe these changes don't break any of them. cheers, grant > --- > > drivers/net/usb/r8152.c | 14 ++++++++++++++ > 1 file changed, 14 insertions(+) > > diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c > index d6edf0254599..aca7dd7b4090 100644 > --- a/drivers/net/usb/r8152.c > +++ b/drivers/net/usb/r8152.c > @@ -3000,6 +3000,8 @@ static void rtl8152_nic_reset(struct r8152 *tp) > ocp_write_byte(tp, MCU_TYPE_PLA, PLA_CR, CR_RST); > > for (i = 0; i < 1000; i++) { > + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) > + return; > if (!(ocp_read_byte(tp, MCU_TYPE_PLA, PLA_CR) & CR_RST)) > break; > usleep_range(100, 400); > @@ -3329,6 +3331,8 @@ static void rtl_disable(struct r8152 *tp) > rxdy_gated_en(tp, true); > > for (i = 0; i < 1000; i++) { > + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) > + return; > ocp_data = ocp_read_byte(tp, MCU_TYPE_PLA, PLA_OOB_CTRL); > if ((ocp_data & FIFO_EMPTY) == FIFO_EMPTY) > break; > @@ -3336,6 +3340,8 @@ static void rtl_disable(struct r8152 *tp) > } > > for (i = 0; i < 1000; i++) { > + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) > + return; > if (ocp_read_word(tp, MCU_TYPE_PLA, PLA_TCR0) & TCR0_TX_EMPTY) > break; > usleep_range(1000, 2000); > @@ -5499,6 +5505,8 @@ static void wait_oob_link_list_ready(struct r8152 *tp) > int i; > > for (i = 0; i < 1000; i++) { > + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) > + return; > ocp_data = ocp_read_byte(tp, MCU_TYPE_PLA, PLA_OOB_CTRL); > if (ocp_data & LINK_LIST_READY) > break; > @@ -5513,6 +5521,8 @@ static void r8156b_wait_loading_flash(struct r8152 *tp) > int i; > > for (i = 0; i < 100; i++) { > + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) > + return; > if (ocp_read_word(tp, MCU_TYPE_USB, USB_GPHY_CTRL) & GPHY_PATCH_DONE) > break; > usleep_range(1000, 2000); > @@ -5635,6 +5645,8 @@ static int r8153_pre_firmware_1(struct r8152 *tp) > for (i = 0; i < 104; i++) { > u32 ocp_data = ocp_read_byte(tp, MCU_TYPE_USB, USB_WDT1_CTRL); > > + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) > + return -ENODEV; > if (!(ocp_data & WTD1_EN)) > break; > usleep_range(1000, 2000); > @@ -5791,6 +5803,8 @@ static void r8153_aldps_en(struct r8152 *tp, bool enable) > data &= ~EN_ALDPS; > ocp_reg_write(tp, OCP_POWER_CFG, data); > for (i = 0; i < 20; i++) { > + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) > + return; > usleep_range(1000, 2000); > if (ocp_read_word(tp, MCU_TYPE_PLA, 0xe000) & 0x0100) > break; > -- > 2.43.0.rc0.421.g78406f8d94-goog >
On Fri, 2023-11-17 at 13:08 -0800, Douglas Anderson wrote: > Previous commits added checks for RTL8152_INACCESSIBLE in the loops in > the driver. There are still a few more that keep tripping the driver > up in error cases and make things take longer than they should. Add > those in. > > Signed-off-by: Douglas Anderson <dianders@chromium.org> I think this deserves a 'Fixes' tag. Please add it. Additionally please insert the target tree in the subj prefix when re- postin (in this case 'net') You can retain the already collected reviewed-by tags. Thanks, Paolo
Hi, On Tue, Nov 21, 2023 at 2:28 AM Paolo Abeni <pabeni@redhat.com> wrote: > > On Fri, 2023-11-17 at 13:08 -0800, Douglas Anderson wrote: > > Previous commits added checks for RTL8152_INACCESSIBLE in the loops in > > the driver. There are still a few more that keep tripping the driver > > up in error cases and make things take longer than they should. Add > > those in. > > > > Signed-off-by: Douglas Anderson <dianders@chromium.org> > > I think this deserves a 'Fixes' tag. Please add it. Sure, I can add it. It didn't feel worth it to me since there's no real functional issue--just that it takes a little longer for these loops to exit out, but it shouldn't hurt. I guess that means breaking this commit into several depending on when the offending loop was added. > Additionally please insert the target tree in the subj prefix when re- > postin (in this case 'net') Funny, I just followed the tags for other commits to this file and the "net:" prefix isn't common. I guess this should be "net: usb: r8152". I can add it when I post v2.
On Tue, Nov 21, 2023 at 09:55:46AM -0800, Doug Anderson wrote: > Hi, > > On Tue, Nov 21, 2023 at 2:28 AM Paolo Abeni <pabeni@redhat.com> wrote: > > > > On Fri, 2023-11-17 at 13:08 -0800, Douglas Anderson wrote: > > > Previous commits added checks for RTL8152_INACCESSIBLE in the loops in > > > the driver. There are still a few more that keep tripping the driver > > > up in error cases and make things take longer than they should. Add > > > those in. > > > > > > Signed-off-by: Douglas Anderson <dianders@chromium.org> > > > > I think this deserves a 'Fixes' tag. Please add it. > > Sure, I can add it. It didn't feel worth it to me since there's no > real functional issue--just that it takes a little longer for these > loops to exit out, but it shouldn't hurt. I guess that means breaking > this commit into several depending on when the offending loop was > added. > > > > Additionally please insert the target tree in the subj prefix when re- > > postin (in this case 'net') > > Funny, I just followed the tags for other commits to this file and the > "net:" prefix isn't common. I guess this should be "net: usb: r8152". > I can add it when I post v2. Hi Doug, unfortunately prefix can have more than one meaning here. The target tree, often either net or net-next, should go in the [] part of the subject. In this case I think what you want is: [PATCH net n/m v2] Add RTL8152_INACCESSIBLE checks to more loops
diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c index d6edf0254599..aca7dd7b4090 100644 --- a/drivers/net/usb/r8152.c +++ b/drivers/net/usb/r8152.c @@ -3000,6 +3000,8 @@ static void rtl8152_nic_reset(struct r8152 *tp) ocp_write_byte(tp, MCU_TYPE_PLA, PLA_CR, CR_RST); for (i = 0; i < 1000; i++) { + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) + return; if (!(ocp_read_byte(tp, MCU_TYPE_PLA, PLA_CR) & CR_RST)) break; usleep_range(100, 400); @@ -3329,6 +3331,8 @@ static void rtl_disable(struct r8152 *tp) rxdy_gated_en(tp, true); for (i = 0; i < 1000; i++) { + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) + return; ocp_data = ocp_read_byte(tp, MCU_TYPE_PLA, PLA_OOB_CTRL); if ((ocp_data & FIFO_EMPTY) == FIFO_EMPTY) break; @@ -3336,6 +3340,8 @@ static void rtl_disable(struct r8152 *tp) } for (i = 0; i < 1000; i++) { + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) + return; if (ocp_read_word(tp, MCU_TYPE_PLA, PLA_TCR0) & TCR0_TX_EMPTY) break; usleep_range(1000, 2000); @@ -5499,6 +5505,8 @@ static void wait_oob_link_list_ready(struct r8152 *tp) int i; for (i = 0; i < 1000; i++) { + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) + return; ocp_data = ocp_read_byte(tp, MCU_TYPE_PLA, PLA_OOB_CTRL); if (ocp_data & LINK_LIST_READY) break; @@ -5513,6 +5521,8 @@ static void r8156b_wait_loading_flash(struct r8152 *tp) int i; for (i = 0; i < 100; i++) { + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) + return; if (ocp_read_word(tp, MCU_TYPE_USB, USB_GPHY_CTRL) & GPHY_PATCH_DONE) break; usleep_range(1000, 2000); @@ -5635,6 +5645,8 @@ static int r8153_pre_firmware_1(struct r8152 *tp) for (i = 0; i < 104; i++) { u32 ocp_data = ocp_read_byte(tp, MCU_TYPE_USB, USB_WDT1_CTRL); + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) + return -ENODEV; if (!(ocp_data & WTD1_EN)) break; usleep_range(1000, 2000); @@ -5791,6 +5803,8 @@ static void r8153_aldps_en(struct r8152 *tp, bool enable) data &= ~EN_ALDPS; ocp_reg_write(tp, OCP_POWER_CFG, data); for (i = 0; i < 20; i++) { + if (test_bit(RTL8152_INACCESSIBLE, &tp->flags)) + return; usleep_range(1000, 2000); if (ocp_read_word(tp, MCU_TYPE_PLA, 0xe000) & 0x0100) break;
Previous commits added checks for RTL8152_INACCESSIBLE in the loops in the driver. There are still a few more that keep tripping the driver up in error cases and make things take longer than they should. Add those in. Signed-off-by: Douglas Anderson <dianders@chromium.org> --- drivers/net/usb/r8152.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+)