Here are backports of the three patches that failed to apply to 5.15 due to trivial context conflicts.
Hopefully they apply to the older stable trees as well as-is.
Note that the last patch depends on features that were not added until 5.9 as mentioned in the commit message. Note that the author of that patch did not add a stable tag for this one, but backporting shouldn't hurt.
Johan
Johan Hovold (3): usb: dwc3: fix PHY disable sequence usb: dwc3: qcom: fix use-after-free on runtime-PM wakeup usb: dwc3: disable USB core PHY management
drivers/usb/dwc3/core.c | 19 ++++++++++--------- drivers/usb/dwc3/dwc3-qcom.c | 14 +++++++++++++- drivers/usb/dwc3/host.c | 11 +++++++++++ 3 files changed, 34 insertions(+), 10 deletions(-)
From: Johan Hovold johan+linaro@kernel.org
commit d2ac7bef95c9ead307801ccb6cb6dfbeb14247bf upstream.
Generic PHYs must be powered-off before they can be tore down.
Similarly, suspending legacy PHYs after having powered them off makes no sense.
Fix the dwc3_core_exit() (e.g. called during suspend) and open-coded dwc3_probe() error-path sequences that got this wrong.
Note that this makes dwc3_core_exit() match the dwc3_core_init() error path with respect to powering off the PHYs.
Fixes: 03c1fd622f72 ("usb: dwc3: core: add phy cleanup for probe error handling") Fixes: c499ff71ff2a ("usb: dwc3: core: re-factor init and exit paths") Cc: stable@vger.kernel.org # 4.8 Reviewed-by: Andrew Halaney ahalaney@redhat.com Reviewed-by: Matthias Kaehlcke mka@chromium.org Reviewed-by: Manivannan Sadhasivam manivannan.sadhasivam@linaro.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220804151001.23612-2-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context to 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org --- drivers/usb/dwc3/core.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-)
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index cfac5503aa66..9c24cf46b9a0 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -731,15 +731,16 @@ static void dwc3_core_exit(struct dwc3 *dwc) { dwc3_event_buffers_cleanup(dwc);
+ usb_phy_set_suspend(dwc->usb2_phy, 1); + usb_phy_set_suspend(dwc->usb3_phy, 1); + phy_power_off(dwc->usb2_generic_phy); + phy_power_off(dwc->usb3_generic_phy); + usb_phy_shutdown(dwc->usb2_phy); usb_phy_shutdown(dwc->usb3_phy); phy_exit(dwc->usb2_generic_phy); phy_exit(dwc->usb3_generic_phy);
- usb_phy_set_suspend(dwc->usb2_phy, 1); - usb_phy_set_suspend(dwc->usb3_phy, 1); - phy_power_off(dwc->usb2_generic_phy); - phy_power_off(dwc->usb3_generic_phy); clk_bulk_disable_unprepare(dwc->num_clks, dwc->clks); reset_control_assert(dwc->reset); } @@ -1662,16 +1663,16 @@ static int dwc3_probe(struct platform_device *pdev) dwc3_debugfs_exit(dwc); dwc3_event_buffers_cleanup(dwc);
- usb_phy_shutdown(dwc->usb2_phy); - usb_phy_shutdown(dwc->usb3_phy); - phy_exit(dwc->usb2_generic_phy); - phy_exit(dwc->usb3_generic_phy); - usb_phy_set_suspend(dwc->usb2_phy, 1); usb_phy_set_suspend(dwc->usb3_phy, 1); phy_power_off(dwc->usb2_generic_phy); phy_power_off(dwc->usb3_generic_phy);
+ usb_phy_shutdown(dwc->usb2_phy); + usb_phy_shutdown(dwc->usb3_phy); + phy_exit(dwc->usb2_generic_phy); + phy_exit(dwc->usb3_generic_phy); + dwc3_ulpi_exit(dwc);
err4:
On Tue, Sep 06, 2022 at 02:07:00PM +0200, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit d2ac7bef95c9ead307801ccb6cb6dfbeb14247bf upstream.
Generic PHYs must be powered-off before they can be tore down.
Similarly, suspending legacy PHYs after having powered them off makes no sense.
Fix the dwc3_core_exit() (e.g. called during suspend) and open-coded dwc3_probe() error-path sequences that got this wrong.
Note that this makes dwc3_core_exit() match the dwc3_core_init() error path with respect to powering off the PHYs.
Fixes: 03c1fd622f72 ("usb: dwc3: core: add phy cleanup for probe error handling") Fixes: c499ff71ff2a ("usb: dwc3: core: re-factor init and exit paths") Cc: stable@vger.kernel.org # 4.8 Reviewed-by: Andrew Halaney ahalaney@redhat.com Reviewed-by: Matthias Kaehlcke mka@chromium.org Reviewed-by: Manivannan Sadhasivam manivannan.sadhasivam@linaro.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220804151001.23612-2-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context to 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org
drivers/usb/dwc3/core.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-)
This one did not apply to 4.9.y, 4.14.y, or 4.19.y :(
On Tue, Sep 06, 2022 at 02:16:24PM +0200, Greg Kroah-Hartman wrote:
On Tue, Sep 06, 2022 at 02:07:00PM +0200, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit d2ac7bef95c9ead307801ccb6cb6dfbeb14247bf upstream.
Generic PHYs must be powered-off before they can be tore down.
Similarly, suspending legacy PHYs after having powered them off makes no sense.
Fix the dwc3_core_exit() (e.g. called during suspend) and open-coded dwc3_probe() error-path sequences that got this wrong.
Note that this makes dwc3_core_exit() match the dwc3_core_init() error path with respect to powering off the PHYs.
Fixes: 03c1fd622f72 ("usb: dwc3: core: add phy cleanup for probe error handling") Fixes: c499ff71ff2a ("usb: dwc3: core: re-factor init and exit paths") Cc: stable@vger.kernel.org # 4.8 Reviewed-by: Andrew Halaney ahalaney@redhat.com Reviewed-by: Matthias Kaehlcke mka@chromium.org Reviewed-by: Manivannan Sadhasivam manivannan.sadhasivam@linaro.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220804151001.23612-2-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context to 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org
drivers/usb/dwc3/core.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-)
This one did not apply to 4.9.y, 4.14.y, or 4.19.y :(
Perhaps someone who cares about these old trees can do the backports. Should be as trivial. Can't be the patch submitters responsibility to maintain 8 stable trees.
Johan
On Tue, Sep 06, 2022 at 02:25:25PM +0200, Johan Hovold wrote:
On Tue, Sep 06, 2022 at 02:16:24PM +0200, Greg Kroah-Hartman wrote:
On Tue, Sep 06, 2022 at 02:07:00PM +0200, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit d2ac7bef95c9ead307801ccb6cb6dfbeb14247bf upstream.
Generic PHYs must be powered-off before they can be tore down.
Similarly, suspending legacy PHYs after having powered them off makes no sense.
Fix the dwc3_core_exit() (e.g. called during suspend) and open-coded dwc3_probe() error-path sequences that got this wrong.
Note that this makes dwc3_core_exit() match the dwc3_core_init() error path with respect to powering off the PHYs.
Fixes: 03c1fd622f72 ("usb: dwc3: core: add phy cleanup for probe error handling") Fixes: c499ff71ff2a ("usb: dwc3: core: re-factor init and exit paths") Cc: stable@vger.kernel.org # 4.8 Reviewed-by: Andrew Halaney ahalaney@redhat.com Reviewed-by: Matthias Kaehlcke mka@chromium.org Reviewed-by: Manivannan Sadhasivam manivannan.sadhasivam@linaro.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220804151001.23612-2-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context to 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org
drivers/usb/dwc3/core.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-)
This one did not apply to 4.9.y, 4.14.y, or 4.19.y :(
Perhaps someone who cares about these old trees can do the backports. Should be as trivial. Can't be the patch submitters responsibility to maintain 8 stable trees.
I agree! :)
From: Johan Hovold johan+linaro@kernel.org
commit a872ab303d5ddd4c965f9cd868677781a33ce35a upstream.
The Qualcomm dwc3 runtime-PM implementation checks the xhci platform-device pointer in the wakeup-interrupt handler to determine whether the controller is in host mode and if so triggers a resume.
After a role switch in OTG mode the xhci platform-device would have been freed and the next wakeup from runtime suspend would access the freed memory.
Note that role switching is executed from a freezable workqueue, which guarantees that the pointer is stable during suspend.
Also note that runtime PM has been broken since commit 2664deb09306 ("usb: dwc3: qcom: Honor wakeup enabled/disabled state"), which incidentally also prevents this issue from being triggered.
Fixes: a4333c3a6ba9 ("usb: dwc3: Add Qualcomm DWC3 glue driver") Cc: stable@vger.kernel.org # 4.18 Reviewed-by: Matthias Kaehlcke mka@chromium.org Reviewed-by: Manivannan Sadhasivam manivannan.sadhasivam@linaro.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220804151001.23612-5-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context for 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org --- drivers/usb/dwc3/dwc3-qcom.c | 14 +++++++++++++- drivers/usb/dwc3/host.c | 1 + 2 files changed, 14 insertions(+), 1 deletion(-)
diff --git a/drivers/usb/dwc3/dwc3-qcom.c b/drivers/usb/dwc3/dwc3-qcom.c index 873bf5041117..d0352daab012 100644 --- a/drivers/usb/dwc3/dwc3-qcom.c +++ b/drivers/usb/dwc3/dwc3-qcom.c @@ -296,6 +296,14 @@ static void dwc3_qcom_interconnect_exit(struct dwc3_qcom *qcom) icc_put(qcom->icc_path_apps); }
+/* Only usable in contexts where the role can not change. */ +static bool dwc3_qcom_is_host(struct dwc3_qcom *qcom) +{ + struct dwc3 *dwc = platform_get_drvdata(qcom->dwc3); + + return dwc->xhci; +} + static void dwc3_qcom_disable_interrupts(struct dwc3_qcom *qcom) { if (qcom->hs_phy_irq) { @@ -411,7 +419,11 @@ static irqreturn_t qcom_dwc3_resume_irq(int irq, void *data) if (qcom->pm_suspended) return IRQ_HANDLED;
- if (dwc->xhci) + /* + * This is safe as role switching is done from a freezable workqueue + * and the wakeup interrupts are disabled as part of resume. + */ + if (dwc3_qcom_is_host(qcom)) pm_runtime_resume(&dwc->xhci->dev);
return IRQ_HANDLED; diff --git a/drivers/usb/dwc3/host.c b/drivers/usb/dwc3/host.c index f29a264635aa..2078e9d70292 100644 --- a/drivers/usb/dwc3/host.c +++ b/drivers/usb/dwc3/host.c @@ -130,4 +130,5 @@ int dwc3_host_init(struct dwc3 *dwc) void dwc3_host_exit(struct dwc3 *dwc) { platform_device_unregister(dwc->xhci); + dwc->xhci = NULL; }
On Tue, Sep 06, 2022 at 02:07:01PM +0200, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit a872ab303d5ddd4c965f9cd868677781a33ce35a upstream.
The Qualcomm dwc3 runtime-PM implementation checks the xhci platform-device pointer in the wakeup-interrupt handler to determine whether the controller is in host mode and if so triggers a resume.
After a role switch in OTG mode the xhci platform-device would have been freed and the next wakeup from runtime suspend would access the freed memory.
Note that role switching is executed from a freezable workqueue, which guarantees that the pointer is stable during suspend.
Also note that runtime PM has been broken since commit 2664deb09306 ("usb: dwc3: qcom: Honor wakeup enabled/disabled state"), which incidentally also prevents this issue from being triggered.
Fixes: a4333c3a6ba9 ("usb: dwc3: Add Qualcomm DWC3 glue driver") Cc: stable@vger.kernel.org # 4.18 Reviewed-by: Matthias Kaehlcke mka@chromium.org Reviewed-by: Manivannan Sadhasivam manivannan.sadhasivam@linaro.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220804151001.23612-5-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context for 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org
drivers/usb/dwc3/dwc3-qcom.c | 14 +++++++++++++- drivers/usb/dwc3/host.c | 1 + 2 files changed, 14 insertions(+), 1 deletion(-)
This one did not apply to 5.4.y or 4.19.y :(
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
Fixes: 4e88d4c08301 ("usb: add a flag to skip PHY initialization to struct usb_hcd") Fixes: 178a0bce05cb ("usb: core: hcd: integrate the PHY wrapper into the HCD core") Tested-by: Matthias Kaehlcke mka@chromium.org Cc: stable stable@kernel.org Reviewed-by: Matthias Kaehlcke mka@chromium.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220825131836.19769-1-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context to 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org --- drivers/usb/dwc3/host.c | 10 ++++++++++ 1 file changed, 10 insertions(+)
diff --git a/drivers/usb/dwc3/host.c b/drivers/usb/dwc3/host.c index 2078e9d70292..85165a972076 100644 --- a/drivers/usb/dwc3/host.c +++ b/drivers/usb/dwc3/host.c @@ -10,8 +10,13 @@ #include <linux/acpi.h> #include <linux/platform_device.h>
+#include "../host/xhci-plat.h" #include "core.h"
+static const struct xhci_plat_priv dwc3_xhci_plat_priv = { + .quirks = XHCI_SKIP_PHY_INIT, +}; + static int dwc3_host_get_irq(struct dwc3 *dwc) { struct platform_device *dwc3_pdev = to_platform_device(dwc->dev); @@ -87,6 +92,11 @@ int dwc3_host_init(struct dwc3 *dwc) goto err; }
+ ret = platform_device_add_data(xhci, &dwc3_xhci_plat_priv, + sizeof(dwc3_xhci_plat_priv)); + if (ret) + goto err; + memset(props, 0, sizeof(struct property_entry) * ARRAY_SIZE(props));
if (dwc->usb3_lpm_capable)
On Tue, Sep 06, 2022 at 02:07:02PM +0200, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
Fixes: 4e88d4c08301 ("usb: add a flag to skip PHY initialization to struct usb_hcd") Fixes: 178a0bce05cb ("usb: core: hcd: integrate the PHY wrapper into the HCD core") Tested-by: Matthias Kaehlcke mka@chromium.org Cc: stable stable@kernel.org Reviewed-by: Matthias Kaehlcke mka@chromium.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220825131836.19769-1-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context to 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org
drivers/usb/dwc3/host.c | 10 ++++++++++ 1 file changed, 10 insertions(+)
Breaks the build on 4.19.y :(
On Tue, Sep 06, 2022 at 02:18:08PM +0200, Greg Kroah-Hartman wrote:
On Tue, Sep 06, 2022 at 02:07:02PM +0200, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
Fixes: 4e88d4c08301 ("usb: add a flag to skip PHY initialization to struct usb_hcd") Fixes: 178a0bce05cb ("usb: core: hcd: integrate the PHY wrapper into the HCD core") Tested-by: Matthias Kaehlcke mka@chromium.org Cc: stable stable@kernel.org Reviewed-by: Matthias Kaehlcke mka@chromium.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220825131836.19769-1-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context to 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org
drivers/usb/dwc3/host.c | 10 ++++++++++ 1 file changed, 10 insertions(+)
Breaks the build on 4.19.y :(
It's OK to not to backport this one.
Johan
Hi Johan,
On 2022-09-06 14:07, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
[adding also Samsung/ODROID device tree authors Krzysztof and Marek]
For some reason, this commit seems to break detection of the USB to S-ATA controller on ODROID-HC1 devices (Exynos 5422).
We have a known to work OS release of v5.15.60, and known to not be working of v5.15.67. By reverting suspicious commits, I was able to pinpoint the problem to this particular commit.
From what I understand, on that particular hardware the S-ATA controller power is controlled via the V-BUS signal VBUSCTRL_U2 (Schematic [1]). Presumably this signal is no longer controlled with this change.
This came up in our HAOS issue #2153 [2].
-- Stefan
[1] https://dn.odroid.com/5422/ODROID-HC1_MC1/Schematics/HC1_MAIN_REV0.1_2017063... [2] https://github.com/home-assistant/operating-system/issues/2153
Fixes: 4e88d4c08301 ("usb: add a flag to skip PHY initialization to struct usb_hcd") Fixes: 178a0bce05cb ("usb: core: hcd: integrate the PHY wrapper into the HCD core") Tested-by: Matthias Kaehlcke mka@chromium.org Cc: stable stable@kernel.org Reviewed-by: Matthias Kaehlcke mka@chromium.org Signed-off-by: Johan Hovold johan+linaro@kernel.org Link: https://lore.kernel.org/r/20220825131836.19769-1-johan+linaro@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org [ johan: adjust context to 5.15 ] Signed-off-by: Johan Hovold johan+linaro@kernel.org
drivers/usb/dwc3/host.c | 10 ++++++++++ 1 file changed, 10 insertions(+)
diff --git a/drivers/usb/dwc3/host.c b/drivers/usb/dwc3/host.c index 2078e9d70292..85165a972076 100644 --- a/drivers/usb/dwc3/host.c +++ b/drivers/usb/dwc3/host.c @@ -10,8 +10,13 @@ #include <linux/acpi.h> #include <linux/platform_device.h> +#include "../host/xhci-plat.h" #include "core.h" +static const struct xhci_plat_priv dwc3_xhci_plat_priv = {
- .quirks = XHCI_SKIP_PHY_INIT,
+};
static int dwc3_host_get_irq(struct dwc3 *dwc) { struct platform_device *dwc3_pdev = to_platform_device(dwc->dev); @@ -87,6 +92,11 @@ int dwc3_host_init(struct dwc3 *dwc) goto err; }
- ret = platform_device_add_data(xhci, &dwc3_xhci_plat_priv,
sizeof(dwc3_xhci_plat_priv));- if (ret)
goto err;- memset(props, 0, sizeof(struct property_entry) * ARRAY_SIZE(props));
if (dwc->usb3_lpm_capable)
On Tue, Oct 18, 2022 at 05:27:24PM +0200, Stefan Agner wrote:
Hi Johan,
On 2022-09-06 14:07, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
[adding also Samsung/ODROID device tree authors Krzysztof and Marek]
For some reason, this commit seems to break detection of the USB to S-ATA controller on ODROID-HC1 devices (Exynos 5422).
We have a known to work OS release of v5.15.60, and known to not be working of v5.15.67. By reverting suspicious commits, I was able to pinpoint the problem to this particular commit.
From what I understand, on that particular hardware the S-ATA controller power is controlled via the V-BUS signal VBUSCTRL_U2 (Schematic [1]). Presumably this signal is no longer controlled with this change.
This came up in our HAOS issue #2153 [2].
Thanks for the report and sorry about the breakage. This wasn't supposed to go to stable but Greg thought otherwise (and I helped with the backporting to prevent autosel from pulling in even more changes).
But at least this way we found out sooner that this platform depends on having both USB core and dwc3 managing the same PHY.
I think this may be related to the calibration calls added to dwc3 and later removed again by commits:
d8c80bb3b55b ("phy: exynos5-usbdrd: Calibrate LOS levels for exynos5420/5800") a0a465569b45 ("usb: dwc3: remove generic PHY calibrate() calls")
The removal explicitly mentions that the expectation is that USB core will do the PHY calibration.
There could be other changes in the sequencing of events that this platform has been implicitly relying on, but as a start, could try adding the missing calibration calls (patch below) and see if that makes a difference?
Johan
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index 46cf8edf7f93..f8a0e340eb63 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -198,6 +198,8 @@ static void __dwc3_set_mode(struct work_struct *work) otg_set_vbus(dwc->usb2_phy->otg, true); phy_set_mode(dwc->usb2_generic_phy, PHY_MODE_USB_HOST); phy_set_mode(dwc->usb3_generic_phy, PHY_MODE_USB_HOST); + phy_calibrate(dwc->usb2_generic_phy); + phy_calibrate(dwc->usb3_generic_phy); if (dwc->dis_split_quirk) { reg = dwc3_readl(dwc->regs, DWC3_GUCTL3); reg |= DWC3_GUCTL3_SPLITDISABLE; @@ -1369,6 +1371,8 @@ static int dwc3_core_init_mode(struct dwc3 *dwc) otg_set_vbus(dwc->usb2_phy->otg, true); phy_set_mode(dwc->usb2_generic_phy, PHY_MODE_USB_HOST); phy_set_mode(dwc->usb3_generic_phy, PHY_MODE_USB_HOST); + phy_calibrate(dwc->usb2_generic_phy); + phy_calibrate(dwc->usb3_generic_phy);
ret = dwc3_host_init(dwc); if (ret)
Hi Johan,
On 2022-10-19 10:59, Johan Hovold wrote:
On Tue, Oct 18, 2022 at 05:27:24PM +0200, Stefan Agner wrote:
Hi Johan,
On 2022-09-06 14:07, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
[adding also Samsung/ODROID device tree authors Krzysztof and Marek]
For some reason, this commit seems to break detection of the USB to S-ATA controller on ODROID-HC1 devices (Exynos 5422).
We have a known to work OS release of v5.15.60, and known to not be working of v5.15.67. By reverting suspicious commits, I was able to pinpoint the problem to this particular commit.
From what I understand, on that particular hardware the S-ATA controller power is controlled via the V-BUS signal VBUSCTRL_U2 (Schematic [1]). Presumably this signal is no longer controlled with this change.
This came up in our HAOS issue #2153 [2].
Thanks for the report and sorry about the breakage. This wasn't supposed to go to stable but Greg thought otherwise (and I helped with the backporting to prevent autosel from pulling in even more changes).
But at least this way we found out sooner that this platform depends on having both USB core and dwc3 managing the same PHY.
I think this may be related to the calibration calls added to dwc3 and later removed again by commits:
d8c80bb3b55b ("phy: exynos5-usbdrd: Calibrate LOS levels for exynos5420/5800") a0a465569b45 ("usb: dwc3: remove generic PHY calibrate() calls")
The removal explicitly mentions that the expectation is that USB core will do the PHY calibration.
There could be other changes in the sequencing of events that this platform has been implicitly relying on, but as a start, could try adding the missing calibration calls (patch below) and see if that makes a difference?
The patch below did not apply to 5.15.74 directly, but I think I was able to get the corrected patch applied (see below)
That said, I do not have direct access to that hardware, but I created a build and asked the user test it.
Best regards, Stefan
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index c32ca691bcc7..964f512603ec 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -196,6 +196,8 @@ static void __dwc3_set_mode(struct work_struct *work) otg_set_vbus(dwc->usb2_phy->otg, true); phy_set_mode(dwc->usb2_generic_phy, PHY_MODE_USB_HOST); phy_set_mode(dwc->usb3_generic_phy, PHY_MODE_USB_HOST); + phy_calibrate(dwc->usb2_generic_phy); + phy_calibrate(dwc->usb3_generic_phy); if (dwc->dis_split_quirk) { reg = dwc3_readl(dwc->regs, DWC3_GUCTL3); reg |= DWC3_GUCTL3_SPLITDISABLE; @@ -1227,6 +1229,8 @@ static int dwc3_core_init_mode(struct dwc3 *dwc) otg_set_vbus(dwc->usb2_phy->otg, true); phy_set_mode(dwc->usb2_generic_phy, PHY_MODE_USB_HOST); phy_set_mode(dwc->usb3_generic_phy, PHY_MODE_USB_HOST); + phy_calibrate(dwc->usb2_generic_phy); + phy_calibrate(dwc->usb3_generic_phy);
ret = dwc3_host_init(dwc); if (ret)
-- Stefan
On Fri, Oct 21, 2022 at 12:06:12AM +0200, Stefan Agner wrote:
On 2022-10-19 10:59, Johan Hovold wrote:
On Tue, Oct 18, 2022 at 05:27:24PM +0200, Stefan Agner wrote:
On 2022-09-06 14:07, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
For some reason, this commit seems to break detection of the USB to S-ATA controller on ODROID-HC1 devices (Exynos 5422).
I think this may be related to the calibration calls added to dwc3 and later removed again by commits:
d8c80bb3b55b ("phy: exynos5-usbdrd: Calibrate LOS levels for exynos5420/5800") a0a465569b45 ("usb: dwc3: remove generic PHY calibrate() calls")
The removal explicitly mentions that the expectation is that USB core will do the PHY calibration.
There could be other changes in the sequencing of events that this platform has been implicitly relying on, but as a start, could try adding the missing calibration calls (patch below) and see if that makes a difference?
The patch below did not apply to 5.15.74 directly, but I think I was able to get the corrected patch applied (see below)
Looks good to me.
That said, I do not have direct access to that hardware, but I created a build and asked the user test it.
Thanks, let me know how it goes.
Johan
On 2022-10-21 08:54, Johan Hovold wrote:
On Fri, Oct 21, 2022 at 12:06:12AM +0200, Stefan Agner wrote:
On 2022-10-19 10:59, Johan Hovold wrote:
On Tue, Oct 18, 2022 at 05:27:24PM +0200, Stefan Agner wrote:
On 2022-09-06 14:07, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
For some reason, this commit seems to break detection of the USB to S-ATA controller on ODROID-HC1 devices (Exynos 5422).
I think this may be related to the calibration calls added to dwc3 and later removed again by commits:
d8c80bb3b55b ("phy: exynos5-usbdrd: Calibrate LOS levels for exynos5420/5800") a0a465569b45 ("usb: dwc3: remove generic PHY calibrate() calls")
The removal explicitly mentions that the expectation is that USB core will do the PHY calibration.
There could be other changes in the sequencing of events that this platform has been implicitly relying on, but as a start, could try adding the missing calibration calls (patch below) and see if that makes a difference?
The patch below did not apply to 5.15.74 directly, but I think I was able to get the corrected patch applied (see below)
Looks good to me.
That said, I do not have direct access to that hardware, but I created a build and asked the user test it.
Thanks, let me know how it goes.
The user reports the S-ATA disk is *not* recognized with that patch applied.
-- Stefan
On Wed, Oct 26, 2022 at 03:11:00PM +0200, Stefan Agner wrote:
On 2022-10-21 08:54, Johan Hovold wrote:
On Fri, Oct 21, 2022 at 12:06:12AM +0200, Stefan Agner wrote:
On 2022-10-19 10:59, Johan Hovold wrote:
On Tue, Oct 18, 2022 at 05:27:24PM +0200, Stefan Agner wrote:
On 2022-09-06 14:07, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
For some reason, this commit seems to break detection of the USB to S-ATA controller on ODROID-HC1 devices (Exynos 5422).
I think this may be related to the calibration calls added to dwc3 and later removed again by commits:
d8c80bb3b55b ("phy: exynos5-usbdrd: Calibrate LOS levels for exynos5420/5800") a0a465569b45 ("usb: dwc3: remove generic PHY calibrate() calls")
The removal explicitly mentions that the expectation is that USB core will do the PHY calibration.
There could be other changes in the sequencing of events that this platform has been implicitly relying on, but as a start, could try adding the missing calibration calls (patch below) and see if that makes a difference?
The patch below did not apply to 5.15.74 directly, but I think I was able to get the corrected patch applied (see below)
The user reports the S-ATA disk is *not* recognized with that patch applied.
I just noticed a mistake in the instrumentation patch I sent you. Could you try moving the calibrations calls after dwc3_host_init() (e.g. as in the second chunk in the diff below)?
As mentioned in the commit message for a0a465569b45 ("usb: dwc3: remove generic PHY calibrate() calls"), this may not work if the xhci-plat driver is built as a module and there are some corner cases that it does not cover.
It seems we should revert the offending commit and then try to find some time to untangle this mess, but please check if the below addresses the issue first so we know what the problem is.
I'll prepare a revert in the meantime.
Johan
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index 31156d4dec9f..37d49a394912 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -197,6 +197,8 @@ static void __dwc3_set_mode(struct work_struct *work) otg_set_vbus(dwc->usb2_phy->otg, true); phy_set_mode(dwc->usb2_generic_phy, PHY_MODE_USB_HOST); phy_set_mode(dwc->usb3_generic_phy, PHY_MODE_USB_HOST); + phy_calibrate(dwc->usb2_generic_phy); + phy_calibrate(dwc->usb3_generic_phy); if (dwc->dis_split_quirk) { reg = dwc3_readl(dwc->regs, DWC3_GUCTL3); reg |= DWC3_GUCTL3_SPLITDISABLE; @@ -1391,6 +1393,9 @@ static int dwc3_core_init_mode(struct dwc3 *dwc) ret = dwc3_host_init(dwc); if (ret) return dev_err_probe(dev, ret, "failed to initialize host\n"); + + phy_calibrate(dwc->usb2_generic_phy); + phy_calibrate(dwc->usb3_generic_phy); break; case USB_DR_MODE_OTG: INIT_WORK(&dwc->drd_work, __dwc3_set_mode);
On Thu, Oct 27, 2022 at 02:45:15PM +0200, Johan Hovold wrote:
On Wed, Oct 26, 2022 at 03:11:00PM +0200, Stefan Agner wrote:
The user reports the S-ATA disk is *not* recognized with that patch applied.
I just noticed a mistake in the instrumentation patch I sent you. Could you try moving the calibrations calls after dwc3_host_init() (e.g. as in the second chunk in the diff below)?
As mentioned in the commit message for a0a465569b45 ("usb: dwc3: remove generic PHY calibrate() calls"), this may not work if the xhci-plat driver is built as a module and there are some corner cases that it does not cover.
It seems we should revert the offending commit and then try to find some time to untangle this mess, but please check if the below addresses the issue first so we know what the problem is.
I'll prepare a revert in the meantime.
I've now posted the revert, but please do check if the below patch was enough to resolve the immediate issue.
Johan
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index 31156d4dec9f..37d49a394912 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -197,6 +197,8 @@ static void __dwc3_set_mode(struct work_struct *work) otg_set_vbus(dwc->usb2_phy->otg, true); phy_set_mode(dwc->usb2_generic_phy, PHY_MODE_USB_HOST); phy_set_mode(dwc->usb3_generic_phy, PHY_MODE_USB_HOST);
phy_calibrate(dwc->usb2_generic_phy);phy_calibrate(dwc->usb3_generic_phy); if (dwc->dis_split_quirk) { reg = dwc3_readl(dwc->regs, DWC3_GUCTL3); reg |= DWC3_GUCTL3_SPLITDISABLE;@@ -1391,6 +1393,9 @@ static int dwc3_core_init_mode(struct dwc3 *dwc) ret = dwc3_host_init(dwc); if (ret) return dev_err_probe(dev, ret, "failed to initialize host\n");
phy_calibrate(dwc->usb2_generic_phy);phy_calibrate(dwc->usb3_generic_phy); break; case USB_DR_MODE_OTG: INIT_WORK(&dwc->drd_work, __dwc3_set_mode);
Hi Johan,
On 03.11.2022 15:49, Johan Hovold wrote:
On Thu, Oct 27, 2022 at 02:45:15PM +0200, Johan Hovold wrote:
On Wed, Oct 26, 2022 at 03:11:00PM +0200, Stefan Agner wrote:
The user reports the S-ATA disk is *not* recognized with that patch applied.
I just noticed a mistake in the instrumentation patch I sent you. Could you try moving the calibrations calls after dwc3_host_init() (e.g. as in the second chunk in the diff below)?
As mentioned in the commit message for a0a465569b45 ("usb: dwc3: remove generic PHY calibrate() calls"), this may not work if the xhci-plat driver is built as a module and there are some corner cases that it does not cover.
It seems we should revert the offending commit and then try to find some time to untangle this mess, but please check if the below addresses the issue first so we know what the problem is.
I'll prepare a revert in the meantime.
I've now posted the revert, but please do check if the below patch was enough to resolve the immediate issue.
The below patch was a half-fix. It worked only if both dwc3 and xhci_plat_hcd were compiled into the kernel. Afair Debian-based distros used xhci compiled as a module, so this didn't work for that case due to timing issues.
Johan
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index 31156d4dec9f..37d49a394912 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -197,6 +197,8 @@ static void __dwc3_set_mode(struct work_struct *work) otg_set_vbus(dwc->usb2_phy->otg, true); phy_set_mode(dwc->usb2_generic_phy, PHY_MODE_USB_HOST); phy_set_mode(dwc->usb3_generic_phy, PHY_MODE_USB_HOST);
phy_calibrate(dwc->usb2_generic_phy);phy_calibrate(dwc->usb3_generic_phy); if (dwc->dis_split_quirk) { reg = dwc3_readl(dwc->regs, DWC3_GUCTL3); reg |= DWC3_GUCTL3_SPLITDISABLE;@@ -1391,6 +1393,9 @@ static int dwc3_core_init_mode(struct dwc3 *dwc) ret = dwc3_host_init(dwc); if (ret) return dev_err_probe(dev, ret, "failed to initialize host\n");
phy_calibrate(dwc->usb2_generic_phy);phy_calibrate(dwc->usb3_generic_phy); break; case USB_DR_MODE_OTG: INIT_WORK(&dwc->drd_work, __dwc3_set_mode);
Best regards
On Thu, Nov 03, 2022 at 04:18:12PM +0100, Marek Szyprowski wrote:
Hi Johan,
On 03.11.2022 15:49, Johan Hovold wrote:
On Thu, Oct 27, 2022 at 02:45:15PM +0200, Johan Hovold wrote:
On Wed, Oct 26, 2022 at 03:11:00PM +0200, Stefan Agner wrote:
The user reports the S-ATA disk is *not* recognized with that patch applied.
I just noticed a mistake in the instrumentation patch I sent you. Could you try moving the calibrations calls after dwc3_host_init() (e.g. as in the second chunk in the diff below)?
As mentioned in the commit message for a0a465569b45 ("usb: dwc3: remove generic PHY calibrate() calls"), this may not work if the xhci-plat driver is built as a module and there are some corner cases that it does not cover.
It seems we should revert the offending commit and then try to find some time to untangle this mess, but please check if the below addresses the issue first so we know what the problem is.
I'll prepare a revert in the meantime.
I've now posted the revert, but please do check if the below patch was enough to resolve the immediate issue.
The below patch was a half-fix. It worked only if both dwc3 and xhci_plat_hcd were compiled into the kernel. Afair Debian-based distros used xhci compiled as a module, so this didn't work for that case due to timing issues.
Yeah, I mentioned that above too. The intention here was just to confirm the hypothesis that the regression was due to the missing calibration calls.
Johan
Dear All,
On 18.10.2022 17:27, Stefan Agner wrote:
Hi Johan,
On 2022-09-06 14:07, Johan Hovold wrote:
From: Johan Hovold johan+linaro@kernel.org
commit 6000b8d900cd5f52fbcd0776d0cc396e88c8c2ea upstream.
The dwc3 driver manages its PHYs itself so the USB core PHY management needs to be disabled.
Use the struct xhci_plat_priv hack added by commits 46034a999c07 ("usb: host: xhci-plat: add platform data support") and f768e718911e ("usb: host: xhci-plat: add priv quirk for skip PHY initialization") to propagate the setting for now.
[adding also Samsung/ODROID device tree authors Krzysztof and Marek]
For some reason, this commit seems to break detection of the USB to S-ATA controller on ODROID-HC1 devices (Exynos 5422).
We have a known to work OS release of v5.15.60, and known to not be working of v5.15.67. By reverting suspicious commits, I was able to pinpoint the problem to this particular commit.
From what I understand, on that particular hardware the S-ATA controller
power is controlled via the V-BUS signal VBUSCTRL_U2 (Schematic [1]). Presumably this signal is no longer controlled with this change.
This came up in our HAOS issue #2153 [2].
I confirm this issue and I've managed to reproduce it locally. The mainline is also affected. I will try to prepare a proper patch soon.
Best regards
linux-stable-mirror@lists.linaro.org
-
Greg Kroah-Hartman -
Johan Hovold -
Marek Szyprowski -
Stefan Agner