| Message ID | 20230808032510.31754-1-luke.lu@libre.computer (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | usb: dwc3: meson-g12a: do post init to fix broken usb after resumption | expand |
On Tue, Aug 08, 2023 at 03:25:10AM +0000, Luke Lu wrote: > Device connected to usb otg port of GXL-based boards can not be > recognised after resumption, doesn't recover even if disconnect and > reconnect the device. dmesg shows it disconnects during resumption. > > [ 41.492911] usb 1-2: USB disconnect, device number 3 > [ 41.499346] usb 1-2: unregistering device > [ 41.511939] usb 1-2: unregistering interface 1-2:1.0 > > Calling usb_post_init() will fix this issue, and it's tested and > verified on libretech's aml-s905x-cc board. > > Signed-off-by: Luke Lu <luke.lu@libre.computer> What commit id does this fix? Should it also go to stable kernels? thanks, greg k-h
On Tue, Aug 8, 2023 at 12:56 AM Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote: > > On Tue, Aug 08, 2023 at 03:25:10AM +0000, Luke Lu wrote: > > Device connected to usb otg port of GXL-based boards can not be > > recognised after resumption, doesn't recover even if disconnect and > > reconnect the device. dmesg shows it disconnects during resumption. > > > > [ 41.492911] usb 1-2: USB disconnect, device number 3 > > [ 41.499346] usb 1-2: unregistering device > > [ 41.511939] usb 1-2: unregistering interface 1-2:1.0 > > > > Calling usb_post_init() will fix this issue, and it's tested and > > verified on libretech's aml-s905x-cc board. > > > > Signed-off-by: Luke Lu <luke.lu@libre.computer> > > What commit id does this fix? Should it also go to stable kernels? Hi Greg, Suspend device tree flag has not been enabled in-tree for the Amlogic GXL (S805X/S905X/S905D) series due it having PSCI 0.2 and thus never untested. Most of the suspend resume code was only tested on newer G12A/B series. This and "[PATCH v3] net: phy: meson-gxl: implement meson_gxl_phy_resume()" fixes USB and Ethernet after resume. HDMI requires another fix. It would be great if they could go into stable since we are carrying these two patches in our tree on 6.1 and they're applicable back to 5.4 but I am not sure what the criteria is. Best, Da > > thanks, > > greg k-h
HI Greg: On Tue, Aug 8, 2023 at 4:56 AM Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote: > > On Tue, Aug 08, 2023 at 03:25:10AM +0000, Luke Lu wrote: > > Device connected to usb otg port of GXL-based boards can not be > > recognised after resumption, doesn't recover even if disconnect and > > reconnect the device. dmesg shows it disconnects during resumption. > > > > [ 41.492911] usb 1-2: USB disconnect, device number 3 > > [ 41.499346] usb 1-2: unregistering device > > [ 41.511939] usb 1-2: unregistering interface 1-2:1.0 > > > > Calling usb_post_init() will fix this issue, and it's tested and > > verified on libretech's aml-s905x-cc board. > > > > Signed-off-by: Luke Lu <luke.lu@libre.computer> > > What commit id does this fix? Using "git blame" to explore the history, found dwc3_meson_g12a_resume() was introduced along with the file of drivers/usb/dwc3/dwc3-meson-g12a.c. as Da Xue pointed out, the suspend/resume was never tested in GXL based SoC, so it's broken since the beginning.. For the Fixes tag, I think it's proper to use "5b0ba0caaf3a: (usb: dwc3: meson-g12a: refactor usb init)" since the usb_post_init() function was introduced in this commit and this patch will depend on it. > Should it also go to stable kernels? > Yes, It would be great if the patch can go to stable tree, thanks for suggesting this I have it tested on 6.1-lts tree. Btw, I would wait for a few days for more comments and CC to stable in next v2 if no objection. > thanks, > > greg k-h Luke
On Tue, Aug 08, 2023 at 01:33:45AM -0400, Da Xue wrote: > On Tue, Aug 8, 2023 at 12:56 AM Greg Kroah-Hartman > <gregkh@linuxfoundation.org> wrote: > > > > On Tue, Aug 08, 2023 at 03:25:10AM +0000, Luke Lu wrote: > > > Device connected to usb otg port of GXL-based boards can not be > > > recognised after resumption, doesn't recover even if disconnect and > > > reconnect the device. dmesg shows it disconnects during resumption. > > > > > > [ 41.492911] usb 1-2: USB disconnect, device number 3 > > > [ 41.499346] usb 1-2: unregistering device > > > [ 41.511939] usb 1-2: unregistering interface 1-2:1.0 > > > > > > Calling usb_post_init() will fix this issue, and it's tested and > > > verified on libretech's aml-s905x-cc board. > > > > > > Signed-off-by: Luke Lu <luke.lu@libre.computer> > > > > What commit id does this fix? Should it also go to stable kernels? > > Hi Greg, > > Suspend device tree flag has not been enabled in-tree for the Amlogic > GXL (S805X/S905X/S905D) series due it having PSCI 0.2 and thus never > untested. Most of the suspend resume code was only tested on newer > G12A/B series. > > This and "[PATCH v3] net: phy: meson-gxl: implement > meson_gxl_phy_resume()" fixes USB and Ethernet after resume. HDMI > requires another fix. > > It would be great if they could go into stable since we are carrying > these two patches in our tree on 6.1 and they're applicable back to > 5.4 but I am not sure what the criteria is. https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html has the criteria. If you think it should go there, then please resend with a cc: stable tag in the signed-off-by area. thanks, greg k-h
On Tue, Aug 08, 2023 at 06:09:45AM +0000, Luke Lu wrote: > HI Greg: > > On Tue, Aug 8, 2023 at 4:56 AM Greg Kroah-Hartman > <gregkh@linuxfoundation.org> wrote: > > > > On Tue, Aug 08, 2023 at 03:25:10AM +0000, Luke Lu wrote: > > > Device connected to usb otg port of GXL-based boards can not be > > > recognised after resumption, doesn't recover even if disconnect and > > > reconnect the device. dmesg shows it disconnects during resumption. > > > > > > [ 41.492911] usb 1-2: USB disconnect, device number 3 > > > [ 41.499346] usb 1-2: unregistering device > > > [ 41.511939] usb 1-2: unregistering interface 1-2:1.0 > > > > > > Calling usb_post_init() will fix this issue, and it's tested and > > > verified on libretech's aml-s905x-cc board. > > > > > > Signed-off-by: Luke Lu <luke.lu@libre.computer> > > > > What commit id does this fix? > Using "git blame" to explore the history, found dwc3_meson_g12a_resume() was > introduced along with the file of drivers/usb/dwc3/dwc3-meson-g12a.c. > as Da Xue pointed out, the suspend/resume was never tested in GXL based SoC, > so it's broken since the beginning.. > > For the Fixes tag, I think it's proper to use "5b0ba0caaf3a: (usb: > dwc3: meson-g12a: refactor usb init)" > since the usb_post_init() function was introduced in this commit and > this patch will depend on it. > > > Should it also go to stable kernels? > > > Yes, It would be great if the patch can go to stable tree, thanks for > suggesting this > I have it tested on 6.1-lts tree. > > Btw, I would wait for a few days for more comments and CC to stable in > next v2 if no objection. No worries, I'll drop this from my review queue and wait for a v2. thanks, greg k-h
diff --git a/drivers/usb/dwc3/dwc3-meson-g12a.c b/drivers/usb/dwc3/dwc3-meson-g12a.c index e99c7489dba0..2c07c038b584 100644 --- a/drivers/usb/dwc3/dwc3-meson-g12a.c +++ b/drivers/usb/dwc3/dwc3-meson-g12a.c @@ -926,6 +926,12 @@ static int __maybe_unused dwc3_meson_g12a_resume(struct device *dev) return ret; } + if (priv->drvdata->usb_post_init) { + ret = priv->drvdata->usb_post_init(priv); + if (ret) + return ret; + } + return 0; }
Device connected to usb otg port of GXL-based boards can not be recognised after resumption, doesn't recover even if disconnect and reconnect the device. dmesg shows it disconnects during resumption. [ 41.492911] usb 1-2: USB disconnect, device number 3 [ 41.499346] usb 1-2: unregistering device [ 41.511939] usb 1-2: unregistering interface 1-2:1.0 Calling usb_post_init() will fix this issue, and it's tested and verified on libretech's aml-s905x-cc board. Signed-off-by: Luke Lu <luke.lu@libre.computer> --- drivers/usb/dwc3/dwc3-meson-g12a.c | 6 ++++++ 1 file changed, 6 insertions(+)