Message ID | 20240307062052.2319851-1-quic_kriskura@quicinc.com |
---|---|
Headers | show |
Series | Add multiport support for DWC3 controllers | expand |
On Thu, Mar 07, 2024 at 11:50:45AM +0530, Krishna Kurapati wrote: > All DWC3 Multi Port controllers that exist today only support host mode. > Temporarily map XHCI address space for host-only controllers and parse > XHCI Extended Capabilities registers to read number of usb2 ports and > usb3 ports present on multiport controller. Each USB Port is at least HS > capable. > > The port info for usb2 and usb3 phy are identified as num_usb2_ports > and num_usb3_ports. The intention is as follows: > > Wherever we need to perform phy operations like: > > LOOP_OVER_NUMBER_OF_AVAILABLE_PORTS() > { > phy_set_mode(dwc->usb2_generic_phy[i], PHY_MODE_USB_HOST); > phy_set_mode(dwc->usb3_generic_phy[i], PHY_MODE_USB_HOST); > } > > If number of usb2 ports is 3, loop can go from index 0-2 for > usb2_generic_phy. If number of usb3-ports is 2, we don't know for sure, > if the first 2 ports are SS capable or some other ports like (2 and 3) > are SS capable. So instead, num_usb2_ports is used to loop around all > phy's (both hs and ss) for performing phy operations. If any > usb3_generic_phy turns out to be NULL, phy operation just bails out. > num_usb3_ports is used to modify GUSB3PIPECTL registers while setting up > phy's as we need to know how many SS capable ports are there for this. > > Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com> Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
On Thu, Mar 07, 2024 at 11:50:47AM +0530, Krishna Kurapati wrote: > Currently the DWC3 driver supports only single port controller > which requires at least one HS PHY and at most one SS PHY. > > But the DWC3 USB controller can be connected to multiple ports and > each port can have their own PHYs. Each port of the multiport > controller can either be HS+SS capable or HS only capable > Proper quantification of them is required to modify GUSB2PHYCFG > and GUSB3PIPECTL registers appropriately. > > Add support for detecting, obtaining and configuring PHYs supported > by a multiport controller. Limit support to multiport controllers > with up to four ports for now (e.g. as needed for SC8280XP). > > Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com> Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
On Thu, Mar 07, 2024 at 11:50:49AM +0530, Krishna Kurapati wrote: > The logic for requesting interrupts is duplicated for each interrupt. In > the upcoming patches that introduces support for multiport, it would be > better to clean up the duplication before reading mulitport related > interrupts. > > Refactor interrupt setup call by adding a new helper function for > requesting the wakeup interrupts. To simplify implementation, make > the display name same as the interrupt name expected in DT. > > Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com> Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
On Thu, Mar 07, 2024 at 11:50:50AM +0530, Krishna Kurapati wrote: > On multiport supported controllers, each port has its own DP/DM > and SS (if super speed capable) interrupts. As per the bindings, > their interrupt names differ from standard ones having "_x" added > as suffix (x indicates port number). Identify from the interrupt > names whether the controller is a multiport controller or not. > Refactor dwc3_qcom_setup_irq() call to parse multiport interrupts > along with non-multiport ones accordingly.. > > Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com> > --- > drivers/usb/dwc3/dwc3-qcom.c | 146 +++++++++++++++++++++++++++-------- > 1 file changed, 112 insertions(+), 34 deletions(-) This is much better. Just a couple of nits below. > +static int dwc3_qcom_find_num_ports(struct platform_device *pdev) > +{ > + const char *irq_name; > + int port_index; > + int irq; > + > + irq = platform_get_irq_byname_optional(pdev, "qusb2_phy"); > + if (irq > 0) > + return 1; > + > + irq = platform_get_irq_byname_optional(pdev, "dp_hs_phy_irq"); > + if (irq > 0) > + return 1; > + > + for (port_index = 0; port_index < DWC3_MAX_PORTS; port_index++) { > + irq_name = devm_kasprintf(&pdev->dev, GFP_KERNEL, "dp_hs_phy_%d", port_index + 1); Please use a stack allocated buffer for these strings as we don't need them any more after this function returns. > + if (!irq_name) > + return -ENOMEM; > + > + irq = platform_get_irq_byname_optional(pdev, irq_name); > + if (irq <= 0) > + return port_index; > + } > + > + return port_index; I think explicitly returning DWC3_MAX_PORTS here would be more readable. > +} > + > +static int dwc3_qcom_setup_irq(struct platform_device *pdev) > +{ > + struct dwc3_qcom *qcom = platform_get_drvdata(pdev); > + bool is_multiport; > + int ret; > + int i; > + > + qcom->num_ports = dwc3_qcom_find_num_ports(pdev); > + if (qcom->num_ports < 0) > + return -ENOMEM; Just return 'ret' directly. > + > + is_multiport = (qcom->num_ports > 1) ? true : false; And no need for the ternary operator: is_multiport = (qcom->num_ports > 1); > + > + for (i = 0; i < qcom->num_ports; i++) { > + ret = dwc3_qcom_setup_port_irq(pdev, i, is_multiport); > if (ret) > return ret; > - qcom->ss_phy_irq = irq; > } > > return 0; With that fixed: Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
On Thu, Mar 07, 2024 at 11:50:51AM +0530, Krishna Kurapati wrote: > DWC3 Qcom wrapper currently supports only wakeup configuration > for single port controllers. Read speed of each port connected > to the controller and enable wakeup for each of them accordingly. > > Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com> Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
On Thu, Mar 07, 2024 at 11:50:52AM +0530, Krishna Kurapati wrote: > Power event IRQ is used for wakeup in cases: > a) where the controller is super speed capable and missing an > ss_phy interrupt. > b) where the GIC is not capable of detecting DP/DM hs phy irq's. > > Power event IRQ stat register indicates whether high speed phy > entered and exited L2 successfully during suspend and resume. > Indicate the same for all ports of multiport. > > Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com> > @@ -471,9 +480,12 @@ static int dwc3_qcom_resume(struct dwc3_qcom *qcom, bool wakeup) > if (ret) > dev_warn(qcom->dev, "failed to enable interconnect: %d\n", ret); > > + for (i = 0; i < qcom->num_ports; i++) { > /* Clear existing events from PHY related to L2 in/out */ This comment either needs to be moved above the for-loop or be indented one level in. > - dwc3_qcom_setbits(qcom->qscratch_base, PWR_EVNT_IRQ_STAT_REG, > - PWR_EVNT_LPM_IN_L2_MASK | PWR_EVNT_LPM_OUT_L2_MASK); > + dwc3_qcom_setbits(qcom->qscratch_base, > + pwr_evnt_irq_stat_reg[i], > + PWR_EVNT_LPM_IN_L2_MASK | PWR_EVNT_LPM_OUT_L2_MASK); > + } > > qcom->is_suspended = false; With the above fixed: Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
On 3/25/2024 6:18 PM, Johan Hovold wrote: > On Thu, Mar 07, 2024 at 11:50:50AM +0530, Krishna Kurapati wrote: >> On multiport supported controllers, each port has its own DP/DM >> and SS (if super speed capable) interrupts. As per the bindings, >> their interrupt names differ from standard ones having "_x" added >> as suffix (x indicates port number). Identify from the interrupt >> names whether the controller is a multiport controller or not. >> Refactor dwc3_qcom_setup_irq() call to parse multiport interrupts >> along with non-multiport ones accordingly.. >> >> Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com> >> --- >> drivers/usb/dwc3/dwc3-qcom.c | 146 +++++++++++++++++++++++++++-------- >> 1 file changed, 112 insertions(+), 34 deletions(-) > > This is much better. Just a couple of nits below. > >> +static int dwc3_qcom_find_num_ports(struct platform_device *pdev) >> +{ >> + const char *irq_name; >> + int port_index; >> + int irq; >> + >> + irq = platform_get_irq_byname_optional(pdev, "qusb2_phy"); >> + if (irq > 0) >> + return 1; >> + >> + irq = platform_get_irq_byname_optional(pdev, "dp_hs_phy_irq"); >> + if (irq > 0) >> + return 1; >> + >> + for (port_index = 0; port_index < DWC3_MAX_PORTS; port_index++) { >> + irq_name = devm_kasprintf(&pdev->dev, GFP_KERNEL, "dp_hs_phy_%d", port_index + 1); > > Please use a stack allocated buffer for these strings as we don't need > them any more after this function returns. > Sure. Will replace this with "char irq_name[14]". >> + if (!irq_name) >> + return -ENOMEM; >> + >> + irq = platform_get_irq_byname_optional(pdev, irq_name); >> + if (irq <= 0) >> + return port_index; >> + } >> + >> + return port_index; > > I think explicitly returning DWC3_MAX_PORTS here would be more readable. > Will make this change. >> +} >> + >> +static int dwc3_qcom_setup_irq(struct platform_device *pdev) >> +{ >> + struct dwc3_qcom *qcom = platform_get_drvdata(pdev); >> + bool is_multiport; >> + int ret; >> + int i; >> + >> + qcom->num_ports = dwc3_qcom_find_num_ports(pdev); >> + if (qcom->num_ports < 0) >> + return -ENOMEM; > > Just return 'ret' directly. Sure, will init ret to -ENOMEM and return ret here. > >> + >> + is_multiport = (qcom->num_ports > 1) ? true : false; > > And no need for the ternary operator: ACK. > > is_multiport = (qcom->num_ports > 1); > >> + >> + for (i = 0; i < qcom->num_ports; i++) { >> + ret = dwc3_qcom_setup_port_irq(pdev, i, is_multiport); >> if (ret) >> return ret; >> - qcom->ss_phy_irq = irq; >> } >> >> return 0; > > With that fixed: > > Reviewed-by: Johan Hovold <johan+linaro@kernel.org> Thanks for the review Regards, Krishna,
On 3/25/2024 6:22 PM, Johan Hovold wrote: > On Thu, Mar 07, 2024 at 11:50:52AM +0530, Krishna Kurapati wrote: >> Power event IRQ is used for wakeup in cases: >> a) where the controller is super speed capable and missing an >> ss_phy interrupt. >> b) where the GIC is not capable of detecting DP/DM hs phy irq's. >> >> Power event IRQ stat register indicates whether high speed phy >> entered and exited L2 successfully during suspend and resume. >> Indicate the same for all ports of multiport. >> >> Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com> > >> @@ -471,9 +480,12 @@ static int dwc3_qcom_resume(struct dwc3_qcom *qcom, bool wakeup) >> if (ret) >> dev_warn(qcom->dev, "failed to enable interconnect: %d\n", ret); >> >> + for (i = 0; i < qcom->num_ports; i++) { >> /* Clear existing events from PHY related to L2 in/out */ > > This comment either needs to be moved above the for-loop or be indented > one level in. Ah yes, missed it. Will move it out of the loop. > >> - dwc3_qcom_setbits(qcom->qscratch_base, PWR_EVNT_IRQ_STAT_REG, >> - PWR_EVNT_LPM_IN_L2_MASK | PWR_EVNT_LPM_OUT_L2_MASK); >> + dwc3_qcom_setbits(qcom->qscratch_base, >> + pwr_evnt_irq_stat_reg[i], >> + PWR_EVNT_LPM_IN_L2_MASK | PWR_EVNT_LPM_OUT_L2_MASK); >> + } >> >> qcom->is_suspended = false; > > With the above fixed: > > Reviewed-by: Johan Hovold <johan+linaro@kernel.org> Thanks for the review. Regards, Krishna,
On Mon, Mar 25, 2024 at 06:45:07PM +0530, Krishna Kurapati PSSNV wrote: > >> +static int dwc3_qcom_setup_irq(struct platform_device *pdev) > >> +{ > >> + struct dwc3_qcom *qcom = platform_get_drvdata(pdev); > >> + bool is_multiport; > >> + int ret; > >> + int i; > >> + > >> + qcom->num_ports = dwc3_qcom_find_num_ports(pdev); > >> + if (qcom->num_ports < 0) > >> + return -ENOMEM; > > > > Just return 'ret' directly. > > Sure, will init ret to -ENOMEM and return ret here. I meant that you should return whatever error dwc3_qcom_find_num_ports() returns, so perhaps something like: ret = dwc3_qcom_find_num_ports(pdev); if (ret < 0) return ret; qcom->num_ports = ret; It looks like dwc3_qcom_find_num_ports() can also return 0 (e.g. on malformed DT), which also needs to be handled somehow. I missed that earlier. Johan
On 3/25/2024 6:53 PM, Johan Hovold wrote: > On Mon, Mar 25, 2024 at 06:45:07PM +0530, Krishna Kurapati PSSNV wrote: >>>> +static int dwc3_qcom_setup_irq(struct platform_device *pdev) >>>> +{ >>>> + struct dwc3_qcom *qcom = platform_get_drvdata(pdev); >>>> + bool is_multiport; >>>> + int ret; >>>> + int i; >>>> + >>>> + qcom->num_ports = dwc3_qcom_find_num_ports(pdev); >>>> + if (qcom->num_ports < 0) >>>> + return -ENOMEM; >>> >>> Just return 'ret' directly. >> >> Sure, will init ret to -ENOMEM and return ret here. > Hi Johan, > I meant that you should return whatever error dwc3_qcom_find_num_ports() > returns, so perhaps something like: > Got it. Any error that might come up in interrupt reading as well. > > ret = dwc3_qcom_find_num_ports(pdev); > if (ret < 0) > return ret; > > qcom->num_ports = ret; > > It looks like dwc3_qcom_find_num_ports() can also return 0 (e.g. on > malformed DT), which also needs to be handled somehow. I missed that > earlier. > From what I remember, Konrad mentioned that we might not need to support incomplete or improper DT [1]. Also since this is close to getting merged, can we take up any changes for Malformed DT handling later given that only one or two devices are present and less likely to be given a malformed DT. [1]: https://lore.kernel.org/all/c8d77d4f-6696-4dc9-8030-daf1d10b114b@linaro.org/ Regards, Krishna,
On Mon, Mar 25, 2024 at 10:59:49PM +0530, Krishna Kurapati PSSNV wrote: > On 3/25/2024 6:53 PM, Johan Hovold wrote: > > ret = dwc3_qcom_find_num_ports(pdev); > > if (ret < 0) > > return ret; > > > > qcom->num_ports = ret; > > > > It looks like dwc3_qcom_find_num_ports() can also return 0 (e.g. on > > malformed DT), which also needs to be handled somehow. I missed that > > earlier. > > From what I remember, Konrad mentioned that we might not need to > support incomplete or improper DT [1]. You still need to make sure that the driver doesn't misbehave on malformed input. > Also since this is close to getting merged, can we take up any changes > for Malformed DT handling later given that only one or two devices are > present and less likely to be given a malformed DT. No, and I'm a bit disappointed that you're still trying to take short cuts to getting these patches merged after I've told you repeatedly that that is not how upstream works. Fortunately, this is easily fixed and would both simplify the code further and reduce the risk of breaking backwards compatibility. Just change the logic in dwc3_qcom_find_num_ports() so that it returns 1 if "dp_hs_phy_1" is missing, and otherwise you determine the number of ports by iterating from 2 to DWC3_MAX_PORTS - 1. Johan
On 3/26/2024 1:15 AM, Johan Hovold wrote: > Just change the logic in dwc3_qcom_find_num_ports() so that it returns 1 > if "dp_hs_phy_1" is missing, and otherwise you determine the number of > ports by iterating from 2 to DWC3_MAX_PORTS - 1. Hi Johan, I made this change and it works. Removed any return value check for the find_num_ports call as it can return only 1/2/3/4 now. --- irq = platform_get_irq_byname_optional(pdev, "qusb2_phy"); if (irq > 0) return 1; irq = platform_get_irq_byname_optional(pdev, "dp_hs_phy_irq"); if (irq > 0) return 1; irq = platform_get_irq_byname_optional(pdev, "dp_hs_phy_1"); if (irq <= 0) return 1; for (port_index = 1; port_index < DWC3_MAX_PORTS - 1; port_index++) { sprintf(irq_name, "dp_hs_phy_%d", port_index + 1); irq = platform_get_irq_byname_optional(pdev, irq_name); if (irq <= 0) return port_index; } return DWC3_MAX_PORTS; --- Let me know if this is fine and I can push out v17. Regards, Krishna,
On Tue, Mar 26, 2024 at 01:41:52PM +0530, Krishna Kurapati PSSNV wrote: > On 3/26/2024 1:15 AM, Johan Hovold wrote: > > > Just change the logic in dwc3_qcom_find_num_ports() so that it returns 1 > > if "dp_hs_phy_1" is missing, and otherwise you determine the number of > > ports by iterating from 2 to DWC3_MAX_PORTS - 1. > I made this change and it works. Removed any return value check for the > find_num_ports call as it can return only 1/2/3/4 now. > > --- > irq = platform_get_irq_byname_optional(pdev, "qusb2_phy"); > if (irq > 0) > return 1; > > irq = platform_get_irq_byname_optional(pdev, "dp_hs_phy_irq"); > if (irq > 0) > return 1; As I mentioned above, these two lookups are no longer needed and should be removed. > irq = platform_get_irq_byname_optional(pdev, "dp_hs_phy_1"); > if (irq <= 0) > return 1; Just assume it's a single port controller unless "dp_hs_phy_1" is present. > for (port_index = 1; port_index < DWC3_MAX_PORTS - 1; > port_index++) { I think this would be more readable if you use port (num) as iterator (2..DWC3_MAX_PORTS) as you're returning a number of ports. > sprintf(irq_name, "dp_hs_phy_%d", port_index + 1); Then this would use just "port"; > > irq = platform_get_irq_byname_optional(pdev, irq_name); > if (irq <= 0) > return port_index; And return "port - 1" here. > } > > return DWC3_MAX_PORTS; Johan
On 3/26/2024 1:58 PM, Johan Hovold wrote: > On Tue, Mar 26, 2024 at 01:41:52PM +0530, Krishna Kurapati PSSNV wrote: >> On 3/26/2024 1:15 AM, Johan Hovold wrote: >> >>> Just change the logic in dwc3_qcom_find_num_ports() so that it returns 1 >>> if "dp_hs_phy_1" is missing, and otherwise you determine the number of >>> ports by iterating from 2 to DWC3_MAX_PORTS - 1. > >> I made this change and it works. Removed any return value check for the >> find_num_ports call as it can return only 1/2/3/4 now. >> >> --- >> irq = platform_get_irq_byname_optional(pdev, "qusb2_phy"); >> if (irq > 0) >> return 1; >> >> irq = platform_get_irq_byname_optional(pdev, "dp_hs_phy_irq"); >> if (irq > 0) >> return 1; > > As I mentioned above, these two lookups are no longer needed and should > be removed. > >> irq = platform_get_irq_byname_optional(pdev, "dp_hs_phy_1"); >> if (irq <= 0) >> return 1; > > Just assume it's a single port controller unless "dp_hs_phy_1" is > present. > >> for (port_index = 1; port_index < DWC3_MAX_PORTS - 1; >> port_index++) { > > I think this would be more readable if you use port (num) as iterator > (2..DWC3_MAX_PORTS) as you're returning a number of ports. > >> sprintf(irq_name, "dp_hs_phy_%d", port_index + 1); > > Then this would use just "port"; > >> >> irq = platform_get_irq_byname_optional(pdev, irq_name); >> if (irq <= 0) >> return port_index; > > And return "port - 1" here. > >> } >> >> return DWC3_MAX_PORTS; Ok,Thanks. With the above fixed, can I put your RB tag given on [1] or will you be providing another one. Asking because when you gave that RB, this malformed DT thing wasn't there. So wanted to check with you before pushing the code. [1]: https://lore.kernel.org/all/ZgFyukBXIIwZo7v-@hovoldconsulting.com/ Regards, Krishna,
On Tue, Mar 26, 2024 at 03:36:27PM +0530, Krishna Kurapati PSSNV wrote: > On 3/26/2024 1:58 PM, Johan Hovold wrote: > > On Tue, Mar 26, 2024 at 01:41:52PM +0530, Krishna Kurapati PSSNV wrote: > >> On 3/26/2024 1:15 AM, Johan Hovold wrote: > >> > >>> Just change the logic in dwc3_qcom_find_num_ports() so that it returns 1 > >>> if "dp_hs_phy_1" is missing, and otherwise you determine the number of > >>> ports by iterating from 2 to DWC3_MAX_PORTS - 1. > With the above fixed, can I put your RB tag given on [1] or will you be > providing another one. Asking because when you gave that RB, this > malformed DT thing wasn't there. So wanted to check with you before > pushing the code. Thanks for checking. I'm fine with you adding the tag after fixing the above now that we've discussed it. Johan