| Message ID | 1355279525.7216.154.camel@yhuang-dev (mailing list archive) |
|---|---|
| State | New, archived |
| Delegated to: | Bjorn Helgaas |
| Headers | show |
On Tue, Dec 11, 2012 at 6:32 PM, Huang Ying <ying.huang@intel.com> wrote: > On Fri, 2012-12-07 at 22:00 +0100, Rafael J. Wysocki wrote: >> On Thursday, December 06, 2012 04:28:08 PM Sarah Sharp wrote: >> > On Thu, Dec 06, 2012 at 01:43:32AM +0100, Rafael J. Wysocki wrote: >> > > On Wednesday, December 05, 2012 04:33:44 PM Sarah Sharp wrote: >> > > > Wakeup from D3 works fine on the 3.5.0 kernel, but fails on 3.6.2. I >> > > > haven't fully bisected yet. >> > > > >> > > > In debugging, I found that if you only enable runtime suspend for the >> > > > NEC host controller, the host successfully comes out of D3 when you plug >> > > > in a USB device. However, if you enable runtime PM for the parent PCIe root >> > > > port, it stops working. Disabling D3cold for both devices did not help. >> > > > >> > > > It looks like a PCI issue, so what sort of debugging info do you need >> > > > from me? >> > > >> > > It looks like this is related to one of the following commits: >> > >> > > Generally, please try to bisect changes in drivers/pci between v3.5 and v3.6. >> > >> > Ok, I ran git bisect with only the drivers/pci directory as a target. >> > >> > > ee85f54 ACPI/PM: specify lowest allowed state for device sleep state >> > >> > git bisect ended up identifying this as the bad patch, although >> > reverting just that patch after the bisect finished didn't seem to help. >> > However, it does make sense that this would be the culprit patch, if >> > Huang Ying's theory about the PME polling is correct. >> >> Well, we surely don't handle this particular case correctly, and it shouldn't >> be very difficult to fix, so let's hope that this is the culprit. :-) >> > > Please try the patch attached. With it, USB host controller can be > waken up with USB card reader plugging in on my test machine. Potentially silly question: is this better than waking up the PCIe port when polling the device (and putting it back to sleep again)? --Andy > > Best Regards, > Huang Ying > -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Dec 12, 2012 at 10:34 AM, Andrew Lutomirski <luto@mit.edu> wrote: > On Tue, Dec 11, 2012 at 6:32 PM, Huang Ying <ying.huang@intel.com> wrote: >> On Fri, 2012-12-07 at 22:00 +0100, Rafael J. Wysocki wrote: >>> On Thursday, December 06, 2012 04:28:08 PM Sarah Sharp wrote: >>> > On Thu, Dec 06, 2012 at 01:43:32AM +0100, Rafael J. Wysocki wrote: >>> > > On Wednesday, December 05, 2012 04:33:44 PM Sarah Sharp wrote: >>> > > > Wakeup from D3 works fine on the 3.5.0 kernel, but fails on 3.6.2. I >>> > > > haven't fully bisected yet. >>> > > > >>> > > > In debugging, I found that if you only enable runtime suspend for the >>> > > > NEC host controller, the host successfully comes out of D3 when you plug >>> > > > in a USB device. However, if you enable runtime PM for the parent PCIe root >>> > > > port, it stops working. Disabling D3cold for both devices did not help. >>> > > > >>> > > > It looks like a PCI issue, so what sort of debugging info do you need >>> > > > from me? >>> > > >>> > > It looks like this is related to one of the following commits: >>> > >>> > > Generally, please try to bisect changes in drivers/pci between v3.5 and v3.6. >>> > >>> > Ok, I ran git bisect with only the drivers/pci directory as a target. >>> > >>> > > ee85f54 ACPI/PM: specify lowest allowed state for device sleep state >>> > >>> > git bisect ended up identifying this as the bad patch, although >>> > reverting just that patch after the bisect finished didn't seem to help. >>> > However, it does make sense that this would be the culprit patch, if >>> > Huang Ying's theory about the PME polling is correct. >>> >>> Well, we surely don't handle this particular case correctly, and it shouldn't >>> be very difficult to fix, so let's hope that this is the culprit. :-) >>> >> >> Please try the patch attached. With it, USB host controller can be >> waken up with USB card reader plugging in on my test machine. > > Potentially silly question: is this better than waking up the PCIe > port when polling the device (and putting it back to sleep again)? Put PCI devices into/out of low power state can be very time consuming. For D3hot, it is at least 20ms, and for D3cold, it is at least 200ms. Best Regards, Huang Ying -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Dec 12, 2012 at 10:32:05AM +0800, Huang Ying wrote: > Please try the patch attached. With it, USB host controller can be > waken up with USB card reader plugging in on my test machine. This patch works for me. When I enable runtime PM for both the root port and the host controller, and I see the xHCI host go into PCI runtime suspend, but there's no longer a message about the power state changing to D3cold. USB device connects while the PCI host is suspended work now, as well as remote wakeup. Tested-by: Sarah Sharp <sarah.a.sharp@linux.intel.com> Please mark this for stable. Thanks, Sarah Sharp -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Subject: [BUGFIX] PCIe/PM: Do not suspend port if any subordinate device need PME polling In Ulrich reported that his USB3 cardreader does not work reliably when connected to the USB3 port. It turns out that USB3 controller failed to be waken up when plugging in the USB3 cardreader. Further experiment found that the USB3 host controller can only be waken up via polling, while not via PME interrupt. But if the PCIe port the USB3 host controller connected is suspended, we can not poll the USB3 host controller because its config space is not accessible if the PCIe port is put into low power state. To solve the issue, the PCIe port will not be suspended if any subordinate device need PME polling. Reported-by: Ulrich Eckhardt <usb@uli-eckhardt.de> Signed-off-by: Huang Ying <ying.huang@intel.com> --- drivers/pci/pcie/portdrv_pci.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) --- a/drivers/pci/pcie/portdrv_pci.c +++ b/drivers/pci/pcie/portdrv_pci.c @@ -134,10 +134,26 @@ static int pcie_port_runtime_resume(stru return 0; } +static int pci_dev_pme_poll(struct pci_dev *pdev, void *data) +{ + int *pme_poll = data; + *pme_poll = *pme_poll || pdev->pme_poll; + return 0; +} + static int pcie_port_runtime_idle(struct device *dev) { + struct pci_dev *pdev = to_pci_dev(dev); + int pme_poll = false; + + /* + * If any subordinate device needs pme poll, we should keep + * the port in D0, because we need port in D0 to poll it. + */ + pci_walk_bus(pdev->subordinate, pci_dev_pme_poll, &pme_poll); /* Delay for a short while to prevent too frequent suspend/resume */ - pm_schedule_suspend(dev, 10); + if (!pme_poll) + pm_schedule_suspend(dev, 10); return -EBUSY; } #else