| Message ID | 20221010025912.1042437-1-koba.ko@canonical.com |
|---|---|
| Headers | show |
| Series | Can only reach PC3 when ethernet is plugged r8169 | expand |
Koba Ko kirjoitti 10.10.2022 klo 5.59: > BugLink: https://bugs.launchpad.net/bugs/1946433 > > [Impact] > System only can reach PC3, and it affects power consumption alot. > > [Fix] > Kaiheng implemented a dynamic ASPM for r8169, it not only fixes the PC state issue, but also fixes network speed issue. > V7: > https://patchwork.kernel.org/project/netdevbpf/patch/20211016075442.650311-5-kai.heng.feng@canonical.com/ > V6: > https://patchwork.ozlabs.org/project/linux-pci/cover/20211007161552.272771-1-kai.heng.feng@canonical.com/ > > [Test] > Verified on 2 different systems which has PC state issue and has network speed issue, these patches fix both issues. > > [Where problems could occur] > It toggles ASPM on and off depends on the network traffic during runtime, I don't think it'll lead to any regressions. Some potential issues have been addressed during the patch submitting. It's v6 now and accepted by upstream. > > > Kai-Heng Feng (3): > [OEM-6.0/OEM-5.17] r8169: Enable chip-specific ASPM regardless of > PCIe ASPM status > [OEM-6.0/OEM-5.17] r8169: Use mutex to guard config register > locking > [OEM-6.0/OEM-5.17] r8169: Implement dynamic ASPM mechanism > > drivers/net/ethernet/realtek/r8169_main.c | 77 ++++++++++++++++++----- > drivers/pci/pcie/aspm.c | 1 + > 2 files changed, 62 insertions(+), 16 deletions(-) > These are not upstream yet (why?), so should be marked as SAUCE.
On Mon, Oct 10, 2022 at 4:31 PM Timo Aaltonen <tjaalton@ubuntu.com> wrote: > > Koba Ko kirjoitti 10.10.2022 klo 5.59: > > BugLink: https://bugs.launchpad.net/bugs/1946433 > > > > [Impact] > > System only can reach PC3, and it affects power consumption alot. > > > > [Fix] > > Kaiheng implemented a dynamic ASPM for r8169, it not only fixes the PC state issue, but also fixes network speed issue. > > V7: > > https://patchwork.kernel.org/project/netdevbpf/patch/20211016075442.650311-5-kai.heng.feng@canonical.com/ > > V6: > > https://patchwork.ozlabs.org/project/linux-pci/cover/20211007161552.272771-1-kai.heng.feng@canonical.com/ > > > > [Test] > > Verified on 2 different systems which has PC state issue and has network speed issue, these patches fix both issues. > > > > [Where problems could occur] > > It toggles ASPM on and off depends on the network traffic during runtime, I don't think it'll lead to any regressions. Some potential issues have been addressed during the patch submitting. It's v6 now and accepted by upstream. > > > > > > Kai-Heng Feng (3): > > [OEM-6.0/OEM-5.17] r8169: Enable chip-specific ASPM regardless of > > PCIe ASPM status > > [OEM-6.0/OEM-5.17] r8169: Use mutex to guard config register > > locking > > [OEM-6.0/OEM-5.17] r8169: Implement dynamic ASPM mechanism > > > > drivers/net/ethernet/realtek/r8169_main.c | 77 ++++++++++++++++++----- > > drivers/pci/pcie/aspm.c | 1 + > > 2 files changed, 62 insertions(+), 16 deletions(-) > > > > These are not upstream yet (why?), so should be marked as SAUCE. As per Kai-Heng, these don't get merged. I'm confused, the same patches that have been applied on J/OEM-5.14 don't be marked as SAUCE. will send v2 and marked as SAUCE. > > > -- > t >
Koba Ko kirjoitti 10.10.2022 klo 13.33: > On Mon, Oct 10, 2022 at 4:31 PM Timo Aaltonen <tjaalton@ubuntu.com> wrote: >> >> Koba Ko kirjoitti 10.10.2022 klo 5.59: >>> BugLink: https://bugs.launchpad.net/bugs/1946433 >>> >>> [Impact] >>> System only can reach PC3, and it affects power consumption alot. >>> >>> [Fix] >>> Kaiheng implemented a dynamic ASPM for r8169, it not only fixes the PC state issue, but also fixes network speed issue. >>> V7: >>> https://patchwork.kernel.org/project/netdevbpf/patch/20211016075442.650311-5-kai.heng.feng@canonical.com/ >>> V6: >>> https://patchwork.ozlabs.org/project/linux-pci/cover/20211007161552.272771-1-kai.heng.feng@canonical.com/ >>> >>> [Test] >>> Verified on 2 different systems which has PC state issue and has network speed issue, these patches fix both issues. >>> >>> [Where problems could occur] >>> It toggles ASPM on and off depends on the network traffic during runtime, I don't think it'll lead to any regressions. Some potential issues have been addressed during the patch submitting. It's v6 now and accepted by upstream. >>> >>> >>> Kai-Heng Feng (3): >>> [OEM-6.0/OEM-5.17] r8169: Enable chip-specific ASPM regardless of >>> PCIe ASPM status >>> [OEM-6.0/OEM-5.17] r8169: Use mutex to guard config register >>> locking >>> [OEM-6.0/OEM-5.17] r8169: Implement dynamic ASPM mechanism >>> >>> drivers/net/ethernet/realtek/r8169_main.c | 77 ++++++++++++++++++----- >>> drivers/pci/pcie/aspm.c | 1 + >>> 2 files changed, 62 insertions(+), 16 deletions(-) >>> >> >> These are not upstream yet (why?), so should be marked as SAUCE. > > As per Kai-Heng, these don't get merged. > I'm confused, the same patches that have been applied on J/OEM-5.14 > don't be marked as SAUCE. True, my bad. And they also got applied to unstable which is why they're also in jammy and up. > will send v2 and marked as SAUCE. thanks!
Koba Ko kirjoitti 10.10.2022 klo 5.59: > BugLink: https://bugs.launchpad.net/bugs/1946433 > > [Impact] > System only can reach PC3, and it affects power consumption alot. > > [Fix] > Kaiheng implemented a dynamic ASPM for r8169, it not only fixes the PC state issue, but also fixes network speed issue. > V7: > https://patchwork.kernel.org/project/netdevbpf/patch/20211016075442.650311-5-kai.heng.feng@canonical.com/ > V6: > https://patchwork.ozlabs.org/project/linux-pci/cover/20211007161552.272771-1-kai.heng.feng@canonical.com/ > > [Test] > Verified on 2 different systems which has PC state issue and has network speed issue, these patches fix both issues. > > [Where problems could occur] > It toggles ASPM on and off depends on the network traffic during runtime, I don't think it'll lead to any regressions. Some potential issues have been addressed during the patch submitting. It's v6 now and accepted by upstream. > > > Kai-Heng Feng (3): > [OEM-6.0/OEM-5.17] r8169: Enable chip-specific ASPM regardless of > PCIe ASPM status > [OEM-6.0/OEM-5.17] r8169: Use mutex to guard config register > locking > [OEM-6.0/OEM-5.17] r8169: Implement dynamic ASPM mechanism > > drivers/net/ethernet/realtek/r8169_main.c | 77 ++++++++++++++++++----- > drivers/pci/pcie/aspm.c | 1 + > 2 files changed, 62 insertions(+), 16 deletions(-) > there's a newer version
BugLink: https://bugs.launchpad.net/bugs/1946433 [Impact] System only can reach PC3, and it affects power consumption alot. [Fix] Kaiheng implemented a dynamic ASPM for r8169, it not only fixes the PC state issue, but also fixes network speed issue. V7: https://patchwork.kernel.org/project/netdevbpf/patch/20211016075442.650311-5-kai.heng.feng@canonical.com/ V6: https://patchwork.ozlabs.org/project/linux-pci/cover/20211007161552.272771-1-kai.heng.feng@canonical.com/ [Test] Verified on 2 different systems which has PC state issue and has network speed issue, these patches fix both issues. [Where problems could occur] It toggles ASPM on and off depends on the network traffic during runtime, I don't think it'll lead to any regressions. Some potential issues have been addressed during the patch submitting. It's v6 now and accepted by upstream. Kai-Heng Feng (3): [OEM-6.0/OEM-5.17] r8169: Enable chip-specific ASPM regardless of PCIe ASPM status [OEM-6.0/OEM-5.17] r8169: Use mutex to guard config register locking [OEM-6.0/OEM-5.17] r8169: Implement dynamic ASPM mechanism drivers/net/ethernet/realtek/r8169_main.c | 77 ++++++++++++++++++----- drivers/pci/pcie/aspm.c | 1 + 2 files changed, 62 insertions(+), 16 deletions(-)