linux-kernel - Re: [RFC] [PATCH net-next v5 2/3] r8169: Use PCIe ASPM status for NIC ASPM enablement

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 1 Oct 2021 12:32:18 +0800
From:   Kai-Heng Feng <kai.heng.feng@...onical.com>
To:     Bjorn Helgaas <helgaas@...nel.org>
Cc:     Heiner Kallweit <hkallweit1@...il.com>,
        nic_swsd <nic_swsd@...ltek.com>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        David Miller <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Anthony Wong <anthony.wong@...onical.com>,
        Linux Netdev List <netdev@...r.kernel.org>,
        Linux PCI <linux-pci@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [RFC] [PATCH net-next v5 2/3] r8169: Use PCIe ASPM status for NIC
 ASPM enablement

On Fri, Sep 17, 2021 at 11:26 PM Bjorn Helgaas <helgaas@...nel.org> wrote:
>
> On Fri, Sep 17, 2021 at 12:09:08PM +0800, Kai-Heng Feng wrote:
> > On Fri, Sep 17, 2021 at 1:07 AM Bjorn Helgaas <helgaas@...nel.org> wrote:
> > > On Thu, Sep 16, 2021 at 11:44:16PM +0800, Kai-Heng Feng wrote:
> > > > Because ASPM control may not be granted by BIOS while ASPM is enabled,
> > > > and ASPM can be enabled via sysfs, so use pcie_aspm_enabled() directly
> > > > to check current ASPM enable status.
> > > >
> > > > Signed-off-by: Kai-Heng Feng <kai.heng.feng@...onical.com>
> > > > ---
> > > > v5:
> > > >  - New patch.
> > > >
> > > >  drivers/net/ethernet/realtek/r8169_main.c | 13 ++++++++-----
> > > >  1 file changed, 8 insertions(+), 5 deletions(-)
> > > >
> > > > diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c
> > > > index 0199914440abc..6f1a9bec40c05 100644
> > > > --- a/drivers/net/ethernet/realtek/r8169_main.c
> > > > +++ b/drivers/net/ethernet/realtek/r8169_main.c
> > > > @@ -622,7 +622,6 @@ struct rtl8169_private {
> > > >       } wk;
> > > >
> > > >       unsigned supports_gmii:1;
> > > > -     unsigned aspm_manageable:1;
> > > >       dma_addr_t counters_phys_addr;
> > > >       struct rtl8169_counters *counters;
> > > >       struct rtl8169_tc_offsets tc_offset;
> > > > @@ -2664,8 +2663,13 @@ static void rtl_enable_exit_l1(struct rtl8169_private *tp)
> > > >
> > > >  static void rtl_hw_aspm_clkreq_enable(struct rtl8169_private *tp, bool enable)
> > > >  {
> > > > -     /* Don't enable ASPM in the chip if OS can't control ASPM */
> > > > -     if (enable && tp->aspm_manageable) {
> > > > +     struct pci_dev *pdev = tp->pci_dev;
> > > > +
> > > > +     /* Don't enable ASPM in the chip if PCIe ASPM isn't enabled */
> > > > +     if (!pcie_aspm_enabled(pdev) && enable)
> > > > +             return;
> > >
> > > What happens when the user enables or disables ASPM via sysfs (see
> > > https://git.kernel.org/linus/72ea91afbfb0)?
> > >
> > > The driver is not going to know about that change.
> >
> > So it's still better to fold this patch into next one? So the periodic
> > delayed_work can toggle ASPM accordingly.
>
> No, my point is that the user can enable/disable ASPM via sysfs, and
> the driver will not know anything about it.  There's no callback that
> tells the driver when this happens.
>
> My question is whether this code works when that happens.  I doubt it
> works, because if ASPM is not enabled at this moment, you return
> without doing enabling ASPM in the chip below.
>
> If the user subsequently enables ASPM via sysfs, the chip setup below
> will not be done.
>
> If there's chip-specific setup to make ASPM work, I think the
> chip-specific part needs to be done unconditionally.

So it's either adding a callback to notify driver about ASPM change,
or doing chip-specific ASPM unconditionally.
Which one do you prefer?

Kai-Heng

>
> > > > +     if (enable) {
> > > >               RTL_W8(tp, Config5, RTL_R8(tp, Config5) | ASPM_en);
> > > >               RTL_W8(tp, Config2, RTL_R8(tp, Config2) | ClkReqEn);
> > > >       } else {
> > > > @@ -5272,8 +5276,7 @@ static int rtl_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
> > > >       /* Disable ASPM L1 as that cause random device stop working
> > > >        * problems as well as full system hangs for some PCIe devices users.
> > > >        */
> > > > -     rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L1);
> > > > -     tp->aspm_manageable = !rc;
> > > > +     pci_disable_link_state(pdev, PCIE_LINK_STATE_L1);
> > > >
> > > >       /* enable device (incl. PCI PM wakeup and hotplug setup) */
> > > >       rc = pcim_enable_device(pdev);
> > > > --
> > > > 2.32.0
> > > >