netdev - Re: [PATCH v3 wireless-drivers 1/2] mt76: mt76x2e: disable pcie

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Sun, 27 Oct 2019 22:40:30 +0100
From:   Oleksandr Natalenko <oleksandr@...alenko.name>
To:     Lorenzo Bianconi <lorenzo@...nel.org>
Cc:     kvalo@...eaurora.org, linux-wireless@...r.kernel.org, nbd@....name,
        hkallweit1@...il.com, sgruszka@...hat.com,
        lorenzo.bianconi@...hat.com, netdev@...r.kernel.org
Subject: Re: [PATCH v3 wireless-drivers 1/2] mt76: mt76x2e: disable pcie_aspm
 by default

Hi.

On 27.10.2019 20:53, Lorenzo Bianconi wrote:
> On same device (e.g. U7612E-H1) PCIE_ASPM causes continuous mcu hangs 
> and
> instability. Since mt76x2 series does not manage PCIE PS states, first 
> we
> try to disable ASPM using pci_disable_link_state. If it fails, we will
> disable PCIE PS configuring PCI registers.
> This patch has been successfully tested on U7612E-H1 mini-pice card
> 
> Tested-by: Oleksandr Natalenko <oleksandr@...alenko.name>

For this revision, not yet ;) (see below).

> Signed-off-by: Felix Fietkau <nbd@....name>
> Signed-off-by: Lorenzo Bianconi <lorenzo@...nel.org>
> ---
>  drivers/net/wireless/mediatek/mt76/Makefile   |  2 +
>  drivers/net/wireless/mediatek/mt76/mt76.h     |  1 +
>  .../net/wireless/mediatek/mt76/mt76x2/pci.c   |  2 +
>  drivers/net/wireless/mediatek/mt76/pci.c      | 46 +++++++++++++++++++
>  4 files changed, 51 insertions(+)
>  create mode 100644 drivers/net/wireless/mediatek/mt76/pci.c
> 
> diff --git a/drivers/net/wireless/mediatek/mt76/Makefile
> b/drivers/net/wireless/mediatek/mt76/Makefile
> index 4d03596e891f..d7a1ddc9e407 100644
> --- a/drivers/net/wireless/mediatek/mt76/Makefile
> +++ b/drivers/net/wireless/mediatek/mt76/Makefile
> @@ -8,6 +8,8 @@ mt76-y := \
>  	mmio.o util.o trace.o dma.o mac80211.o debugfs.o eeprom.o \
>  	tx.o agg-rx.o mcu.o
> 
> +mt76-$(CONFIG_PCI) += pci.o
> +
>  mt76-usb-y := usb.o usb_trace.o
> 
>  CFLAGS_trace.o := -I$(src)
> diff --git a/drivers/net/wireless/mediatek/mt76/mt76.h
> b/drivers/net/wireless/mediatek/mt76/mt76.h
> index 570c159515a0..dc468ed9434a 100644
> --- a/drivers/net/wireless/mediatek/mt76/mt76.h
> +++ b/drivers/net/wireless/mediatek/mt76/mt76.h
> @@ -578,6 +578,7 @@ bool __mt76_poll_msec(struct mt76_dev *dev, u32
> offset, u32 mask, u32 val,
>  #define mt76_poll_msec(dev, ...) __mt76_poll_msec(&((dev)->mt76), 
> __VA_ARGS__)
> 
>  void mt76_mmio_init(struct mt76_dev *dev, void __iomem *regs);
> +void mt76_pci_disable_aspm(struct pci_dev *pdev);
> 
>  static inline u16 mt76_chip(struct mt76_dev *dev)
>  {
> diff --git a/drivers/net/wireless/mediatek/mt76/mt76x2/pci.c
> b/drivers/net/wireless/mediatek/mt76/mt76x2/pci.c
> index 73c3104f8858..cf611d1b817c 100644
> --- a/drivers/net/wireless/mediatek/mt76/mt76x2/pci.c
> +++ b/drivers/net/wireless/mediatek/mt76/mt76x2/pci.c
> @@ -81,6 +81,8 @@ mt76pci_probe(struct pci_dev *pdev, const struct
> pci_device_id *id)
>  	/* RG_SSUSB_CDR_BR_PE1D = 0x3 */
>  	mt76_rmw_field(dev, 0x15c58, 0x3 << 6, 0x3);
> 
> +	mt76_pci_disable_aspm(pdev);
> +
>  	return 0;
> 
>  error:
> diff --git a/drivers/net/wireless/mediatek/mt76/pci.c
> b/drivers/net/wireless/mediatek/mt76/pci.c
> new file mode 100644
> index 000000000000..04c5a692bc85
> --- /dev/null
> +++ b/drivers/net/wireless/mediatek/mt76/pci.c
> @@ -0,0 +1,46 @@
> +// SPDX-License-Identifier: ISC
> +/*
> + * Copyright (C) 2019 Lorenzo Bianconi <lorenzo@...nel.org>
> + */
> +
> +#include <linux/pci.h>

FYI, I had to #include <linux/pci-aspm.h> on 5.3 kernel because this is 
where pci_disable_link_state() is defined. It is not needed on 5.4+ 
since the declaration got moved to pci.h; this is just a note for those 
who are going to test these changes on 5.3.

I'm still building the kernel, though. Will get back with the results 
later.

> +
> +void mt76_pci_disable_aspm(struct pci_dev *pdev)
> +{
> +	struct pci_dev *parent = pdev->bus->self;
> +	u16 aspm_conf, parent_aspm_conf = 0;
> +
> +	pcie_capability_read_word(pdev, PCI_EXP_LNKCTL, &aspm_conf);
> +	aspm_conf &= PCI_EXP_LNKCTL_ASPMC;
> +	if (parent) {
> +		pcie_capability_read_word(parent, PCI_EXP_LNKCTL,
> +					  &parent_aspm_conf);
> +		parent_aspm_conf &= PCI_EXP_LNKCTL_ASPMC;
> +	}
> +
> +	if (!aspm_conf && (!parent || !parent_aspm_conf)) {
> +		/* aspm already disabled */
> +		return;
> +	}
> +
> +	dev_info(&pdev->dev, "disabling ASPM %s %s\n",
> +		 (aspm_conf & PCI_EXP_LNKCTL_ASPM_L0S) ? "L0s" : "",
> +		 (aspm_conf & PCI_EXP_LNKCTL_ASPM_L1) ? "L1" : "");
> +
> +	if (IS_ENABLED(CONFIG_PCIEASPM)) {
> +		int err;
> +
> +		err = pci_disable_link_state(pdev, aspm_conf);
> +		if (!err)
> +			return;
> +	}
> +
> +	/* both device and parent should have the same ASPM setting.
> +	 * disable ASPM in downstream component first and then upstream.
> +	 */
> +	pcie_capability_clear_word(pdev, PCI_EXP_LNKCTL, aspm_conf);
> +	if (parent)
> +		pcie_capability_clear_word(parent, PCI_EXP_LNKCTL,
> +					   aspm_conf);
> +}
> +EXPORT_SYMBOL_GPL(mt76_pci_disable_aspm);

-- 
   Oleksandr Natalenko (post-factum)