[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <db172ecd-879b-b35b-2f02-e63dd9c956e2@i2se.com>
Date: Wed, 3 May 2023 08:06:33 +0200
From: Stefan Wahren <stefan.wahren@...e.com>
To: Jim Quinlan <jim2101024@...il.com>, linux-pci@...r.kernel.org,
Nicolas Saenz Julienne <nsaenz@...nel.org>,
Bjorn Helgaas <bhelgaas@...gle.com>,
Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
Cyril Brulebois <kibi@...ian.org>,
Phil Elwell <phil@...pberrypi.com>,
bcm-kernel-feedback-list@...adcom.com, james.quinlan@...adcom.com
Cc: Florian Fainelli <f.fainelli@...il.com>,
Lorenzo Pieralisi <lpieralisi@...nel.org>,
Krzysztof WilczyĆski <kw@...ux.com>,
Rob Herring <robh@...nel.org>,
"moderated list:BROADCOM BCM2711/BCM2835 ARM ARCHITECTURE"
<linux-rpi-kernel@...ts.infradead.org>,
"moderated list:BROADCOM BCM2711/BCM2835 ARM ARCHITECTURE"
<linux-arm-kernel@...ts.infradead.org>,
open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v4 3/5] PCI: brcmstb: Set PCIe transaction completion
timeout
Hi Jim,
Am 29.04.23 um 00:34 schrieb Jim Quinlan:
> Since the STB PCIe HW will cause a CPU abort on a PCIe transaction
> completion timeout abort, we might as well extend the default timeout
> limit. Further, different devices and systems may requires a larger or
> smaller amount commensurate with their L1SS exit time, so the property
> "brcm,completion-timeout-us" may be used to set a custom timeout value.
>
> Tested-by: Florian Fainelli <f.fainelli@...il.com>
> Signed-off-by: Jim Quinlan <jim2101024@...il.com>
> ---
> drivers/pci/controller/pcie-brcmstb.c | 30 +++++++++++++++++++++++++++
> 1 file changed, 30 insertions(+)
>
> diff --git a/drivers/pci/controller/pcie-brcmstb.c b/drivers/pci/controller/pcie-brcmstb.c
> index c4b076ea5180..c2cb683447ac 100644
> --- a/drivers/pci/controller/pcie-brcmstb.c
> +++ b/drivers/pci/controller/pcie-brcmstb.c
> @@ -1080,6 +1080,35 @@ static void brcm_config_clkreq(struct brcm_pcie *pcie)
> writel(clkreq_set, pcie->base + PCIE_MISC_HARD_PCIE_HARD_DEBUG);
> }
>
> +static void brcm_config_completion_timeout(struct brcm_pcie *pcie)
> +{
> + /* TIMEOUT register is two registers before RGR1_SW_INIT_1 */
> + const char *fmt = "brcm,completion-timeout-us clamped to region [%u..%u]\n";
> + const unsigned int REG_OFFSET = PCIE_RGR1_SW_INIT_1(pcie) - 8;
> + const u32 timeout_us_min = 16;
> + const u32 timeout_us_max = 19884107;
> + u32 timeout_us = 1000000; /* Our default, 1 second */
> + int rval, ret;
> +
> + ret = of_property_read_u32(pcie->np, "brcm,completion-timeout-us",
> + &timeout_us);
> + if (ret && ret != -EINVAL)
> + dev_err(pcie->dev, "malformed/invalid 'brcm,completion-timeout-us'\n");
> +
> + /* If needed, clamp the requested timeout value and issue a warning */
> + if (timeout_us < timeout_us_min) {
> + timeout_us = timeout_us_min;
> + dev_warn(pcie->dev, fmt, timeout_us_min, timeout_us_max);
> + } else if (timeout_us > timeout_us_max) {
> + timeout_us = timeout_us_max;
> + dev_warn(pcie->dev, fmt, timeout_us_min, timeout_us_max);
> + }
> +
> + /* Each unit in timeout register is 1/216,000,000 seconds */
> + rval = 216 * timeout_us;
> + writel(rval, pcie->base + REG_OFFSET);
i don't think "int" is the proper type for rval here
> +}
> +
> static int brcm_pcie_start_link(struct brcm_pcie *pcie)
> {
> struct device *dev = pcie->dev;
> @@ -1110,6 +1139,7 @@ static int brcm_pcie_start_link(struct brcm_pcie *pcie)
> return -ENODEV;
> }
>
> + brcm_config_completion_timeout(pcie);
> brcm_config_clkreq(pcie);
>
> if (pcie->gen)
Powered by blists - more mailing lists