[PATCH] brcmfmac: fix regression in parsing NVRAM for multiple devices

[PATCH] brcmfmac: fix regression in parsing NVRAM for multiple devices

[Rafał Miłecki]

From: Rafał Miłecki <rafal@milecki.pl>

NVRAM is designed to work with Broadcom's SDK Linux kernel which fakes
PCI domain 0 for all internal MMIO devices. Since official Linux kernel
uses platform devices for that purpose there is a mismatch in numbering
PCI domains.

There used to be a fix for that problem but it was accidentally dropped
during the last firmware loading rework. That resulted in brcmfmac not
being able to extract device specific NVRAM content and all kind of
calibration problems.

Reported-by: Aditya Xavier <adityaxavier@gmail.com>
Fixes: 2baa3aaee27f ("brcmfmac: introduce brcmf_fw_alloc_request() function")
Cc: stable@vger.kernel.org # v4.17+
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c
index 45928b5b8d97..4fffa6988087 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c
@@ -1785,7 +1785,8 @@ brcmf_pcie_prepare_fw_request(struct brcmf_pciedev_info *devinfo)
 	fwreq->items[BRCMF_PCIE_FW_CODE].type = BRCMF_FW_TYPE_BINARY;
 	fwreq->items[BRCMF_PCIE_FW_NVRAM].type = BRCMF_FW_TYPE_NVRAM;
 	fwreq->items[BRCMF_PCIE_FW_NVRAM].flags = BRCMF_FW_REQF_OPTIONAL;
-	fwreq->domain_nr = pci_domain_nr(devinfo->pdev->bus);
+	/* NVRAM reserves PCI domain 0 for Broadcom's SDK faked bus */
+	fwreq->domain_nr = pci_domain_nr(devinfo->pdev->bus) + 1;
 	fwreq->bus_nr = devinfo->pdev->bus->number;
 
 	return fwreq;
-- 
2.13.7

Re: [PATCH] brcmfmac: fix regression in parsing NVRAM for multiple devices

[Kalle Valo]

Rafa=C5=82 Mi=C5=82ecki <zajec5@gmail.com> writes:

> From: Rafa=C5=82 Mi=C5=82ecki <rafal@milecki.pl>
>
> NVRAM is designed to work with Broadcom's SDK Linux kernel which fakes
> PCI domain 0 for all internal MMIO devices. Since official Linux kernel
> uses platform devices for that purpose there is a mismatch in numbering
> PCI domains.
>
> There used to be a fix for that problem but it was accidentally dropped
> during the last firmware loading rework. That resulted in brcmfmac not
> being able to extract device specific NVRAM content and all kind of
> calibration problems.
>
> Reported-by: Aditya Xavier <adityaxavier@gmail.com>
> Fixes: 2baa3aaee27f ("brcmfmac: introduce brcmf_fw_alloc_request() functi=
on")
> Cc: stable@vger.kernel.org # v4.17+
> Signed-off-by: Rafa=C5=82 Mi=C5=82ecki <rafal@milecki.pl>

I'll queue this to 4.18 (if there's still time). I just came from
vacation and have some catching up to do :)

--=20
Kalle Valo

Re: [PATCH] brcmfmac: fix regression in parsing NVRAM for multiple devices

[Arend van Spriel]

On 7/22/2018 11:46 PM, Rafał Miłecki wrote:
> From: Rafał Miłecki <rafal@milecki.pl>
>
> NVRAM is designed to work with Broadcom's SDK Linux kernel which fakes
> PCI domain 0 for all internal MMIO devices. Since official Linux kernel
> uses platform devices for that purpose there is a mismatch in numbering
> PCI domains.
>
> There used to be a fix for that problem but it was accidentally dropped
> during the last firmware loading rework. That resulted in brcmfmac not
> being able to extract device specific NVRAM content and all kind of
> calibration problems.
>
> Reported-by: Aditya Xavier <adityaxavier@gmail.com>
> Fixes: 2baa3aaee27f ("brcmfmac: introduce brcmf_fw_alloc_request() function")
> Cc: stable@vger.kernel.org # v4.17+

oops. my bad.

Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com>
> Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
> ---
>   drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c
> index 45928b5b8d97..4fffa6988087 100644
> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c
> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c
> @@ -1785,7 +1785,8 @@ brcmf_pcie_prepare_fw_request(struct brcmf_pciedev_info *devinfo)
>   	fwreq->items[BRCMF_PCIE_FW_CODE].type = BRCMF_FW_TYPE_BINARY;
>   	fwreq->items[BRCMF_PCIE_FW_NVRAM].type = BRCMF_FW_TYPE_NVRAM;
>   	fwreq->items[BRCMF_PCIE_FW_NVRAM].flags = BRCMF_FW_REQF_OPTIONAL;
> -	fwreq->domain_nr = pci_domain_nr(devinfo->pdev->bus);
> +	/* NVRAM reserves PCI domain 0 for Broadcom's SDK faked bus */
> +	fwreq->domain_nr = pci_domain_nr(devinfo->pdev->bus) + 1;
>   	fwreq->bus_nr = devinfo->pdev->bus->number;
>
>   	return fwreq;
>

Re: [PATCH] brcmfmac: fix regression in parsing NVRAM for multiple devices

[Kalle Valo]

Rafał Miłecki wrote:

> From: Rafał Miłecki <rafal@milecki.pl>
> 
> NVRAM is designed to work with Broadcom's SDK Linux kernel which fakes
> PCI domain 0 for all internal MMIO devices. Since official Linux kernel
> uses platform devices for that purpose there is a mismatch in numbering
> PCI domains.
> 
> There used to be a fix for that problem but it was accidentally dropped
> during the last firmware loading rework. That resulted in brcmfmac not
> being able to extract device specific NVRAM content and all kind of
> calibration problems.
> 
> Reported-by: Aditya Xavier <adityaxavier@gmail.com>
> Fixes: 2baa3aaee27f ("brcmfmac: introduce brcmf_fw_alloc_request() function")
> Cc: stable@vger.kernel.org # v4.17+
> Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
> Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com>

Patch applied to wireless-drivers.git, thanks.

299b6365a3b7 brcmfmac: fix regression in parsing NVRAM for multiple devices

-- 
https://patchwork.kernel.org/patch/10539435/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches