On 6-7-2016 15:42, Arnd Bergmann wrote: > On Wednesday, July 6, 2016 10:08:55 AM CEST Arend Van Spriel wrote: >> On Tue, Jul 5, 2016 at 3:43 PM, Arnd Bergmann <arnd@xxxxxxxx> wrote: >>> On Monday, July 4, 2016 8:36:05 PM CEST Arend van Spriel wrote: >>>> On 04-07-16 16:54, Arnd Bergmann wrote: >>>>> On Monday, July 4, 2016 11:08:38 AM CEST Arend Van Spriel wrote: >>>>> >>>>> In drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c I already see >>>>> over a dozen different chips being supported, bcm4329 is only one of >>>>> them. In particular, there seem to be some that have various modules: >>>>> >>>>> BRCMF_FW_NVRAM_ENTRY(BRCM_CC_43241_CHIP_ID, 0x0000001F, 43241B0), >>>>> BRCMF_FW_NVRAM_ENTRY(BRCM_CC_43241_CHIP_ID, 0x00000020, 43241B4), >>>>> BRCMF_FW_NVRAM_ENTRY(BRCM_CC_43241_CHIP_ID, 0xFFFFFFC0, 43241B5), >>>>> >>>>> So if you have a bcm43241, that compatible string probably should >>>>> include both brcm,bcm43241-b4-fmac and brcm,bcm43241-fmac, possibly also >>>>> brcm,bcm4329-fmac, to show that it is a superset of the programming >>>>> interface of that one. >>>> >>>> Hi Arnd, >>>> >>>> I have to disagree here. The compatible string "brcm,bcm4329-fmac" is >>>> chosen as the bcm4329 chip was the first supported and we never added >>>> others as there is no other programming required. For all supported >>>> devices the same device tree properties apply and are handled the same. >>>> As such there is no need to come up with a new compatible string. >>> >>> Generally speaking, the way that the compatible strings work is that you >>> add a new one whenever you get a new piece of hardware, and you can leave >>> the first one as a fallback so you don't have to change the driver. >>> >>> Adding the new string for the actual device is important though, >>> since you might only discover later that they are not 100% compatible >>> and that you in fact need to know the difference. >>> >>> For discoverable buses like sdio or usb, it may actually be better >>> to not identify the device through the compatible property at all, >>> and instead use a string that is generated from the actual identifier >>> as the primary key, as e.g. documented in >>> Documentation/devicetree/bindings/usb/usb-device.txt >> >> Well, that is my point. We do not need to identify the device here. We >> can obtain that info, ie. chip id and revision, from the device itself >> as our wifi chips have a discoverable AXI backplane. What is missing >> is that we have no way to tell what board/module this device is >> integrated on, which makes it impossible to select the correct nvram >> file. The model property can fill that gap just nicely. > > We have multiple inconsistencies here, and it's not this driver that is > particularly messed up, but I think using the model property here > would make it worse: > > All existing uses of the model property in arch/arm/boot/dts and most of > the ones in arch/powerpc/boot/dts are against the intended usage in > one way or another, but adding different kind of incorrect usage won't > improve that. > > The only way I can see the model property being used correctly would > be to have it match the first entry in the compatible property, but > that is completely redundant, so we tend to omit it, except for the > root node in which it is required. For the root node however, the > historic practice that has crept in on ARM is to put something completely > different in there, which is a human-readable description of the > machine rather than something we can use as a unique indentifier. > > I'd just consider the "model" property burned, and not use it for anything > that doesn't already use it, just like we handle "device_type": a few > things require it, nothing else should use it. If that is the agreed approach in devicetree arena I am fine with it. I have been unaware of this and just looked at the suggestion from Jonas seeing a solution to the problem at hand. > I agree with your point that the "compatible" property in case of brcmfmac > is really odd because it is not required to identify the hardware when > the SDIO device identification is sufficient, and we just need it to guard > the definition of the other properties. However it seems that now we > actually do need to identify the hardware because we cannot read the > board ID and revision that is supposed to come from nvram but also needed > to read the nvram contents from a file. The board ID and rev in the nvram may not be used if the device has these stored in OTP. However, the problem is that the device need firmware+nvram loaded into it before we can start the firmware on the device. Now if we were to add boardtype and boardrev properties in the binding to identify the hardware, I suppose a new compatible value would be required. Regards, Arend -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
