From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail-vk0-x22c.google.com ([2607:f8b0:400c:c05::22c]) by bombadil.infradead.org with esmtps (Exim 4.80.1 #2 (Red Hat Linux)) id 1aWogD-0006BO-FG for barebox@lists.infradead.org; Fri, 19 Feb 2016 17:17:22 +0000 Received: by mail-vk0-x22c.google.com with SMTP id k196so80871514vka.0 for ; Fri, 19 Feb 2016 09:17:00 -0800 (PST) MIME-Version: 1.0 In-Reply-To: <1455823875.996.47.camel@rtred1test09.kymeta.local> References: <1455672559-25061-1-git-send-email-andrew.smirnov@gmail.com> <1455672559-25061-16-git-send-email-andrew.smirnov@gmail.com> <1455676867.18517.64.camel@rtred1test09.kymeta.local> <1455823875.996.47.camel@rtred1test09.kymeta.local> Date: Fri, 19 Feb 2016 09:17:00 -0800 Message-ID: From: Andrey Smirnov List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "barebox" Errors-To: barebox-bounces+u.kleine-koenig=pengutronix.de@lists.infradead.org Subject: Re: [PATCH 15/18] [RFC] net: eth: Always use DEVICE_ID_DYNAMIC To: Trent Piepho Cc: "barebox@lists.infradead.org" On Thu, Feb 18, 2016 at 11:30 AM, Trent Piepho wrote: > On Tue, 2016-02-16 at 20:16 -0800, Andrey Smirnov wrote: >> >> This patch solves the problem by forcing all Ethernet adapters to >> >> use dynamic ID allocation. >> > >> > A lot of systems depend on aliases/ethernet0 pointing to the proper >> > ethernet adapter. How will they work with this? >> > >> > How will a boot script know which ethernet device to use ethact on? >> > >> > Suppose someone who doesn't know much about device trees and hasn't >> > looked at the SoC datasheet, which in my experience is about 80% of the >> > barebox userbase, comes up to you and says, "This port on the board is >> > labeled eth1, how do tell barebox to use it?" Right now the answer is >> > "ethact eth1" but after this change, which basically eliminates the OF >> > alias system, the answer is???? I think it's going to be something that >> > said developer doesn't like very much. >> >> I agree with your points and this commit isn't really a proper fix, I >> put an RFC tag there to indicate that this patch was more of "Hey >> guys, there's this problem, let's discuss how we solve it" rather than >> "There's your problem, here's the fix". >> >> > >> > Maybe a better solution would be to have dynamically allocated devices >> > not use IDs that have been "reserved" by the existence of an OF alias? >> > There wouldn't be some call to reserve an id, just the alias's existence >> > would be sufficient. Common code that allocates IDs should have all the >> > info it needs to check for an alias and then either use it or allocate >> > an id not already aliased. >> >> It is a fix, but only partial. To play devil's advocate, imagine a >> board that has SoC with two PCIe Ethernet chips (think i210) attached >> to it. Even if you query for aliases you get back to the same problem >> where your silkscreen labels are not providing any good info to inform >> general user what to give to "ethact" (granted it is possible to >> assigned a DT node to PCI device and thus probably possible to create >> an aliases for such interfaces, but it is very clunky in practice not >> many .dts files has any provisions like that) > > Adding an alias for such a device seems reasonable to me. My though > process goes something like this: > > If the ports are a fixed part of the device, even if on a probe-able bus > like PCIe or USB, then we should be able to assign them a name. They > are always there, always in the same place, always connected the same > way, so it's not impossible. > > So we use what's constant, how the port is connected. It will be a > certain device-function in a certain slot on a certain bus on a certain > pci controller. And we make that the name because it's always the same. > > But that name is huge and impossible to remember or type in. So we > create a link, or alias, for it. An easier to use name that points to > full name. I don't necessarily agree with this point: - IMHO, the amount if information needed to be encoded in the name to cover 99% of the use-cases is not that impossible to type or remember, compare e.g. "eth-pci-00-01", "eth-usb-00-01", "eth-spi-00-01" to "eth0", "eth1" - Even if we ignore the first point. IMHO, one's chances to remember that "that one port to the left edge of the board" is "eth0" and "the one to the right edge of the board" is "eth1" are the same as remembering that former is "eth-pci-xx-yy" and latter is "eth-spi-zz-zz", so the only way to make the process painless is to have appropriate silkscreen and maybe labels on the case of the device. At this point it doesn't make that much difference if it is the "old" style name or a "new" one that's captured in those places. > > And thus we arrive at what is basically the OF alias system. > It's not that this can't "solve" this problem, but in my opinion (and I could be completely wrong) the purpose of aliases in DT is to allow flexibility in DT code and the fact it can be used to alter the behavior of BB is somewhat of a serendipitous side-effect. What I don't like about this solution is that it creates a dependency between DT and BB's board's scripts that doesn't have to be there. > >> I think the most sound fix for this problem would be to embed the >> NIC's topology information into its name (not unlike Consistent > > Isn't that basically the same thing as the OF path of a device? Yes and no, it would be a rather abridged version of OF path that omits a number of irrelevant details _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox