From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marek Vasut Date: Thu, 30 Aug 2018 15:42:40 +0200 Subject: [U-Boot] [PATCH V2 1/2] pci: Support parsing PCI controller DT subnodes In-Reply-To: References: <20180824182718.10879-1-marek.vasut+renesas@gmail.com> <5256dc05-92f4-fc4d-ca0a-73447e08d56f@gmail.com> <66e48ae0-1309-1ede-5182-d5de4ceb00dd@gmail.com> Message-ID: <02bb72fd-eef5-c7ca-e546-4b1c2343afbc@gmail.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On 08/30/2018 03:32 PM, Bin Meng wrote: > Hi Marek, > > On Thu, Aug 30, 2018 at 1:07 AM Marek Vasut wrote: >> >> On 08/29/2018 05:15 PM, Bin Meng wrote: >>> +Simon >>> >>> Hi Marek, >>> >>> On Wed, Aug 29, 2018 at 10:22 PM Marek Vasut wrote: >>>> >>>> On 08/24/2018 08:27 PM, Marek Vasut wrote: >>>>> The PCI controller can have DT subnodes describing extra properties >>>>> of particular PCI devices, ie. a PHY attached to an EHCI controller >>>>> on a PCI bus. This patch parses those DT subnodes and assigns a node >>>>> to the PCI device instance, so that the driver can extract details >>>>> from that node and ie. configure the PHY using the PHY subsystem. >>>>> >>>>> Signed-off-by: Marek Vasut >>>>> Cc: Simon Glass >>>>> Cc: Tom Rini >>>> >>>> Well, bump ? >>>> >>>> This is the only missing patch to get my hardware working properly. >>> >>> I don't think we ever had an agreement on the v1 patch. Simon had a >>> long email that pointed out what Linux does seems like a 'fallback' to >>> find a node with no compatible string. >>> >>> Back to this, if we have to go with this way, please create a test >>> case to cover this scenario. >> >> The fact that it works on a particular board is not tested enough? >> Do we need a custom, special, synthetic test ? >> > > I believe that's always been the requirement against the DM code > changes. I was requested in the past when I changed something in the > DM and I see other people were asked to do so. Like Alex said, it does > not mean this patch was not tested enough, but to ensure future > commits won't break this. So, do you have any suggestion how to implement this test ? It seems Alex posed the same question. It doesn't seem to be trivial in the context of sandbox. >> Anyway, any feedback on the patch ? Did you test it ? I again only see >> "do this random stuff and that random stuff" , but zero actual feedback. >> > > If "this and that random stuff" means test case I asked for, please > check my proposal on the v1 patch thread which indicated that a proper > test case should be created. You seems to have missed that. So, any feedback on this actual patch ? -- Best regards, Marek Vasut