Message ID | 1453965300-24844-1-git-send-email-Qianyu.Gong@nxp.com |
---|---|
State | Rejected |
Headers | show |
On Thu, Jan 28, 2016 at 3:15 PM, Gong Qianyu <Qianyu.Gong@nxp.com> wrote: > From: Mingkai Hu <Mingkai.Hu@freescale.com> > > Referring to 'commit ff3e077bd23c ("dm: pci: Add a uclass for PCI")'. > > For legacy PCIe driver, it needs loop to read the vender_id from > devie 0 to devie 255 to check if there is device available. > Reading non-existen device will trigger the "Cannot read bus > configuration: -1" information. > > Signed-off-by: Mingkai Hu <Mingkai.Hu@freescale.com> > Signed-off-by: Gong Qianyu <Qianyu.Gong@nxp.com> > --- Which pci controller are you testing? [snip] Regards, Bin
Hi, > -----Original Message----- > From: Bin Meng [mailto:bmeng.cn@gmail.com] > Sent: Thursday, January 28, 2016 3:30 PM > To: Qianyu Gong <qianyu.gong@nxp.com> > Cc: U-Boot Mailing List <u-boot@lists.denx.de> > Subject: Re: [U-Boot] [PATCH] pci: gate print info of reading vender id with > CONFIG_DM_PCI > > On Thu, Jan 28, 2016 at 3:15 PM, Gong Qianyu <Qianyu.Gong@nxp.com> wrote: > > From: Mingkai Hu <Mingkai.Hu@freescale.com> > > > > Referring to 'commit ff3e077bd23c ("dm: pci: Add a uclass for PCI")'. > > > > For legacy PCIe driver, it needs loop to read the vender_id from devie > > 0 to devie 255 to check if there is device available. > > Reading non-existen device will trigger the "Cannot read bus > > configuration: -1" information. > > > > Signed-off-by: Mingkai Hu <Mingkai.Hu@freescale.com> > > Signed-off-by: Gong Qianyu <Qianyu.Gong@nxp.com> > > --- > > Which pci controller are you testing? > > [snip] > > Regards, > Bin Designware PCIe controller on LS2085A. Regards, Qianyu
On Thu, Jan 28, 2016 at 3:43 PM, Qianyu Gong <qianyu.gong@nxp.com> wrote: > Hi, > >> -----Original Message----- >> From: Bin Meng [mailto:bmeng.cn@gmail.com] >> Sent: Thursday, January 28, 2016 3:30 PM >> To: Qianyu Gong <qianyu.gong@nxp.com> >> Cc: U-Boot Mailing List <u-boot@lists.denx.de> >> Subject: Re: [U-Boot] [PATCH] pci: gate print info of reading vender id with >> CONFIG_DM_PCI >> >> On Thu, Jan 28, 2016 at 3:15 PM, Gong Qianyu <Qianyu.Gong@nxp.com> wrote: >> > From: Mingkai Hu <Mingkai.Hu@freescale.com> >> > >> > Referring to 'commit ff3e077bd23c ("dm: pci: Add a uclass for PCI")'. >> > >> > For legacy PCIe driver, it needs loop to read the vender_id from devie >> > 0 to devie 255 to check if there is device available. >> > Reading non-existen device will trigger the "Cannot read bus >> > configuration: -1" information. >> > >> > Signed-off-by: Mingkai Hu <Mingkai.Hu@freescale.com> >> > Signed-off-by: Gong Qianyu <Qianyu.Gong@nxp.com> >> > --- >> >> Which pci controller are you testing? >> >> [snip] >> >> Regards, >> Bin > > Designware PCIe controller on LS2085A. > Please check commit 7ba34ff09f1ef105521f914e4ad4e4ac19975bac "pci: layerscape: Adjust the return value when ls_pcie_addr_valid() fails" which is already in mainline. Regards, Bin
> -----Original Message----- > From: Bin Meng [mailto:bmeng.cn@gmail.com] > Sent: Thursday, January 28, 2016 4:02 PM > To: Qianyu Gong <qianyu.gong@nxp.com> > Cc: u-boot@lists.denx.de > Subject: Re: [U-Boot] [PATCH] pci: gate print info of reading vender id with > CONFIG_DM_PCI > > On Thu, Jan 28, 2016 at 3:43 PM, Qianyu Gong <qianyu.gong@nxp.com> wrote: > > Hi, > > > >> -----Original Message----- > >> From: Bin Meng [mailto:bmeng.cn@gmail.com] > >> Sent: Thursday, January 28, 2016 3:30 PM > >> To: Qianyu Gong <qianyu.gong@nxp.com> > >> Cc: U-Boot Mailing List <u-boot@lists.denx.de> > >> Subject: Re: [U-Boot] [PATCH] pci: gate print info of reading vender > >> id with CONFIG_DM_PCI > >> > >> On Thu, Jan 28, 2016 at 3:15 PM, Gong Qianyu <Qianyu.Gong@nxp.com> wrote: > >> > From: Mingkai Hu <Mingkai.Hu@freescale.com> > >> > > >> > Referring to 'commit ff3e077bd23c ("dm: pci: Add a uclass for PCI")'. > >> > > >> > For legacy PCIe driver, it needs loop to read the vender_id from > >> > devie > >> > 0 to devie 255 to check if there is device available. > >> > Reading non-existen device will trigger the "Cannot read bus > >> > configuration: -1" information. > >> > > >> > Signed-off-by: Mingkai Hu <Mingkai.Hu@freescale.com> > >> > Signed-off-by: Gong Qianyu <Qianyu.Gong@nxp.com> > >> > --- > >> > >> Which pci controller are you testing? > >> > >> [snip] > >> > >> Regards, > >> Bin > > > > Designware PCIe controller on LS2085A. > > > > Please check commit 7ba34ff09f1ef105521f914e4ad4e4ac19975bac "pci: > layerscape: Adjust the return value when ls_pcie_addr_valid() fails" > which is already in mainline. > > Regards, > Bin Ok. That's nice of you. And we also find another commit cab24b3407189a12 "dm: pci: Convert 'pci' command to driver model" submitted by Simon has fixed the issue. So no need of this patch. Thanks a lot. Regards, Qianyu
diff --git a/common/cmd_pci.c b/common/cmd_pci.c index 4e0951f..d94f90d 100644 --- a/common/cmd_pci.c +++ b/common/cmd_pci.c @@ -324,7 +324,9 @@ void pciinfo(int bus_num, int short_pci_listing) unsigned char header_type; unsigned short vendor_id; pci_dev_t dev; +#ifdef CONFIG_DM_PCI int ret; +#endif if (!hose) return; @@ -348,10 +350,14 @@ void pciinfo(int bus_num, int short_pci_listing) if (pci_skip_dev(hose, dev)) continue; +#ifdef CONFIG_DM_PCI ret = pci_read_config_word(dev, PCI_VENDOR_ID, &vendor_id); if (ret) goto error; +#else + pci_read_config_word(dev, PCI_VENDOR_ID, &vendor_id); +#endif if ((vendor_id == 0xFFFF) || (vendor_id == 0x0000)) continue; @@ -373,8 +379,10 @@ void pciinfo(int bus_num, int short_pci_listing) } return; +#ifdef CONFIG_DM_PCI error: printf("Cannot read bus configuration: %d\n", ret); +#endif } #endif