Hi Simon > -----Original Message----- > From: s...@google.com [mailto:s...@google.com] On Behalf Of Simon Glass > Sent: Friday, November 11, 2016 18:17 > To: Yehuda Yitschak > Cc: Bin Meng; Heiko Schocher; Przemyslaw Marczak; Stefan Roese; Stephen > Warren; U-Boot Mailing List > Subject: Re: [PATCH v2 1/1] cmd: pci: add option to parse and display BAR > information > > Hi, > > On 6 November 2016 at 07:31, <yehu...@marvell.com> wrote: > > From: Yehuda Yitschak <yehu...@marvell.com> > > > > Currently the PCI command only allows to see the BAR register values > > but not the size and actual base address. > > This little extension parses the BAR registers and displays the base, > > size and type of each BAR. > > > > Signed-off-by: Yehuda Yitschak <yehu...@marvell.com> > > --- > > cmd/pci.c | 95 > > > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > +++++ > > 1 file changed, 95 insertions(+) > > Reviewed-by: Simon Glass <s...@chromium.org> > > > > > diff --git a/cmd/pci.c b/cmd/pci.c > > index 2f4978a..51eb536 100644 > > --- a/cmd/pci.c > > +++ b/cmd/pci.c > > @@ -92,6 +92,96 @@ static void pci_show_regs(pci_dev_t dev, struct > > pci_reg_info *regs) } #endif > > > > +#ifdef CONFIG_DM_PCI > > +int pci_bar_show(struct udevice *dev) #else int > > +pci_bar_show(pci_dev_t dev) #endif { > > + u8 header_type; > > + int bar_cnt, bar_id, is_64, is_io, mem_type; > > + u32 base_low, base_high; > > + u32 size_low, size_high; > > + u64 base, size; > > + u32 reg_addr; > > + int prefetchable; > > + > > +#ifdef CONFIG_DM_PCI > > I think you can implement only the DM_PCI case, since we are trying to move > everything to DM_PCI.
Cool. Less clutter > > > + dm_pci_read_config8(dev, PCI_HEADER_TYPE, &header_type); #else > > + pci_read_config_byte(dev, PCI_HEADER_TYPE, &header_type); > > +#endif > > + > > + if (header_type == PCI_HEADER_TYPE_CARDBUS) { > > + printf("CardBus doesn't support BARs\n"); > > + return -1; > > -ENOSYS perhaps. This is -EPERM which seems wrong. > > > + } > > + > > + bar_cnt = (header_type == PCI_HEADER_TYPE_NORMAL) ? 6 : 2; > > + > > + printf("ID Base Size Width > > Type\n"); > > + > > + printf("----------------------------------------------------------\n > > + "); > > + > > + bar_id = 0; > > + reg_addr = PCI_BASE_ADDRESS_0; > > + while (bar_cnt) { > > +#ifdef CONFIG_DM_PCI > > + dm_pci_read_config32(dev, reg_addr, &base_low); > > + dm_pci_write_config32(dev, reg_addr, 0xFFFFFFFF); > > Suggest lower-case hex. > > > + dm_pci_read_config32(dev, reg_addr, &size_low); > > + dm_pci_write_config32(dev, reg_addr, base_low); #else > > + pci_read_config_dword(dev, reg_addr, &base_low); > > + pci_write_config_dword(dev, reg_addr, 0xFFFFFFFF); > > + pci_read_config_dword(dev, reg_addr, &size_low); > > + pci_write_config_dword(dev, reg_addr, base_low); > > +#endif > > + reg_addr += 4; > > + > > + base = base_low & ~0xF; > > + size = size_low & ~0xF; > > + base_high = 0x0; > > + size_high = 0xFFFFFFFF; > > + is_64 = 0; > > + prefetchable = base_low & PCI_BASE_ADDRESS_MEM_PREFETCH; > > + is_io = base_low & PCI_BASE_ADDRESS_SPACE_IO; > > + mem_type = base_low & > PCI_BASE_ADDRESS_MEM_TYPE_MASK; > > + > > + if (mem_type == PCI_BASE_ADDRESS_MEM_TYPE_64) { #ifdef > > +CONFIG_DM_PCI > > + dm_pci_read_config32(dev, reg_addr, &base_high); > > + dm_pci_write_config32(dev, reg_addr, 0xFFFFFFFF); > > + dm_pci_read_config32(dev, reg_addr, &size_high); > > + dm_pci_write_config32(dev, reg_addr, > > +base_high); #else > > + pci_read_config_dword(dev, reg_addr, &base_high); > > + pci_write_config_dword(dev, reg_addr, 0xFFFFFFFF); > > + pci_read_config_dword(dev, reg_addr, &size_high); > > + pci_write_config_dword(dev, reg_addr, > > +base_high); #endif > > + bar_cnt--; > > + reg_addr += 4; > > + is_64 = 1; > > + } > > + > > + base = base | ((u64)base_high << 32); > > + size = size | ((u64)size_high << 32); > > + > > + if ((!is_64 && size_low) || (is_64 && size)) { > > + size = ~size + 1; > > + printf(" %d 0x%016llx 0x%016llx %d %s > > %s\n", > > Can we drop the the address values? It is implied in U-Boot. If not, let's use > %#x. I prefer to keep them. It makes debugging much simpler I will change the format to %#016llx. > > > + bar_id, base, size, is_64 ? 64 : 32, > > + is_io ? "I/O" : "MEM", > > + prefetchable ? "Prefetchable" : ""); > > Check with sandbox, this gives a warning: > > cmd/pci.c: In function ‘pci_bar_show’: > cmd/pci.c:175:11: warning: format ‘%llx’ expects argument of type ‘long long > unsigned int’, but argument 3 has type ‘u64’ [-Wformat=] > prefetchable ? "Prefetchable" : ""); > ^ Strange, I can't see that. What compiler are you using when you get the warning ? I am using gcc-4.8 for armv8, maybe that's why I don't see the warnings I might come down to the built-in definition of __UINT64_TYPE__ which the sandbox arch uses > cmd/pci.c:175:11: warning: format ‘%llx’ expects argument of type ‘long long > unsigned int’, but argument 4 has type ‘u64’ [-Wformat=] > > > + } > > + > > + bar_id++; > > + bar_cnt--; > > + } > > + > > + return 0; > > +} > > + > > static struct pci_reg_info regs_start[] = { > > { "vendor ID", PCI_SIZE_16, PCI_VENDOR_ID }, > > { "device ID", PCI_SIZE_16, PCI_DEVICE_ID }, @@ -573,6 +663,7 > > @@ static int do_pci(cmd_tbl_t *cmdtp, int flag, int argc, char * const > argv[]) > > if (argc > 4) > > value = simple_strtoul(argv[4], NULL, 16); > > case 'h': /* header */ > > + case 'b': /* bars */ > > if (argc < 3) > > goto usage; > > if ((bdf = get_pci_dev(argv[2])) == -1) @@ -641,6 > > +732,8 @@ static int do_pci(cmd_tbl_t *cmdtp, int flag, int argc, char * > const argv[]) > > ret = pci_cfg_write(dev, addr, size, value); #endif > > break; > > + case 'b': /* bars */ > > + return pci_bar_show(dev); > > default: > > ret = CMD_RET_USAGE; > > break; > > @@ -663,6 +756,8 @@ static char pci_help_text[] = #endif > > "pci header b.d.f\n" > > " - show header of PCI device 'bus.device.function'\n" > > + "pci bar b.d.f\n" > > + " - show BARs base and size for device b.d.f'\n" > > "pci display[.b, .w, .l] b.d.f [address] [# of objects]\n" > > " - display PCI configuration space (CFG)\n" > > "pci next[.b, .w, .l] b.d.f address\n" > > -- > > 1.9.1 > > > > Regards, > Simon _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot