On Tue, Aug 22, 2017 at 09:47:27PM +0200, Mark Kettenis wrote:
> > Date: Tue, 22 Aug 2017 22:12:51 +0300
> > From: Artturi Alm <[email protected]>
> >
> > Hi,
> >
> > i came across dwc2 usb node like this:
> >
> > usb@101c0000 {
> > compatible = "snps,dwc2";
> > reg = <0x101c0000 0x40000>;
> > interrupts = <0x0 0x11 0x4>;
> > clocks = <0x2 0x1c9>;
> > clock-names = "otg";
> > dr_mode = "host";
> > phys = <0x6>;
> > phy-names = "usb2-phy";
> > status = "okay";
> > };
> >
> >
> > and as it's at /-root, there's not much i could use to identify between
> > broadcom & rockchip while keeping the code simple for future additions,
> > and not mess w/each SoC compat string individually when applying vendor-
> > specific things..
> >
> > i guess this should not be abused ever alone w/o some && in any _match(),
> > but when used responsibly may make some _attach() funcs prettier:)
> >
> > to export the _vend variant, or keep as strchr hack like below?
> > (untested diff rly, for just the discussion atm.)
>
> All the nodes I have seen have a more specific string included in the
> compatible property. Also, many, if not all of the parameters have
> sane default values or values that can be read from a configuration
> register.
>
> So no, I don't think what you're proposing is a good idea.
>
ok, i'll try to get this fixed in u-boot.
but incase i was unclear, the diff was supposed to allow ie. this where the
specifics obviously doesn't matter beyond "," after vendor in compatible:
diff --git a/sys/dev/fdt/if_dwge_fdt.c b/sys/dev/fdt/if_dwge_fdt.c
index edfe5acb992..a874274c215 100644
--- a/sys/dev/fdt/if_dwge_fdt.c
+++ b/sys/dev/fdt/if_dwge_fdt.c
@@ -120,9 +120,7 @@ dwge_fdt_attach(struct device *parent, struct device *self,
void *aux)
/* Do hardware specific initializations. */
if (OF_is_compatible(faa->fa_node, "allwinner,sun7i-a20-gmac"))
dwge_fdt_attach_allwinner(fsc);
- else if (OF_is_compatible(faa->fa_node, "rockchip,rk3288-gmac"))
- dwge_fdt_attach_rockchip(fsc);
- else if (OF_is_compatible(faa->fa_node, "rockchip,rk3399-gmac"))
+ else if (OF_is_compatible(faa->fa_node, "rockchip"))
dwge_fdt_attach_rockchip(fsc);
/* Enable clock. */
> Regarding your mail from a few days ago. I did try something like
> your diff to get the usb ports on the rk3288-tinker to work but failed
> so far.
>
u-boot is still rather broken for rk3188, and success-rate on even u-boot
finding the attached devices is less than 1/10... but one try got this far:
U-Boot SPL 2017.09-rc2-00151-g2d7cb5b-dirty (Aug 21 2017 - 18:15:17)
Returning to boot ROM...
U-Boot 2017.09-rc2-00151-g2d7cb5b-dirty (Aug 21 2017 - 18:15:17 +0300)
Model: Radxa Rock
DRAM: 2 GiB
MMC: dwmmc@10214000 - probe failed: -22
*** Warning - No MMC card found, using default environment
In: serial@20064000
Out: serial@20064000
Err: serial@20064000
Model: Radxa Rock
Net: Net Initialization Skipped
No ethernet found.
Hit any key to stop autoboot: 0
starting USB...
USB0: Core Release: 2.91a
dwc_otg_core_host_init: Timeout!
dwc_otg_core_host_init: Timeout!
dwc_otg_core_host_init: Timeout!
dwc_otg_core_host_init: Timeout!
dwc_otg_core_host_init: Timeout!
dwc_otg_core_host_init: Timeout!
dwc_otg_core_host_init: Timeout!
dwc_otg_core_host_init: Timeout!
USB1: Core Release: 2.91a
scanning bus 0 for devices... 1 USB Device(s) found
scanning bus 1 for devices... 5 USB Device(s) found
scanning usb for storage devices... 1 Storage Device(s) found
scanning usb for ethernet devices... 0 Ethernet Device(s) found
Device 0: Vendor: Kingston Rev: PMAP Prod: DataTraveler 2.0
Type: Removable Hard Disk
Capacity: 14891.4 MB = 14.5 GB (30497664 x 512)
... is now current device
Scanning usb 0:1...
Found EFI removable media binary efi/boot/bootarm.efi
reading efi/boot/bootarm.efi
64908 bytes read in 70 ms (905.3 KiB/s)
libfdt fdt_check_header(): FDT_ERR_BADMAGIC
## Starting EFI application at 62000000 ...
Scanning disk [email protected]...
Scanning disk usb_mass_storage.lun0...
Found 2 disks
>> OpenBSD/armv7 BOOTARM 0.8
boot>
booting sd0a:/bsd: 3887248+166476+499544 [281845+90+515968+242685]=0x558c3c
OpenBSD/armv7 booting ...
arg0 0xc0858c3c arg1 0x0 arg2 0x68000000
Allocating page tables
freestart = 0x60859000, free_pages = 522151 (0x0007f7a7)
IRQ stack: p0x60887000 v0xc0887000
ABT stack: p0x60888000 v0xc0888000
UND stack: p0x60889000 v0xc0889000
SVC stack: p0x6088a000 v0xc088a000
Creating L1 page table at 0x6085c000
Mapping kernel
Constructing L2 page tables
undefined page pmap [ using 1041048 bytes of bsd ELF symbol table ]
board type: 0
Copyright (c) 1982, 1986, 1989, 1991, 1993
The Regents of the University of California. All rights reserved.
Copyright (c) 1995-2017 OpenBSD. All rights reserved. https://www.OpenBSD.org
OpenBSD 6.2-beta (GENERIC) #4: Mon Aug 21 18:27:38 EEST 2017
[email protected]:/usr/src/sys/arch/armv7/compile/GENERIC
real mem = 2147483648 (2048MB)
avail mem = 2097254400 (2000MB)
mainbus0 at root: Radxa Rock
cpu0 at mainbus0: ARM Cortex-A9 r3p0 (ARMv7)
cpu0: DC enabled IC enabled WB disabled EABT branch prediction enabled
cpu0: 32KB(32b/l,4way) I-cache, 32KB(32b/l,4way) wr-back D-cache
cortex0 at mainbus0
amptimer0 at cortex0: tick rate 396000 KHz
armliicc0 at cortex0: rtl 0 waymask: 0x00000007
simplebus0 at mainbus0: "amba"
ampintc0 at mainbus0 nirq 160, ncpu 4
com0 at mainbus0: ns16550, no working fifo
com1 at mainbus0: ns16550, no working fifo
dwctwo0 at mainbus0
dwdog0 at mainbus0
com2 at mainbus0: ns16550, no working fifo
com2: console
com3 at mainbus0: ns16550, no working fifo
usb0 at dwctwo0: USB revision 2.0
uhub0 at usb0 configuration 1 interface 0 "Broadcom DWC2 root hub" rev
2.00/1.00 addr 1
uhub0: device problem, disabling port 1
vscsi0 at root
scsibus0 at vscsi0: 256 targets
softraid0 at root
scsibus1 at softraid0: 256 targets
boot device: lookup 'sd0a:/bsd' failed.
root device:
use one of: exit
root device:
So, it's broken here on rk3188 too and needs more work.
-Artturi
> > diff --git a/sys/dev/ofw/fdt.c b/sys/dev/ofw/fdt.c
> > index d6a64499e88..f22ed7a9f49 100644
> > --- a/sys/dev/ofw/fdt.c
> > +++ b/sys/dev/ofw/fdt.c
> > @@ -42,6 +42,7 @@ int fdt_translate_reg(void *, struct fdt_reg *);
> > #ifdef DEBUG
> > void fdt_print_node_recurse(void *, int);
> > #endif
> > +static int fdt_is_compatible_vend(void *, const char *);
> >
> > static int tree_inited = 0;
> > static struct fdt tree;
> > @@ -634,6 +635,24 @@ fdt_is_compatible(void *node, const char *name)
> > return 0;
> > }
> >
> > +int
> > +fdt_is_compatible_vend(void *node, const char *name)
> > +{
> > + size_t vlen = strlen(name);
> > + char *data;
> > + int len;
> > +
> > + len = fdt_node_property(node, "compatible", &data);
> > + while (len > 0) {
> > + if (strncmp(data, name, vlen) == 0)
> > + return 1;
> > + len -= strlen(data) + 1;
> > + data += strlen(data) + 1;
> > + }
> > +
> > + return 0;
> > +}
> > +
> > #ifdef DEBUG
> > /*
> > * Debug methods for printing whole tree, particular odes and properies
> > @@ -902,6 +921,8 @@ int
> > OF_is_compatible(int handle, const char *name)
> > {
> > void *node = (char *)tree.header + handle;
> > + if (strchr(name, ',') == NULL)
> > + return fdt_is_compatible_vend(node, name);
> > return (fdt_is_compatible(node, name));
> > }
> >
> >
