Re: [coreboot] proposed change to dtc to allow multiple dts files per single chip

[Carl-Daniel Hailfinger]

On 29.01.2008 17:44, ron minnich wrote:
> Second try.
> In the current version of dtc, if one has the line:
> /config/ = "northbridge/amd/geodelx";
>
> Then the file northbridge/amd/geodelx/dts is read in and processed. 
> Magic(TM) appends the name "/dts" to the path.
>
> This hack is fine with chips that only do one thing. 
> But some (all) northbridge parts play several roles: APIC cluster, PCI domain
> device, and PCI device. The result is a need for more than one dts, since
> there are three possible devices, with three types of IDs, and so on.
>
> To keep things sane, I am proposing to enable multiple dts files in a
> directory, names (e.g., nothing required here):
> domaindts
> pcidts
> apicdts
>
> (of course these names can be anything, this is just an example).
> This change will require a change to the dtc, since we can no longer
> assume just one dts file, and hence need a way to name these different 
> files.
>
> The proposed change is very simple. We now require the full path name 
> for the file, and eliminate the Magic(TM).
>
> So, 
> /config/ = "northbridge/amd/geodelx/pcidts";
>
> will open the pcidts file. 
> /config/ = "northbridge/amd/geodelx/domaindts";
> will open the domain dts.
>
> Maybe we should just call it domain and pci and apic? works for me.
> /config/ = "northbridge/amd/geodelx/domain";
> /config/ = "northbridge/amd/geodelx/pcibridge";
> /config/ = "northbridge/amd/geodelx/apic";
>
> Changes: 
> dtc.c: create a new function, fopenfile, that will only open a path if it 
> really is a file. Modify dtc_open_file to use this function. fopenfile
> assumes "-" means stdin; should it, or should I move that assumption back
> to dtc_open_file?
> dtc.h: add prototypes
> dtc-parser.y: Given a config path, open the path.
> southbridge/amd/cs5536/cs5536.c: example of how C code changes
>
> Signed-off-by: Ronald G. Minnich

Acked-by: Carl-Daniel Hailfinger

Please see the comments below, but they do not have to be addressed for
this commit, just keep them in mind for future commits in that area.
> Index: util/dtc/dtc.c
>
> --- util/dtc/dtc.c	(revision 565)
> +++ util/dtc/dtc.c	(working copy)
> @@ -61,21 +61,47 @@
>    	fill_fullpaths(child, tree->fullpath);
>  }
>
> -static FILE *dtc_open_file(char *fname)
> +/* How stupid is POSIX? This stupid: you can fopen a directory if mode
> is "r", but then there is very little you can do with it except get errors.
> This function now makes sure that the thing you open is a file.
> */
>
> +FILE *fopenfile(char *fname)
>  {
> FILE *f;
> FILE *f = NULL;
> if (streq(fname, "-"))
> if (streq(fname, "-")){
> f = stdin;
> else
> } else {
> struct stat buf;
> if (stat(fname, &buf) < 0) {
> 	errno = EISDIR;
> 	goto no;
> }
> if (! S_ISREG(buf.st_mode)){
> 	errno = EISDIR;
> 	goto no;
> }
> f = fopen(fname, "r");
> }
>
> +no:
> return f;
>
> +}
>
> +FILE *dtc_open_file(char *fname)
> +{
>
> FILE *f = fopenfile(fname);
>
> if (! f)
>   die("Couldn't open "%s": %s\n", fname, strerror(errno));
>
> return f;
>
> }
>
> static void usage(void)
>  {
>    fprintf(stderr, "Usage:\n");
> Index: util/dtc/dtc.h
> ===================================================================
> --- util/dtc/dtc.h	(revision 565)
> +++ util/dtc/dtc.h	(working copy)
> @@ -31,6 +31,8 @@
>  #include <errno.h>
>  #include <unistd.h>
>  #include <netinet/in.h>
> +#include <sys/types.h>
> +#include <sys/stat.h>
>
> /* also covers (part of?) linux's byteswap.h functionality */
>  #include "endian.h"
> @@ -236,5 +238,6 @@
>
> char *join_path(char *path, char *name);
>  void fill_fullpaths(struct node *tree, char *prefix);
>
> +FILE *fopenfile(char *fname);
> +FILE *dtc_open_file(char *fname);
>  #endif /* _DTC_H */
> Index: util/dtc/dtc-parser.y
> ===================================================================
> --- util/dtc/dtc-parser.y	(revision 565)
> +++ util/dtc/dtc-parser.y	(working copy)
> @@ -121,17 +121,16 @@
>  config:        DT_CONFIG '(' 
>    	includepath {
>    		void switchin(FILE *f);
> 	/* switch ... */
> 	char path[1024];
> FILE *f;
> 	/* The need for a cast here is silly */
> 	char *name = (char *)$3.val;
> /* TODO: keep track of which of these we have read in. If we have already done it, then 
>   * don't do it twice. 
>   */
> 	sprintf(path, "%s/dts", $3.val);
> 	f  = fopen(path, "r");
> 	f  = fopenfile(name);
> if (! f){
> 		perror(path);
> 		perror(name);
> 	exit(1);
> }
> switchin(f);
> Index: southbridge/amd/cs5536/cs5536.c
>
> --- southbridge/amd/cs5536/cs5536.c	(revision 565)
> +++ southbridge/amd/cs5536/cs5536.c	(working copy)
> @@ -172,7 +172,7 @@
> @param sb Southbridge config structure.
>
> */
> -static void lpc_init(struct southbridge_amd_cs5536_config *sb)
> +static void lpc_init(struct southbridge_amd_cs5536_dts_config *sb)
>  {
>    struct msr msr;
>
> @@ -220,7 +220,7 @@
> @param sb Southbridge config structure.
>
> */
> -static void uarts_init(struct southbridge_amd_cs5536_config *sb)
> +static void uarts_init(struct southbridge_amd_cs5536_dts_config *sb)
>  {
>    struct msr msr;
>    u16 addr = 0;
> @@ -383,7 +383,7 @@
> @param sb Southbridge config structure.
>
> */
> -static void enable_USB_port4(struct southbridge_amd_cs5536_config *sb)
> +static void enable_USB_port4(struct southbridge_amd_cs5536_dts_config *sb)
>  {
>    u32 *bar;
>    struct msr msr;
> @@ -475,7 +475,7 @@
>  {
>    struct device *dev;
>    struct msr msr;
> struct southbridge_amd_cs5536_config *sb;
>
> struct southbridge_amd_cs5536_dts_config *sb;
> const struct msrinit *csi;
>
> post_code(P80_CHIPSET_INIT);
>
> @@ -486,7 +486,7 @@
>    	       __FUNCTION__);
>    	return;
>    }
> sb = (struct southbridge_amd_cs5536_config *)dev->device_configuration;
> sb = (struct southbridge_amd_cs5536_dts_config *)dev->device_configuration;
>
> #if 0
>    if (!IsS3Resume())
> @@ -548,8 +548,8 @@
>   */
>  static void southbridge_init(struct device *dev)
>  {
> struct southbridge_amd_cs5536_config *sb =
>    (struct southbridge_amd_cs5536_config *)dev->device_configuration;
>
> struct southbridge_amd_cs5536_dts_config *sb =
>    (struct southbridge_amd_cs5536_dts_config *)dev->device_configuration;
>
> /*
> struct device *gpiodev;

struct southbridge_amd_cs5536_dts_config is a rather ling name.
While I agree that we want to reflect the name of the dts in the name of
the struct, we probably want to keep the name as short as possible.
Since _dts is already part of the struct name, maybe we can remove the
_config from the end.

Regards,
Carl-Daniel

proposed change to dtc to allow multiple dts files per single chip

[ron minnich]

Committed revision 566.

I will keep your comments in mind.

thanks!

ron