Refactoring and Endian bug fixes of fdt_support
v2 cover Refactoring and Endian bug fixes of fdt_support
Masahiro Yamada (9):
fdt_support: delete unnecessary DECLARE_GLOBAL_DATA_PTR
fdt_support: refactor with fdt_find_or_add_subnode helper func
fdt_support: delete force argument of fdt_initrd()
fdt_support: delete force argument of fdt_chosen()
fdt_support: refactor fdt_fixup_stdout() function
fdt_support: add 'const' qualifier for unchanged argument
fdt_support: fix an endian bug of fdt_fixup_memory_banks
fdt_support: fix an endian bug of fdt_initrd()
fdt_support: correct the return condition of fdt_initrd()
arch/microblaze/lib/bootm.c | 2 +- common/cmd_fdt.c | 4 +- common/fdt_support.c | 295 +++++++++++++++++++++----------------------- common/image-fdt.c | 4 +- include/fdt_support.h | 4 +- 5 files changed, 151 insertions(+), 158 deletions(-)
Signature · 1 line
1.8.3.2
PATCH v2 1/9 fdt_support: delete unnecessary DECLARE_GLOBAL_DATA_PTR
gd->bd is not used in fdt_support.c.
---
Changes in v2: None
common/fdt_support.c | 5 ----- 1 file changed, 5 deletions(-)
common/fdt_support.c +0 -5
@@ -17,11 +17,6 @@#include <exports.h> /* -* Global data (for the gd->bd) -*/ -DECLARE_GLOBAL_DATA_PTR ; - -/* * Get cells len in bytes * if #NNNN-cells property is 2 then len is 8 * otherwise len is 4
Signature · 1 line
1.8.3.2
PATCH v2 1/9 fdt_support: delete unnecessary DECLARE_GLOBAL_DATA_PTR
Quoted 3 lines · Masahiro Yamada
gd->bdis not used infdt_support.c.Signed-off-by Masahiro Yamada <[email protected]>
PATCH v2 1/9 fdt_support: delete unnecessary DECLARE_GLOBAL_DATA_PTR
In reply to Masahiro YamadaQuoted 4 lines · Masahiro Yamada
gd->bdis not used infdt_support.c.Signed-off-by Masahiro Yamada <[email protected]>Acked-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140619/fe57f513/attachment.pgp>
PATCH v2 2/9 fdt_support: refactor with fdt_find_or_add_subnode helper func
In reply to Masahiro YamadaSome functions in fdt_support.c do the same routine:
search a node with a given name ("chosen", "memory", etc.) or newly create it if it does not exist.
So this commit makes that routine to a helper function.
---
Changes in v2: None
common/fdt_support.c | 71 ++++++++++++++++++++++++++-------------------------- 1 file changed, 36 insertions(+), 35 deletions(-)
common/fdt_support.c +36 -35
@@ -98,6 +98,31 @@ int fdt_find_and_setprop(void *fdt, const char *node, const char *prop,return fdt_setprop (fdt, nodeoff, prop, val, len); } +/** +* fdt_find_or_add_subnode - find or possibly add a subnode of a given node +* @fdt: pointer to the device tree blob +* @parentoffset: structure block offset of a node +* @name: name of the subnode to locate +* +* fdt_subnode_offset() finds a subnode of the node with a given name. +* If the subnode does not exist, it will be created. +*/ +static int fdt_find_or_add_subnode (void *fdt,int parentoffset, +const char *name) +{ +int offset; + + offset =fdt_subnode_offset (fdt, parentoffset, name); + +if (offset == -FDT_ERR_NOTFOUND ) + offset =fdt_add_subnode (fdt, parentoffset, name); + +if (offset <0 ) +printf ("%s: %s: %s\n" , __func__, name,fdt_strerror (offset)); + +return offset; +} +#ifdef CONFIG_OF_STDOUT_VIA_ALIAS #ifdef CONFIG_CONS_INDEX
@@ -160,14 +185,10 @@ int fdt_initrd(void *fdt, ulong initrd_start, ulong initrd_end, int force)const char *path;uint64_t addr, size; -/* Find the "chosen" node. */ - nodeoffset =fdt_path_offset (fdt,"/chosen" ); - -/* If there is no "chosen" node in the blob return */ -if (nodeoffset <0 ) { -printf ("fdt_initrd: %s\n" ,fdt_strerror (nodeoffset)); +/* find or create "/chosen" node. */ + nodeoffset =fdt_find_or_add_subnode (fdt,0 ,"chosen" ); +if (nodeoffset <0 )return nodeoffset; - }/* just return if initrd_start/end aren't valid */ if ((initrd_start ==0 ) || (initrd_end ==0 ))
@@ -233,25 +254,10 @@ int fdt_chosen(void *fdt, int force)return err; } -/* -* Find the "chosen" node. -*/ - nodeoffset =fdt_path_offset (fdt,"/chosen" ); - -/* -* If there is no "chosen" node in the blob, create it. -*/ -if (nodeoffset <0 ) { -/* -* Create a new node "/chosen" (offset 0 is root level) -*/ - nodeoffset =fdt_add_subnode (fdt,0 ,"chosen" ); -if (nodeoffset <0 ) { -printf ("WARNING: could not create /chosen %s.\n" , -fdt_strerror (nodeoffset)); -return nodeoffset; - } - } +/* find or create "/chosen" node. */ + nodeoffset =fdt_find_or_add_subnode (fdt,0 ,"chosen" ); +if (nodeoffset <0 ) +return nodeoffset;/* * Create /chosen properites that don't exist in the fdt.
@@ -393,16 +399,11 @@ int fdt_fixup_memory_banks(void *blob, u64 start[], u64 size[], int banks)return err; } -/* update, or add and update /memory node */ - nodeoffset =fdt_path_offset (blob,"/memory" ); -if (nodeoffset <0 ) { - nodeoffset =fdt_add_subnode (blob,0 ,"memory" ); -if (nodeoffset <0 ) { -printf ("WARNING: could not create /memory: %s.\n" , -fdt_strerror (nodeoffset)); +/* find or create "/memory" node. */ + nodeoffset =fdt_find_or_add_subnode (blob,0 ,"memory" ); +if (nodeoffset <0 )return nodeoffset; - } - } + err =fdt_setprop (blob, nodeoffset,"device_type" ,"memory" ,sizeof ("memory" ));if (err <0 ) {
Signature · 1 line
1.8.3.2
PATCH v2 2/9 fdt_support: refactor with fdt_find_or_add_subnode helper func
Hi Masahiro,
Quoted 18 lines · Masahiro Yamada
Some functions in
fdt_support.cdo the same routine:
search a node with a given name ("chosen", "memory", etc.) or newly create it if it does not exist.So this commit makes that routine to a helper function.
Signed-off-by Masahiro Yamada <[email protected]>---
Changes in v2: None
common/fdt_support.c | 71 ++++++++++++++++++++++++++-------------------------- 1 file changed, 36 insertions(+), 35 deletions(-)common/fdt_support.c +0
-0
[snip]
common/fdt_support.c int fdt_initrd(void *fdt, ulong initrd_start, ulong initrd_end, int force) · line 190
Quoted 13 lines
common/fdt_support.c +3
-6@@ -160,14 +185,10 @@ int fdt_initrd(void *fdt, ulong initrd_start, ulong initrd_end, int force)const char *path;uint64_t addr, size; -/* Find the "chosen" node. */ - nodeoffset =fdt_path_offset (fdt,"/chosen" ); - -/* If there is no "chosen" node in the blob return */ -if (nodeoffset <0 ) { -printf ("fdt_initrd: %s\n" ,fdt_strerror (nodeoffset)); +/* find or create "/chosen" node. */ + nodeoffset =fdt_find_or_add_subnode (fdt,0 ,"chosen" ); +if (nodeoffset <0 )
It looks like you are changing the behaviour here, but I think that is fine. We should create it if it is missing.
Regards,
Simon
PATCH v2 2/9 fdt_support: refactor with fdt_find_or_add_subnode helper func
In reply to Masahiro YamadaQuoted 8 lines · Masahiro Yamada
Some functions in
fdt_support.cdo the same routine:
search a node with a given name ("chosen", "memory", etc.) or newly create it if it does not exist.So this commit makes that routine to a helper function.
Signed-off-by Masahiro Yamada <[email protected]>Acked-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140619/57b06db2/attachment.pgp>
PATCH v2 3/9 fdt_support: delete force argument of fdt_initrd()
In reply to Masahiro YamadaAfter all, we have realized "force" argument is completely useless. fdt_initrd() was always called with force = 1.
We should always want to do the same thing
(set appropriate value to the property)
even if the property already exists.
---
Changes in v2:
- Rebase on commit 0f507779ca
arch/microblaze/lib/bootm.c | 2 +- common/cmd_fdt.c | 2 +- common/fdt_support.c | 35 +++++++++++++++-------------------- common/image-fdt.c | 2 +- include/fdt_support.h | 2 +- 5 files changed, 19 insertions(+), 24 deletions(-)
arch/microblaze/lib/bootm.c +1 -1
@@ -58,7 +58,7 @@ int do_bootm_linux(int flag, int argc, char * const argv[],/* fixup the initrd now that we know where it should be */ if (images->rd_start && images->rd_end && of_flat_tree) ret =fdt_initrd (of_flat_tree, images->rd_start, - images->rd_end,1 ); + images->rd_end);if (ret)return 1 ;
common/cmd_fdt.c +1 -1
@@ -582,7 +582,7 @@ static int do_fdt(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) }fdt_chosen (working_fdt,1 ); -fdt_initrd (working_fdt, initrd_start, initrd_end,1 ); +fdt_initrd (working_fdt, initrd_start, initrd_end);#if defined(CONFIG_FIT_SIGNATURE ) } else if (strncmp(argv[1], "che", 3) == 0) {
common/fdt_support.c +15 -20
@@ -177,12 +177,11 @@ static int fdt_fixup_stdout(void *fdt, int chosenoff) }#endif -int fdt_initrd (void *fdt,ulong initrd_start,ulong initrd_end,int force) +int fdt_initrd (void *fdt,ulong initrd_start,ulong initrd_end) {int nodeoffset, addr_cell_len;int err, j, total;fdt64_t tmp; -const char *path;uint64_t addr, size;/* find or create "/chosen" node. */
@@ -216,26 +215,22 @@ int fdt_initrd(void *fdt, ulong initrd_start, ulong initrd_end, int force) addr_cell_len =get_cells_len (fdt,"#address-cells" ); - path =fdt_getprop (fdt, nodeoffset,"linux,initrd-start" ,NULL ); -if ((path ==NULL ) || force) { -write_cell ((u8 *)&tmp, initrd_start, addr_cell_len); - err =fdt_setprop (fdt, nodeoffset, -"linux,initrd-start" , &tmp, addr_cell_len); -if (err <0 ) { -printf ("WARNING: " -"could not set linux,initrd-start %s.\n" , -fdt_strerror (err)); -return err; - } -write_cell ((u8 *)&tmp, initrd_end, addr_cell_len); - err =fdt_setprop (fdt, nodeoffset, +write_cell ((u8 *)&tmp, initrd_start, addr_cell_len); + err =fdt_setprop (fdt, nodeoffset, +"linux,initrd-start" , &tmp, addr_cell_len); +if (err <0 ) { +printf ("WARNING: could not set linux,initrd-start %s.\n" , +fdt_strerror (err)); +return err; + } +write_cell ((u8 *)&tmp, initrd_end, addr_cell_len); + err =fdt_setprop (fdt, nodeoffset,"linux,initrd-end" , &tmp, addr_cell_len); -if (err <0 ) { -printf ("WARNING: could not set linux,initrd-end %s.\n" , -fdt_strerror (err)); +if (err <0 ) { +printf ("WARNING: could not set linux,initrd-end %s.\n" , +fdt_strerror (err)); -return err; - } +return err; }return 0 ;
common/image-fdt.c +1 -1
@@ -483,7 +483,7 @@ int image_setup_libfdt(bootm_headers_t *images, void *blob,/* Create a new LMB reservation */ lmb_reserve (lmb, (ulong )blob, of_size); -fdt_initrd (blob, *initrd_start, *initrd_end,1 ); +fdt_initrd (blob, *initrd_start, *initrd_end);if (!ft_verify_fdt (blob))return -1 ;
include/fdt_support.h +1 -1
@@ -15,7 +15,7 @@u32 fdt_getprop_u32_default (const void *fdt,const char *path,const char *prop,const u32 dflt);int fdt_chosen (void *fdt,int force); -int fdt_initrd (void *fdt,ulong initrd_start,ulong initrd_end,int force); +int fdt_initrd (void *fdt,ulong initrd_start,ulong initrd_end);void do_fixup_by_path (void *fdt,const char *path,const char *prop,const void *val,int len,int create); void do_fixup_by_path_u32(void *fdt, const char *path, const char *prop,
Signature · 1 line
1.8.3.2
PATCH v2 3/9 fdt_support: delete force argument of fdt_initrd()
Quoted 8 lines · Masahiro Yamada
After all, we have realized "force" argument is completely useless.
fdt_initrd()was always called with force = 1.We should always want to do the same thing
(set appropriate value to the property)
even if the property already exists.Signed-off-by Masahiro Yamada <[email protected]>
Quoted 5 lines
---
Changes in v2:
- Rebase on commit0f507779ca
PATCH v2 3/9 fdt_support: delete force argument of fdt_initrd()
In reply to Masahiro YamadaQuoted 9 lines · Masahiro Yamada
After all, we have realized "force" argument is completely useless.
fdt_initrd()was always called with force = 1.We should always want to do the same thing
(set appropriate value to the property)
even if the property already exists.Signed-off-by Masahiro Yamada <[email protected]>Acked-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140619/6d1fc0a9/attachment.pgp>
PATCH v2 4/9 fdt_support: delete force argument of fdt_chosen()
In reply to Masahiro YamadaAfter all, we have realized "force" argument is completely useless. fdt_chosen() was always called with force = 1.
We should always want to do the same thing
(set appropriate value to the property)
even if the property already exists.
---
Changes in v2: None
common/cmd_fdt.c | 2 +- common/fdt_support.c | 38 ++++++++++++-------------------------- common/image-fdt.c | 2 +- include/fdt_support.h | 2 +- 4 files changed, 15 insertions(+), 29 deletions(-)
common/cmd_fdt.c +1 -1
@@ -581,7 +581,7 @@ static int do_fdt(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) initrd_end =simple_strtoul (argv[3 ],NULL ,16 ); } -fdt_chosen (working_fdt,1 ); +fdt_chosen (working_fdt);fdt_initrd (working_fdt, initrd_start, initrd_end);#if defined(CONFIG_FIT_SIGNATURE )
common/fdt_support.c +12 -26
@@ -236,12 +236,11 @@ int fdt_initrd(void *fdt, ulong initrd_start, ulong initrd_end)return 0 ; } -int fdt_chosen (void *fdt,int force) +int fdt_chosen (void *fdt) {int nodeoffset;int err;char *str;/* used to set string properties */ -const char *path; err =fdt_check_header (fdt);if (err <0 ) {
@@ -254,38 +253,25 @@ int fdt_chosen(void *fdt, int force)if (nodeoffset <0 )return nodeoffset; -/* -* Create /chosen properites that don't exist in the fdt. -* If the property exists, update it only if the "force" parameter -* is true. -*/ str =getenv ("bootargs" );if (str !=NULL ) { - path =fdt_getprop (fdt, nodeoffset,"bootargs" ,NULL ); -if ((path ==NULL ) || force) { - err =fdt_setprop (fdt, nodeoffset, -"bootargs" , str,strlen (str)+1 ); -if (err <0 ) -printf ("WARNING: could not set bootargs %s.\n" , -fdt_strerror (err)); - } + err =fdt_setprop (fdt, nodeoffset, +"bootargs" , str,strlen (str)+1 ); +if (err <0 ) +printf ("WARNING: could not set bootargs %s.\n" , +fdt_strerror (err)); }#ifdef CONFIG_OF_STDOUT_VIA_ALIAS - path =fdt_getprop (fdt, nodeoffset,"linux,stdout-path" ,NULL ); -if ((path ==NULL ) || force) - err =fdt_fixup_stdout (fdt, nodeoffset); + err =fdt_fixup_stdout (fdt, nodeoffset);#endif #ifdef OF_STDOUT_PATH - path =fdt_getprop (fdt, nodeoffset,"linux,stdout-path" ,NULL ); -if ((path ==NULL ) || force) { - err =fdt_setprop (fdt, nodeoffset, -"linux,stdout-path" ,OF_STDOUT_PATH ,strlen (OF_STDOUT_PATH )+1 ); -if (err <0 ) -printf ("WARNING: could not set linux,stdout-path %s.\n" , -fdt_strerror (err)); - } + err =fdt_setprop (fdt, nodeoffset,"linux,stdout-path" , +OF_STDOUT_PATH ,strlen (OF_STDOUT_PATH )+1 ); +if (err <0 ) +printf ("WARNING: could not set linux,stdout-path %s.\n" , +fdt_strerror (err));#endif return err;
common/image-fdt.c +1 -1
@@ -457,7 +457,7 @@ int image_setup_libfdt(bootm_headers_t *images, void *blob,ulong *initrd_end = &images->initrd_end;int ret; -if (fdt_chosen (blob,1 ) <0 ) { +if (fdt_chosen (blob) <0 ) {puts ("ERROR: /chosen node create failed" );puts (" - must RESET the board to recover.\n" );return -1 ;
include/fdt_support.h +1 -1
@@ -14,7 +14,7 @@u32 fdt_getprop_u32_default (const void *fdt,const char *path,const char *prop,const u32 dflt); -int fdt_chosen (void *fdt,int force); +int fdt_chosen (void *fdt);int fdt_initrd (void *fdt,ulong initrd_start,ulong initrd_end);void do_fixup_by_path (void *fdt,const char *path,const char *prop,const void *val,int len,int create);
Signature · 1 line
1.8.3.2
PATCH v2 4/9 fdt_support: delete force argument of fdt_chosen()
Quoted 8 lines · Masahiro Yamada
After all, we have realized "force" argument is completely useless.
fdt_chosen()was always called with force = 1.We should always want to do the same thing
(set appropriate value to the property)
even if the property already exists.Signed-off-by Masahiro Yamada <[email protected]>
Quoted 9 lines
---
Changes in v2: None
common/cmd_fdt.c | 2 +- common/fdt_support.c | 38 ++++++++++++-------------------------- common/image-fdt.c | 2 +- include/fdt_support.h | 2 +- 4 files changed, 15 insertions(+), 29 deletions(-)
PATCH v2 4/9 fdt_support: delete force argument of fdt_chosen()
In reply to Masahiro YamadaQuoted 9 lines · Masahiro Yamada
After all, we have realized "force" argument is completely useless.
fdt_chosen()was always called with force = 1.We should always want to do the same thing
(set appropriate value to the property)
even if the property already exists.Signed-off-by Masahiro Yamada <[email protected]>Acked-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140619/8776cb06/attachment.pgp>
PATCH v2 5/9 fdt_support: refactor fdt_fixup_stdout() function
In reply to Masahiro Yamada - Do not use a deep indentation. We have only 80-character on each line and 1 indentation consumes 8 spaces. Before the code moves far to the right, you should consider to fix your code. See Linux Documentation/CodingStyle.
- Add CONFIG_OF_STDOUT_VIA_ALIAS and OF_STDOUT_PATH macros only to their definition. Do not add them to both callee and caller. This is a tip to avoid using #ifdef everywhere.
- OF_STDOUT_PATH and CONFIG_OF_STDOUT_VIA_ALIAS are exclusive. If both are defined, the former takes precedence. Do not try to fix-up "linux,stdout-path" property twice.
---
Changes in v2: None
common/fdt_support.c | 85 ++++++++++++++++++++++++++-------------------------- 1 file changed, 42 insertions(+), 43 deletions(-)
common/fdt_support.c +42 -43
@@ -123,9 +123,14 @@ static int fdt_find_or_add_subnode(void *fdt, int parentoffset,return offset; } -#ifdef CONFIG_OF_STDOUT_VIA_ALIAS - -#ifdef CONFIG_CONS_INDEX +/* rename to CONFIG_OF_STDOUT_PATH ? */ +#if defined(OF_STDOUT_PATH ) +static int fdt_fixup_stdout (void *fdt,int chosenoff) +{ +return fdt_setprop (fdt, chosenoff,"linux,stdout-path" , +OF_STDOUT_PATH ,strlen (OF_STDOUT_PATH ) +1 ); +} +#elif defined(CONFIG_OF_STDOUT_VIA_ALIAS ) && defined(CONFIG_CONS_INDEX )static void fdt_fill_multisername (char *sername,size_t maxlen) {const char *outname = stdio_devices[stdout]->name;
@@ -137,44 +142,48 @@ static void fdt_fill_multisername(char *sername, size_t maxlen)if (strcmp (outname +1 ,"serial" ) >0 )strncpy (sername, outname +1 , maxlen); } -#endif static int fdt_fixup_stdout (void *fdt,int chosenoff) { -int err =0 ; -#ifdef CONFIG_CONS_INDEX -int node; +int err; +int aliasoff;char sername[9 ] = {0 }; -const char *path; +const void *path; +int len; +char tmp[256 ];/* long enough */ fdt_fill_multisername (sername,sizeof (sername) -1 );if (!sername[0 ])sprintf (sername,"serial%d" ,CONFIG_CONS_INDEX -1 ); - err = node =fdt_path_offset (fdt,"/aliases" ); -if (node >=0 ) { -int len; - path =fdt_getprop (fdt, node, sername, &len); -if (path) { -char *p =malloc (len); - err = -FDT_ERR_NOSPACE ; -if (p) { -memcpy (p, path, len); - err =fdt_setprop (fdt, chosenoff, -"linux,stdout-path" , p, len); -free (p); - } - }else { - err = len; - } + aliasoff =fdt_path_offset (fdt,"/aliases" ); +if (aliasoff <0 ) { + err = aliasoff; + goto error; } -#endif + + path =fdt_getprop (fdt, aliasoff, sername, &len); +if (!path) { + err = len; + goto error; + } + +/* fdt_setprop may break "path" so we copy it to tmp buffer */ +memcpy (tmp, path, len); + + err =fdt_setprop (fdt, chosenoff,"linux,stdout-path" , tmp, len); +error:if (err <0 )printf ("WARNING: could not set linux,stdout-path %s.\n" , -fdt_strerror (err)); +fdt_strerror (err));return err; } +#else +static int fdt_fixup_stdout (void *fdt,int chosenoff) +{ +return 0 ; +}#endif int fdt_initrd (void *fdt,ulong initrd_start,ulong initrd_end)
@@ -254,27 +263,17 @@ int fdt_chosen(void *fdt)return nodeoffset; str =getenv ("bootargs" ); -if (str !=NULL ) { - err =fdt_setprop (fdt, nodeoffset, -"bootargs" , str,strlen (str)+1 ); -if (err <0 ) +if (str) { + err =fdt_setprop (fdt, nodeoffset,"bootargs" , str, +strlen (str) +1 ); +if (err <0 ) {printf ("WARNING: could not set bootargs %s.\n" ,fdt_strerror (err)); +return err; + } } -#ifdef CONFIG_OF_STDOUT_VIA_ALIAS - err =fdt_fixup_stdout (fdt, nodeoffset); -#endif - -#ifdef OF_STDOUT_PATH - err =fdt_setprop (fdt, nodeoffset,"linux,stdout-path" , -OF_STDOUT_PATH ,strlen (OF_STDOUT_PATH )+1 ); -if (err <0 ) -printf ("WARNING: could not set linux,stdout-path %s.\n" , -fdt_strerror (err)); -#endif - -return err; +return fdt_fixup_stdout (fdt, nodeoffset); } void do_fixup_by_path(void *fdt, const char *path, const char *prop,
Signature · 1 line
1.8.3.2
PATCH v2 5/9 fdt_support: refactor fdt_fixup_stdout() function
Quoted 10 lines · Masahiro Yamada
- Do not use a deep indentation. We have only 80-character on each line and 1 indentation consumes 8 spaces. Before the code moves far to the right, you should consider to fix your code. See Linux
Documentation/CodingStyle.- Add
CONFIG_OF_STDOUT_VIA_ALIASandOF_STDOUT_PATHmacros only to their definition. Do not add them to both callee and caller. This is a tip to avoid using #ifdef everywhere.
Quoted 5 lines
-
OF_STDOUT_PATHandCONFIG_OF_STDOUT_VIA_ALIASare exclusive. If both are defined, the former takes precedence. Do not try to fix-up "linux,stdout-path" property twice.Signed-off-by Masahiro Yamada <[email protected]>
Quoted 6 lines
---
Changes in v2: None
common/fdt_support.c | 85 ++++++++++++++++++++++++++-------------------------- 1 file changed, 42 insertions(+), 43 deletions(-)
PATCH v2 5/9 fdt_support: refactor fdt_fixup_stdout() function
In reply to Masahiro YamadaQuoted 10 lines · Masahiro Yamada
- Do not use a deep indentation. We have only 80-character on each line and 1 indentation consumes 8 spaces. Before the code moves far to the right, you should consider to fix your code. See Linux
Documentation/CodingStyle.- Add
CONFIG_OF_STDOUT_VIA_ALIASandOF_STDOUT_PATHmacros only to their definition. Do not add them to both callee and caller. This is a tip to avoid using #ifdef everywhere.
Quoted 6 lines
-
OF_STDOUT_PATHandCONFIG_OF_STDOUT_VIA_ALIASare exclusive. If both are defined, the former takes precedence. Do not try to fix-up "linux,stdout-path" property twice.Signed-off-by Masahiro Yamada <[email protected]>Acked-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140619/63e920eb/attachment.pgp>
PATCH v2 6/9 fdt_support: add 'const' qualifier for unchanged argument
In reply to Masahiro YamadaIn the next commit, I will add a new function, fdt_pack_reg() which uses get_cells_len().
Beforehand, this commit adds 'const' qualifier to get_cells_len(). Otherwise, a warning message will appear:
warning: passing argument 1 of 'get_cells_len' discards 'const' qualifier from pointer target type [enabled by default]
---
Changes in v2:
- Drop the period at the end of the commit subject
common/fdt_support.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
common/fdt_support.c +2 -2
@@ -21,11 +21,11 @@* if #NNNN-cells property is 2 then len is 8 * otherwise len is 4 */ -static int get_cells_len (void *blob,char *nr_cells_name) +static int get_cells_len (const void *fdt,const char *nr_cells_name) {const fdt32_t *cell; - cell =fdt_getprop (blob,0 , nr_cells_name,NULL ); + cell =fdt_getprop (fdt,0 , nr_cells_name,NULL );if (cell &&fdt32_to_cpu (*cell) ==2 )return 8 ;
Signature · 1 line
1.8.3.2
PATCH v2 6/9 fdt_support: add 'const' qualifier for unchanged argument
Quoted 9 lines · Masahiro Yamada
In the next commit, I will add a new function,
fdt_pack_reg()which usesget_cells_len().Beforehand, this commit adds 'const' qualifier to
get_cells_len(). Otherwise, a warning message will appear:
warning: passing argument 1 of 'get_cells_len' discards 'const' qualifier from pointer target type [enabled by default]Signed-off-by Masahiro Yamada <[email protected]>
Quoted 5 lines
---
Changes in v2:
- Drop the period at the end of the commit subject
PATCH v2 6/9 fdt_support: add 'const' qualifier for unchanged argument
In reply to Masahiro YamadaQuoted 10 lines · Masahiro Yamada
In the next commit, I will add a new function,
fdt_pack_reg()which usesget_cells_len().Beforehand, this commit adds 'const' qualifier to
get_cells_len(). Otherwise, a warning message will appear:
warning: passing argument 1 of 'get_cells_len' discards 'const' qualifier from pointer target type [enabled by default]Signed-off-by Masahiro Yamada <[email protected]>Acked-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140619/54d66f27/attachment.pgp>
PATCH v2 7/9 fdt_support: fix an endian bug of fdt_fixup_memory_banks
In reply to Masahiro YamadaData written to DTB must be converted to big endian order. It is usually done by using cpu_to_fdt32(), cpu_to_fdt64(), etc.
fdt_fixup_memory_banks() invoked write_cell(), which always swaps byte order.
It means the function only worked on little endian architectures.
This commit adds and uses a new helper function, fdt_pack_reg(), which works on both big endian and little endian architrectures.
---
Changes in v2: None
common/fdt_support.c | 42 ++++++++++++++++++++++++++++++------------ 1 file changed, 30 insertions(+), 12 deletions(-)
common/fdt_support.c +30 -12
@@ -354,6 +354,34 @@ void do_fixup_by_compat_u32(void *fdt, const char *compat,do_fixup_by_compat (fdt, compat, prop, &tmp,4 , create); } +/* +* fdt_pack_reg - pack address and size array into the "reg"-suitable stream +*/ +static int fdt_pack_reg (const void *fdt,void *buf,uint64_t *address, +uint64_t *size,int n) +{ +int i; +int address_len =get_cells_len (fdt,"#address-cells" ); +int size_len =get_cells_len (fdt,"#size-cells" ); +char *p = buf; + +for (i =0 ; i < n; i++) { +if (address_len ==8 ) + *(fdt64_t *)p =cpu_to_fdt64 (address[i]); +else + *(fdt32_t *)p =cpu_to_fdt32 (address[i]); + p += address_len; + +if (size_len ==8 ) + *(fdt64_t *)p =cpu_to_fdt64 (size[i]); +else + *(fdt32_t *)p =cpu_to_fdt32 (size[i]); + p += size_len; + } + +return p - (char *)buf; +} +#ifdef CONFIG_NR_DRAM_BANKS #define MEMORY_BANKS_MAX CONFIG_NR_DRAM_BANKS#else
@@ -362,9 +390,8 @@ void do_fixup_by_compat_u32(void *fdt, const char *compat,int fdt_fixup_memory_banks (void *blob,u64 start[],u64 size[],int banks) {int err, nodeoffset; -int addr_cell_len, size_cell_len, len; +int len;u8 tmp[MEMORY_BANKS_MAX *16 ];/* Up to 64-bit address + 64-bit size */ -int bank;if (banks >MEMORY_BANKS_MAX ) { printf("%s: num banks %d exceeds hardcoded limit %d."
@@ -392,16 +419,7 @@ int fdt_fixup_memory_banks(void *blob, u64 start[], u64 size[], int banks)return err; } - addr_cell_len =get_cells_len (blob,"#address-cells" ); - size_cell_len =get_cells_len (blob,"#size-cells" ); - -for (bank =0 , len =0 ; bank < banks; bank++) { -write_cell (tmp + len, start[bank], addr_cell_len); - len += addr_cell_len; - -write_cell (tmp + len, size[bank], size_cell_len); - len += size_cell_len; - } + len =fdt_pack_reg (blob, tmp, start, size, banks); err =fdt_setprop (blob, nodeoffset,"reg" , tmp, len);if (err <0 ) {
Signature · 1 line
1.8.3.2
PATCH v2 7/9 fdt_support: fix an endian bug of fdt_fixup_memory_banks
Quoted 11 lines · Masahiro Yamada
Data written to DTB must be converted to big endian order. It is usually done by using
cpu_to_fdt32(),cpu_to_fdt64(), etc.
fdt_fixup_memory_banks()invokedwrite_cell(), which always swaps byte order.
It means the function only worked on little endian architectures.This commit adds and uses a new helper function,
fdt_pack_reg(), which works on both big endian and little endian architrectures.Signed-off-by Masahiro Yamada <[email protected]>
Quoted 7 lines
---
Changes in v2: None
common/fdt_support.c | 42 ++++++++++++++++++++++++++++++------------ 1 file changed, 30 insertions(+), 12 deletions(-)
PATCH v2 7/9 fdt_support: fix an endian bug of fdt_fixup_memory_banks
In reply to Masahiro YamadaQuoted 7 lines · Masahiro Yamada
Data written to DTB must be converted to big endian order. It is usually done by using
cpu_to_fdt32(),cpu_to_fdt64(), etc.
fdt_fixup_memory_banks()invokedwrite_cell(), which always swaps byte order.
It means the function only worked on little endian architectures.
Quoted 5 lines
This commit adds and uses a new helper function,
fdt_pack_reg(), which works on both big endian and little endian architrectures.Signed-off-by Masahiro Yamada <[email protected]>Acked-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140619/5e757612/attachment.pgp>
PATCH v2 8/9 fdt_support: fix an endian bug of fdt_initrd()
In reply to Masahiro YamadaData written to DTB must be converted to big endian order. It is usually done by using cpu_to_fdt32(), cpu_to_fdt64(), etc.
fdt_initrd() invoked write_cell(), which always swaps byte order. It means the function only worked on little endian architectures. (On big endian architectures, the byte order should be kept as it is)
This commit uses cpu_to_fdt32() and cpu_to_fdt64() and deletes write_cell().
---
Changes in v2: None
common/fdt_support.c | 41 ++++++++++++++++++++--------------------- 1 file changed, 20 insertions(+), 21 deletions(-)
common/fdt_support.c +20 -21
@@ -32,18 +32,6 @@ static int get_cells_len(const void *fdt, const char *nr_cells_name)return 4 ; } -/* -* Write a 4 or 8 byte big endian cell -*/ -static void write_cell (u8 *addr,u64 val,int size) -{ -int shift = (size -1 ) *8 ; -while (size-- >0 ) { - *addr++ = (val >> shift) &0xff ; - shift -=8 ; - } -} -/** * fdt_getprop_u32_default - Find a node and return it's property or a default *
@@ -186,11 +174,21 @@ static int fdt_fixup_stdout(void *fdt, int chosenoff) }#endif +static inline int fdt_setprop_uxx (void *fdt,int nodeoffset,const char *name, +uint64_t val,int is_u64) +{ +if (is_u64) +return fdt_setprop_u64 (fdt, nodeoffset, name, val); +else +return fdt_setprop_u32 (fdt, nodeoffset, name, (uint32_t )val); +} + +int fdt_initrd (void *fdt,ulong initrd_start,ulong initrd_end) { -int nodeoffset, addr_cell_len; +int nodeoffset;int err, j, total; -fdt64_t tmp; +int is_u64;uint64_t addr, size;/* find or create "/chosen" node. */
@@ -222,19 +220,20 @@ int fdt_initrd(void *fdt, ulong initrd_start, ulong initrd_end)return err; } - addr_cell_len =get_cells_len (fdt,"#address-cells" ); + is_u64 = (get_cells_len (fdt,"#address-cells" ) ==8 ); + + err =fdt_setprop_uxx (fdt, nodeoffset,"linux,initrd-start" , + (uint64_t )initrd_start, is_u64); -write_cell ((u8 *)&tmp, initrd_start, addr_cell_len); - err =fdt_setprop (fdt, nodeoffset, -"linux,initrd-start" , &tmp, addr_cell_len);if (err <0 ) {printf ("WARNING: could not set linux,initrd-start %s.\n" ,fdt_strerror (err));return err; } -write_cell ((u8 *)&tmp, initrd_end, addr_cell_len); - err =fdt_setprop (fdt, nodeoffset, -"linux,initrd-end" , &tmp, addr_cell_len); + + err =fdt_setprop_uxx (fdt, nodeoffset,"linux,initrd-end" , + (uint64_t )initrd_end, is_u64); +if (err <0 ) {printf ("WARNING: could not set linux,initrd-end %s.\n" ,fdt_strerror (err));
Signature · 1 line
1.8.3.2
PATCH v2 8/9 fdt_support: fix an endian bug of fdt_initrd()
Quoted 11 lines · Masahiro Yamada
Data written to DTB must be converted to big endian order. It is usually done by using
cpu_to_fdt32(),cpu_to_fdt64(), etc.
fdt_initrd()invokedwrite_cell(), which always swaps byte order. It means the function only worked on little endian architectures. (On big endian architectures, the byte order should be kept as it is)This commit uses
cpu_to_fdt32()andcpu_to_fdt64()and deleteswrite_cell().Signed-off-by Masahiro Yamada <[email protected]>
Quoted 7 lines
---
Changes in v2: None
common/fdt_support.c | 41 ++++++++++++++++++++--------------------- 1 file changed, 20 insertions(+), 21 deletions(-)
PATCH v2 8/9 fdt_support: fix an endian bug of fdt_initrd()
In reply to Masahiro YamadaQuoted 7 lines · Masahiro Yamada
Data written to DTB must be converted to big endian order. It is usually done by using
cpu_to_fdt32(),cpu_to_fdt64(), etc.
fdt_initrd()invokedwrite_cell(), which always swaps byte order. It means the function only worked on little endian architectures. (On big endian architectures, the byte order should be kept as it is)
Quoted 5 lines
This commit uses
cpu_to_fdt32()andcpu_to_fdt64()and deleteswrite_cell().Signed-off-by Masahiro Yamada <[email protected]>Acked-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140619/983ff048/attachment.pgp>
PATCH v2 9/9 fdt_support: correct the return condition of fdt_initrd()
In reply to Masahiro YamadaBefore this commit, fdt_initrd() just returned if initrd start address is zero.
But it is possible if the RAM is located at address 0.
This commit makes the return condition more reasonable:
Just return if the size of initrd is zero.
---
Changes in v2: None
common/fdt_support.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
common/fdt_support.c +4 -4
@@ -191,15 +191,15 @@ int fdt_initrd(void *fdt, ulong initrd_start, ulong initrd_end)int is_u64;uint64_t addr, size; +/* just return if the size of initrd is zero */ +if (initrd_start == initrd_end) +return 0 ; +/* find or create "/chosen" node. */ nodeoffset =fdt_find_or_add_subnode (fdt,0 ,"chosen" );if (nodeoffset <0 )return nodeoffset; -/* just return if initrd_start/end aren't valid */ -if ((initrd_start ==0 ) || (initrd_end ==0 )) -return 0 ; - total =fdt_num_mem_rsv (fdt);/*
Signature · 1 line
1.8.3.2
PATCH v2 9/9 fdt_support: correct the return condition of fdt_initrd()
Quoted 8 lines · Masahiro Yamada
Before this commit,
fdt_initrd()just returned if initrd start address is zero.
But it is possible if the RAM is located at address 0.This commit makes the return condition more reasonable:
Just return if the size of initrd is zero.Signed-off-by Masahiro Yamada <[email protected]>
Quoted 7 lines
---
Changes in v2: None
common/fdt_support.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
PATCH v2 9/9 fdt_support: correct the return condition of fdt_initrd()
In reply to Masahiro YamadaQuoted 9 lines · Masahiro Yamada
Before this commit,
fdt_initrd()just returned if initrd start address is zero.
But it is possible if the RAM is located at address 0.This commit makes the return condition more reasonable:
Just return if the size of initrd is zero.Signed-off-by Masahiro Yamada <[email protected]>Acked-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140619/8ec0bac1/attachment.pgp>
v2 cover Refactoring and Endian bug fixes of fdt_support
In reply to Masahiro YamadaHi Jerry Van Baren,
Quoted 13 lines · On Fri, 18 Apr 2014 17:40:56 +0900 Masahiro Yamada <[email protected]>…
Masahiro Yamada (9):
fdt_support: delete unnecessaryDECLARE_GLOBAL_DATA_PTR
fdt_support: refactor withfdt_find_or_add_subnodehelper func
fdt_support: delete force argument offdt_initrd()
fdt_support: delete force argument offdt_chosen()
fdt_support: refactorfdt_fixup_stdout()function
fdt_support: add 'const' qualifier for unchanged argument
fdt_support: fix an endian bug offdt_fixup_memory_banks
fdt_support: fix an endian bug offdt_initrd()
fdt_support: correct the return condition offdt_initrd()
I think this series is assigned to you.
Could you review it?
More than two months have passed since I posted v1. It was dismissed long time and finally it got not applicable.
I rebased it and posted v2 but still no comments so far.
Best Regards
Masahiro Yamada
v2 cover Refactoring and Endian bug fixes of fdt_support
Hi Jerry Van Baren, Tom
This is the third time I request the review
of my series.
This series was delegated to Jerry Van Baren
and nothing's happened for almost 3 months.
Is he too busy? Or not an active developer any more?
Tom, could you review (and if it looks good, apply this series directly into u-boot/master?)
Some FDT-related patches are delegated to Jerry Van Baren and some are to other custodians.
For example,
http://patchwork.ozlabs.org/patch/338543/
http://patchwork.ozlabs.org/patch/338542/
were delegated to York Sun and applied immediately.
But my patchset has still been dismissed. I'm sad.
Best Regards
Masahiro Yamada
Quoted 37 lines · On Thu, 24 Apr 2014 10:03:35 +0900 Masahiro Yamada <[email protected]>…
Hi Jerry Van Baren,
Quoted 13 lines · On Fri, 18 Apr 2014 17:40:56 +0900 Masahiro Yamada <[email protected]>…
Masahiro Yamada (9):
fdt_support: delete unnecessaryDECLARE_GLOBAL_DATA_PTR
fdt_support: refactor withfdt_find_or_add_subnodehelper func
fdt_support: delete force argument offdt_initrd()
fdt_support: delete force argument offdt_chosen()
fdt_support: refactorfdt_fixup_stdout()function
fdt_support: add 'const' qualifier for unchanged argument
fdt_support: fix an endian bug offdt_fixup_memory_banks
fdt_support: fix an endian bug offdt_initrd()
fdt_support: correct the return condition offdt_initrd()I think this series is assigned to you.
Could you review it?More than two months have passed since I posted v1. It was dismissed long time and finally it got not applicable.
I rebased it and posted v2 but still no comments so far.
Best Regards
Masahiro Yamada_______________________________________________
U-Boot mailing list
U-Boot at lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
Signature · 7 lines
????????? ????LSI???????? ?????????? ?????? ??????? ????? <[email protected]> ?617-8520 ???????????1?? ?? : 050-3783-5420 ?? : 7-664-2813, ?????: 664-772