| Message ID | 20170203160939.27594-2-heiko@sntech.de |
|---|---|
| State | Superseded |
| Delegated to: | Simon Glass |
| Headers | show |
On 3 February 2017 at 08:09, Heiko Stuebner <heiko@sntech.de> wrote: > Right now the u-boot,dm-pre-reloc flag will make each marked node > always appear in both spl and tpl. But systems needing an additional > tpl might have special constraints for each, like the spl needing to > be very tiny. > > So introduce two additional flags to mark nodes for only spl or tpl > environments and introduce a function dm_fdt_pre_reloc to automate > the necessary checks in code instances checking for pre-relocation > flags. > > The behaviour of the original flag stays untouched and still marks > a node for both spl and tpl. > > Signed-off-by: Heiko Stuebner <heiko@sntech.de> > --- > doc/driver-model/README.txt | 4 ++++ > drivers/clk/at91/pmc.c | 3 ++- > drivers/core/root.c | 2 +- > drivers/core/util.c | 29 +++++++++++++++++++++++++++++ > drivers/pinctrl/pinctrl-uclass.c | 3 ++- > include/dm/util.h | 2 ++ > scripts/Makefile.spl | 7 ++++++- > tools/dtoc/dtoc.py | 2 ++ > 8 files changed, 48 insertions(+), 4 deletions(-) Reviewed-by: Simon Glass <sjg@chromium.org> s/u-boot/U-Boot Please add a comment for dm_fdt_pre_reloc() in the header file. Two things to consider: - Should we drop the use of u-boot,dm-pre-reloc in Makefile.spl, and convert all users to your version? This would mean having both u-boot,dm-pre-reloc and u-boot,dm-spl in some cases, I suspect. - Can you use #ifdef in SPL/TPL to reduce code size fractionally in dm_fdt_pre_reloc()? Regards, Simon
Hi Simon, Am Montag, 6. Februar 2017, 07:34:54 CET schrieb Simon Glass: > On 3 February 2017 at 08:09, Heiko Stuebner <heiko@sntech.de> wrote: > > Right now the u-boot,dm-pre-reloc flag will make each marked node > > always appear in both spl and tpl. But systems needing an additional > > tpl might have special constraints for each, like the spl needing to > > be very tiny. > > > > So introduce two additional flags to mark nodes for only spl or tpl > > environments and introduce a function dm_fdt_pre_reloc to automate > > the necessary checks in code instances checking for pre-relocation > > flags. > > > > The behaviour of the original flag stays untouched and still marks > > a node for both spl and tpl. > > > > Signed-off-by: Heiko Stuebner <heiko@sntech.de> > > --- > > > > doc/driver-model/README.txt | 4 ++++ > > drivers/clk/at91/pmc.c | 3 ++- > > drivers/core/root.c | 2 +- > > drivers/core/util.c | 29 +++++++++++++++++++++++++++++ > > drivers/pinctrl/pinctrl-uclass.c | 3 ++- > > include/dm/util.h | 2 ++ > > scripts/Makefile.spl | 7 ++++++- > > tools/dtoc/dtoc.py | 2 ++ > > 8 files changed, 48 insertions(+), 4 deletions(-) > > Reviewed-by: Simon Glass <sjg@chromium.org> > > s/u-boot/U-Boot > > Please add a comment for dm_fdt_pre_reloc() in the header file. > > Two things to consider: > > - Should we drop the use of u-boot,dm-pre-reloc in Makefile.spl, and > convert all users to your version? This would mean having both > u-boot,dm-pre-reloc and u-boot,dm-spl in some cases, I suspect. ^^ u-boot,dm-spl and u-boot,dm-tpl I guess Dropping u-boot,dm-pre-reloc is for you to decide, as you wrote the original implementation ;-) . It will save one fdt-query but might need several iterations as I guess new boards will add new ones after we remove the old ones. > - Can you use #ifdef in SPL/TPL to reduce code size fractionally in > dm_fdt_pre_reloc()? Do you mean moving the + dm_spl = fdt_getprop(blob, offset, "u-boot,dm-spl", NULL); + dm_tpl = fdt_getprop(blob, offset, "u-boot,dm-tpl", NULL); into the existing #ifdef CONFIG_TPL_BUILD / SPL_BUILD blocks or something else? Thanks Heiko
Hi Heiko, On 16 February 2017 at 17:36, Heiko Stübner <heiko@sntech.de> wrote: > > Hi Simon, > > Am Montag, 6. Februar 2017, 07:34:54 CET schrieb Simon Glass: > > On 3 February 2017 at 08:09, Heiko Stuebner <heiko@sntech.de> wrote: > > > Right now the u-boot,dm-pre-reloc flag will make each marked node > > > always appear in both spl and tpl. But systems needing an additional > > > tpl might have special constraints for each, like the spl needing to > > > be very tiny. > > > > > > So introduce two additional flags to mark nodes for only spl or tpl > > > environments and introduce a function dm_fdt_pre_reloc to automate > > > the necessary checks in code instances checking for pre-relocation > > > flags. > > > > > > The behaviour of the original flag stays untouched and still marks > > > a node for both spl and tpl. > > > > > > Signed-off-by: Heiko Stuebner <heiko@sntech.de> > > > --- > > > > > > doc/driver-model/README.txt | 4 ++++ > > > drivers/clk/at91/pmc.c | 3 ++- > > > drivers/core/root.c | 2 +- > > > drivers/core/util.c | 29 +++++++++++++++++++++++++++++ > > > drivers/pinctrl/pinctrl-uclass.c | 3 ++- > > > include/dm/util.h | 2 ++ > > > scripts/Makefile.spl | 7 ++++++- > > > tools/dtoc/dtoc.py | 2 ++ > > > 8 files changed, 48 insertions(+), 4 deletions(-) > > > > Reviewed-by: Simon Glass <sjg@chromium.org> > > > > s/u-boot/U-Boot > > > > Please add a comment for dm_fdt_pre_reloc() in the header file. > > > > Two things to consider: > > > > - Should we drop the use of u-boot,dm-pre-reloc in Makefile.spl, and > > convert all users to your version? This would mean having both > > u-boot,dm-pre-reloc and u-boot,dm-spl in some cases, I suspect. > ^^ u-boot,dm-spl and u-boot,dm-tpl I guess > > Dropping u-boot,dm-pre-reloc is for you to decide, as you wrote the original > implementation ;-) . It will save one fdt-query but might need several > iterations as I guess new boards will add new ones after we remove the old > ones. Let's do it. It is more efficient and I don't expect it to cause many problems if we get it in at the start of the merge window. > > > - Can you use #ifdef in SPL/TPL to reduce code size fractionally in > > dm_fdt_pre_reloc()? > > Do you mean moving the > + dm_spl = fdt_getprop(blob, offset, "u-boot,dm-spl", NULL); > + dm_tpl = fdt_getprop(blob, offset, "u-boot,dm-tpl", NULL); > > into the existing #ifdef CONFIG_TPL_BUILD / SPL_BUILD blocks or something > else? Well I don't think it matters, but the idea is that there would then be no point in checking u-boot,dm-tpl except when CONFIG_TPL_BUILD is defined. Regards, Simon
diff --git a/doc/driver-model/README.txt b/doc/driver-model/README.txt index 1b5ccec4b2..af28b43c7b 100644 --- a/doc/driver-model/README.txt +++ b/doc/driver-model/README.txt @@ -825,6 +825,10 @@ drivers marked with DM_FLAG_PRE_RELOC or the device tree 'u-boot,dm-pre-reloc' flag are initialised prior to relocation. This helps to reduce the driver model overhead. +It is possible to limit this to specific relocation steps, by using +the more specialized 'u-boot,dm-spl' and 'u-boot,dm-tpl' flags +in the devicetree. + Then post relocation we throw that away and re-init driver model again. For drivers which require some sort of continuity between pre- and post-relocation devices, we can provide access to the pre-relocation diff --git a/drivers/clk/at91/pmc.c b/drivers/clk/at91/pmc.c index 76ba91af81..c098efc9f7 100644 --- a/drivers/clk/at91/pmc.c +++ b/drivers/clk/at91/pmc.c @@ -10,6 +10,7 @@ #include <dm/device.h> #include <dm/lists.h> #include <dm/root.h> +#include <dm/util.h> #include "pmc.h" DECLARE_GLOBAL_DATA_PTR; @@ -56,7 +57,7 @@ int at91_clk_sub_device_bind(struct udevice *dev, const char *drv_name) offset > 0; offset = fdt_next_subnode(fdt, offset)) { if (pre_reloc_only && - !fdt_getprop(fdt, offset, "u-boot,dm-pre-reloc", NULL)) + !dm_fdt_pre_reloc(fdt, offset)) continue; /* * If this node has "compatible" property, this is not diff --git a/drivers/core/root.c b/drivers/core/root.c index 9edfc1efb6..45f990ba9e 100644 --- a/drivers/core/root.c +++ b/drivers/core/root.c @@ -205,7 +205,7 @@ int dm_scan_fdt_node(struct udevice *parent, const void *blob, int offset, offset > 0; offset = fdt_next_subnode(blob, offset)) { if (pre_reloc_only && - !fdt_getprop(blob, offset, "u-boot,dm-pre-reloc", NULL)) + !dm_fdt_pre_reloc(blob, offset)) continue; if (!fdtdec_get_is_enabled(blob, offset)) { dm_dbg(" - ignoring disabled device\n"); diff --git a/drivers/core/util.c b/drivers/core/util.c index e01dd06d28..a67434b80d 100644 --- a/drivers/core/util.c +++ b/drivers/core/util.c @@ -5,6 +5,7 @@ */ #include <common.h> +#include <libfdt.h> #include <vsprintf.h> void dm_warn(const char *fmt, ...) @@ -35,3 +36,31 @@ int list_count_items(struct list_head *head) return count; } + +int dm_fdt_pre_reloc(const void *blob, int offset) +{ + bool dm_spl, dm_tpl; + + if (fdt_getprop(blob, offset, "u-boot,dm-pre-reloc", NULL)) + return 1; + + dm_spl = fdt_getprop(blob, offset, "u-boot,dm-spl", NULL); + dm_tpl = fdt_getprop(blob, offset, "u-boot,dm-tpl", NULL); + +#ifdef CONFIG_TPL_BUILD + if (dm_tpl) + return 1; +#elif defined(CONFIG_SPL_BUILD) + if (dm_spl) + return 1; +#else + /* + * In regular builds individual spl and tpl handling both + * count as handled pre-relocation for later second init. + */ + if (dm_spl || dm_tpl) + return 1; +#endif + + return 0; +} diff --git a/drivers/pinctrl/pinctrl-uclass.c b/drivers/pinctrl/pinctrl-uclass.c index 02ab9b4afd..9e0122906c 100644 --- a/drivers/pinctrl/pinctrl-uclass.c +++ b/drivers/pinctrl/pinctrl-uclass.c @@ -12,6 +12,7 @@ #include <dm/lists.h> #include <dm/pinctrl.h> #include <dm/uclass.h> +#include <dm/util.h> DECLARE_GLOBAL_DATA_PTR; @@ -131,7 +132,7 @@ static int pinconfig_post_bind(struct udevice *dev) offset > 0; offset = fdt_next_subnode(fdt, offset)) { if (pre_reloc_only && - !fdt_getprop(fdt, offset, "u-boot,dm-pre-reloc", NULL)) + !dm_fdt_pre_reloc(fdt, offset)) continue; /* * If this node has "compatible" property, this is not diff --git a/include/dm/util.h b/include/dm/util.h index 15daa3d19f..38c023ce66 100644 --- a/include/dm/util.h +++ b/include/dm/util.h @@ -48,4 +48,6 @@ static inline void dm_dump_devres(void) } #endif +int dm_fdt_pre_reloc(const void *blob, int offset); + #endif diff --git a/scripts/Makefile.spl b/scripts/Makefile.spl index c962bbca2c..60bd844e48 100644 --- a/scripts/Makefile.spl +++ b/scripts/Makefile.spl @@ -206,8 +206,13 @@ $(obj)/$(SPL_BIN)-pad.bin: $(obj)/$(SPL_BIN) # 'u-boot,dm-pre-reloc' property and thus are not needed by SPL. The second # pass removes various unused properties from the remaining nodes. # The output is typically a much smaller device tree file. +ifeq ($(CONFIG_TPL_BUILD),y) +fdtgrep_props := -b u-boot,dm-pre-reloc -b u-boot,dm-tpl +else +fdtgrep_props := -b u-boot,dm-pre-reloc -b u-boot,dm-spl +endif quiet_cmd_fdtgrep = FDTGREP $@ - cmd_fdtgrep = $(objtree)/tools/fdtgrep -b u-boot,dm-pre-reloc -RT $< \ + cmd_fdtgrep = $(objtree)/tools/fdtgrep $(fdtgrep_props) -RT $< \ -n /chosen -O dtb | \ $(objtree)/tools/fdtgrep -r -O dtb - -o $@ \ $(addprefix -P ,$(subst $\",,$(CONFIG_OF_SPL_REMOVE_PROPS))) diff --git a/tools/dtoc/dtoc.py b/tools/dtoc/dtoc.py index 11050b66f7..232f8ff739 100755 --- a/tools/dtoc/dtoc.py +++ b/tools/dtoc/dtoc.py @@ -30,6 +30,8 @@ PROP_IGNORE_LIST = [ "status", 'phandle', 'u-boot,dm-pre-reloc', + 'u-boot,dm-tpl', + 'u-boot,dm-spl', ] # C type declarations for the tyues we support
Right now the u-boot,dm-pre-reloc flag will make each marked node always appear in both spl and tpl. But systems needing an additional tpl might have special constraints for each, like the spl needing to be very tiny. So introduce two additional flags to mark nodes for only spl or tpl environments and introduce a function dm_fdt_pre_reloc to automate the necessary checks in code instances checking for pre-relocation flags. The behaviour of the original flag stays untouched and still marks a node for both spl and tpl. Signed-off-by: Heiko Stuebner <heiko@sntech.de> --- doc/driver-model/README.txt | 4 ++++ drivers/clk/at91/pmc.c | 3 ++- drivers/core/root.c | 2 +- drivers/core/util.c | 29 +++++++++++++++++++++++++++++ drivers/pinctrl/pinctrl-uclass.c | 3 ++- include/dm/util.h | 2 ++ scripts/Makefile.spl | 7 ++++++- tools/dtoc/dtoc.py | 2 ++ 8 files changed, 48 insertions(+), 4 deletions(-)