* [PATCH] board_f: show_dram_config: Print also real DRAM size
@ 2022-09-11 9:39 Pali Rohár
2022-09-12 13:34 ` Simon Glass
2022-09-18 11:23 ` [PATCH v2] " Pali Rohár
0 siblings, 2 replies; 11+ messages in thread
From: Pali Rohár @ 2022-09-11 9:39 UTC (permalink / raw)
To: u-boot; +Cc: Simon Glass
32-bit U-Boot builds cannot use more than around 2 GB of DDR memory. But on
some platforms/boards it is possible to connect also 4 GB SODIMM DDR memory.
U-Boot currently prints only effective size of RAM which can use, which may
be misleading as somebody would expect that this line prints total size of
connected DDR modules. So change show_dram_config code to prints both real
and effective DRAM size if they are different. If they are same then print
just one number like before. It is possible that effective size is just few
bytes smaller than the real size, so print both numbers only in case
function print_size() prints formats them differently.
Signed-off-by: Pali Rohár <pali@kernel.org>
---
common/board_f.c | 31 ++++++++++++++++++++++++++++++-
1 file changed, 30 insertions(+), 1 deletion(-)
diff --git a/common/board_f.c b/common/board_f.c
index 9e34fbee147e..3131a06db940 100644
--- a/common/board_f.c
+++ b/common/board_f.c
@@ -54,6 +54,7 @@
#include <asm/sections.h>
#include <dm/root.h>
#include <linux/errno.h>
+#include <linux/log2.h>
/*
* Pointer to initial global data area
@@ -213,6 +214,30 @@ static int announce_dram_init(void)
return 0;
}
+/*
+ * Check if the sizes in their natural units written in decimal format with
+ * one fraction number are same.
+ */
+static int sizes_near(unsigned long long size1, unsigned long long size2)
+{
+ unsigned int size1_scale = ilog2(size1) / 10 * 10;
+ unsigned int size1_val = (10 * size1 + ((1ULL << size1_scale) >> 1)) >> size1_scale;
+ unsigned int size2_scale = ilog2(size2) / 10 * 10;
+ unsigned int size2_val = (10 * size2 + ((1ULL << size2_scale) >> 1)) >> size2_scale;
+
+ if (size1_val == 10240) {
+ size1_val = 10;
+ size1_scale += 10;
+ }
+
+ if (size2_val == 10240) {
+ size2_val = 10;
+ size2_scale += 10;
+ }
+
+ return size1_scale == size2_scale && size1_val == size2_val;
+}
+
static int show_dram_config(void)
{
unsigned long long size;
@@ -229,7 +254,11 @@ static int show_dram_config(void)
}
debug("\nDRAM: ");
- print_size(size, "");
+ print_size(gd->ram_size, "");
+ if (!sizes_near(gd->ram_size, size)) {
+ printf(" (effective ");
+ print_size(size, ")");
+ }
board_add_ram_info(0);
putc('\n');
--
2.20.1
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH] board_f: show_dram_config: Print also real DRAM size
2022-09-11 9:39 [PATCH] board_f: show_dram_config: Print also real DRAM size Pali Rohár
@ 2022-09-12 13:34 ` Simon Glass
2022-09-12 18:56 ` Pali Rohár
2022-09-18 11:23 ` [PATCH v2] " Pali Rohár
1 sibling, 1 reply; 11+ messages in thread
From: Simon Glass @ 2022-09-12 13:34 UTC (permalink / raw)
To: Pali Rohár; +Cc: U-Boot Mailing List
Hi Pali,
On Sun, 11 Sept 2022 at 03:39, Pali Rohár <pali@kernel.org> wrote:
>
> 32-bit U-Boot builds cannot use more than around 2 GB of DDR memory. But on
> some platforms/boards it is possible to connect also 4 GB SODIMM DDR memory.
> U-Boot currently prints only effective size of RAM which can use, which may
> be misleading as somebody would expect that this line prints total size of
> connected DDR modules. So change show_dram_config code to prints both real
> and effective DRAM size if they are different. If they are same then print
> just one number like before. It is possible that effective size is just few
> bytes smaller than the real size, so print both numbers only in case
> function print_size() prints formats them differently.
>
> Signed-off-by: Pali Rohár <pali@kernel.org>
> ---
> common/board_f.c | 31 ++++++++++++++++++++++++++++++-
> 1 file changed, 30 insertions(+), 1 deletion(-)
>
> diff --git a/common/board_f.c b/common/board_f.c
> index 9e34fbee147e..3131a06db940 100644
> --- a/common/board_f.c
> +++ b/common/board_f.c
> @@ -54,6 +54,7 @@
> #include <asm/sections.h>
> #include <dm/root.h>
> #include <linux/errno.h>
> +#include <linux/log2.h>
>
> /*
> * Pointer to initial global data area
> @@ -213,6 +214,30 @@ static int announce_dram_init(void)
> return 0;
> }
>
> +/*
> + * Check if the sizes in their natural units written in decimal format with
> + * one fraction number are same.
> + */
> +static int sizes_near(unsigned long long size1, unsigned long long size2)
> +{
> + unsigned int size1_scale = ilog2(size1) / 10 * 10;
> + unsigned int size1_val = (10 * size1 + ((1ULL << size1_scale) >> 1)) >> size1_scale;
> + unsigned int size2_scale = ilog2(size2) / 10 * 10;
> + unsigned int size2_val = (10 * size2 + ((1ULL << size2_scale) >> 1)) >> size2_scale;
Can you put that expression into a function with a comment, etc.? It
is a bit hard to understand.
> +
> + if (size1_val == 10240) {
> + size1_val = 10;
> + size1_scale += 10;
> + }
> +
> + if (size2_val == 10240) {
> + size2_val = 10;
> + size2_scale += 10;
> + }
If you are doing the same thing to each, why bother? It should not
affect the expression below, should it? :
> +
> + return size1_scale == size2_scale && size1_val == size2_val;
> +}
> +
> static int show_dram_config(void)
> {
> unsigned long long size;
> @@ -229,7 +254,11 @@ static int show_dram_config(void)
> }
> debug("\nDRAM: ");
>
> - print_size(size, "");
> + print_size(gd->ram_size, "");
> + if (!sizes_near(gd->ram_size, size)) {
> + printf(" (effective ");
> + print_size(size, ")");
> + }
> board_add_ram_info(0);
> putc('\n');
>
> --
> 2.20.1
>
Can we make this testable somehow? You could put the new code into a
lib/ function, perhaps, and call it from a C unit test in test/lib ?
Regards,
Simon
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] board_f: show_dram_config: Print also real DRAM size
2022-09-12 13:34 ` Simon Glass
@ 2022-09-12 18:56 ` Pali Rohár
2022-09-12 21:58 ` Sean Anderson
0 siblings, 1 reply; 11+ messages in thread
From: Pali Rohár @ 2022-09-12 18:56 UTC (permalink / raw)
To: Simon Glass; +Cc: U-Boot Mailing List
On Monday 12 September 2022 07:34:47 Simon Glass wrote:
> Hi Pali,
>
> On Sun, 11 Sept 2022 at 03:39, Pali Rohár <pali@kernel.org> wrote:
> >
> > 32-bit U-Boot builds cannot use more than around 2 GB of DDR memory. But on
> > some platforms/boards it is possible to connect also 4 GB SODIMM DDR memory.
> > U-Boot currently prints only effective size of RAM which can use, which may
> > be misleading as somebody would expect that this line prints total size of
> > connected DDR modules. So change show_dram_config code to prints both real
> > and effective DRAM size if they are different. If they are same then print
> > just one number like before. It is possible that effective size is just few
> > bytes smaller than the real size, so print both numbers only in case
> > function print_size() prints formats them differently.
> >
> > Signed-off-by: Pali Rohár <pali@kernel.org>
> > ---
> > common/board_f.c | 31 ++++++++++++++++++++++++++++++-
> > 1 file changed, 30 insertions(+), 1 deletion(-)
> >
> > diff --git a/common/board_f.c b/common/board_f.c
> > index 9e34fbee147e..3131a06db940 100644
> > --- a/common/board_f.c
> > +++ b/common/board_f.c
> > @@ -54,6 +54,7 @@
> > #include <asm/sections.h>
> > #include <dm/root.h>
> > #include <linux/errno.h>
> > +#include <linux/log2.h>
> >
> > /*
> > * Pointer to initial global data area
> > @@ -213,6 +214,30 @@ static int announce_dram_init(void)
> > return 0;
> > }
> >
> > +/*
> > + * Check if the sizes in their natural units written in decimal format with
> > + * one fraction number are same.
> > + */
> > +static int sizes_near(unsigned long long size1, unsigned long long size2)
> > +{
> > + unsigned int size1_scale = ilog2(size1) / 10 * 10;
> > + unsigned int size1_val = (10 * size1 + ((1ULL << size1_scale) >> 1)) >> size1_scale;
> > + unsigned int size2_scale = ilog2(size2) / 10 * 10;
> > + unsigned int size2_val = (10 * size2 + ((1ULL << size2_scale) >> 1)) >> size2_scale;
>
> Can you put that expression into a function with a comment, etc.? It
> is a bit hard to understand.
Ok.
> > +
> > + if (size1_val == 10240) {
> > + size1_val = 10;
> > + size1_scale += 10;
> > + }
> > +
> > + if (size2_val == 10240) {
> > + size2_val = 10;
> > + size2_scale += 10;
> > + }
>
> If you are doing the same thing to each, why bother? It should not
> affect the expression below, should it? :
This is interesting question, and the answer it that it is required and
affects comparison expression below. For example for the case when size1
is below 1GB limit, size2 is above 1GB limit and both values are near.
Imagine that size1 is approaching value 1GB from the left and size2 from
the right side.
> > +
> > + return size1_scale == size2_scale && size1_val == size2_val;
> > +}
> > +
> > static int show_dram_config(void)
> > {
> > unsigned long long size;
> > @@ -229,7 +254,11 @@ static int show_dram_config(void)
> > }
> > debug("\nDRAM: ");
> >
> > - print_size(size, "");
> > + print_size(gd->ram_size, "");
> > + if (!sizes_near(gd->ram_size, size)) {
> > + printf(" (effective ");
> > + print_size(size, ")");
> > + }
> > board_add_ram_info(0);
> > putc('\n');
> >
> > --
> > 2.20.1
> >
>
> Can we make this testable somehow? You could put the new code into a
> lib/ function, perhaps, and call it from a C unit test in test/lib ?
>
> Regards,
> Simon
Meh... I do not know how to test such code. Due to size / optimization
requirements it is not a good idea to make function outside of board_f.c
file.
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] board_f: show_dram_config: Print also real DRAM size
2022-09-12 18:56 ` Pali Rohár
@ 2022-09-12 21:58 ` Sean Anderson
2022-09-14 17:10 ` Simon Glass
2022-09-18 11:30 ` Pali Rohár
0 siblings, 2 replies; 11+ messages in thread
From: Sean Anderson @ 2022-09-12 21:58 UTC (permalink / raw)
To: Pali Rohár, Simon Glass; +Cc: U-Boot Mailing List
On 9/12/22 2:56 PM, Pali Rohár wrote:
> On Monday 12 September 2022 07:34:47 Simon Glass wrote:
>> Hi Pali,
>>
>> On Sun, 11 Sept 2022 at 03:39, Pali Rohár <pali@kernel.org> wrote:
>> >
>> > 32-bit U-Boot builds cannot use more than around 2 GB of DDR memory. But on
>> > some platforms/boards it is possible to connect also 4 GB SODIMM DDR memory.
>> > U-Boot currently prints only effective size of RAM which can use, which may
>> > be misleading as somebody would expect that this line prints total size of
>> > connected DDR modules. So change show_dram_config code to prints both real
>> > and effective DRAM size if they are different. If they are same then print
>> > just one number like before. It is possible that effective size is just few
>> > bytes smaller than the real size, so print both numbers only in case
>> > function print_size() prints formats them differently.
>> >
>> > Signed-off-by: Pali Rohár <pali@kernel.org>
>> > ---
>> > common/board_f.c | 31 ++++++++++++++++++++++++++++++-
>> > 1 file changed, 30 insertions(+), 1 deletion(-)
>> >
>> > diff --git a/common/board_f.c b/common/board_f.c
>> > index 9e34fbee147e..3131a06db940 100644
>> > --- a/common/board_f.c
>> > +++ b/common/board_f.c
>> > @@ -54,6 +54,7 @@
>> > #include <asm/sections.h>
>> > #include <dm/root.h>
>> > #include <linux/errno.h>
>> > +#include <linux/log2.h>
>> >
>> > /*
>> > * Pointer to initial global data area
>> > @@ -213,6 +214,30 @@ static int announce_dram_init(void)
>> > return 0;
>> > }
>> >
>> > +/*
>> > + * Check if the sizes in their natural units written in decimal format with
>> > + * one fraction number are same.
>> > + */
>> > +static int sizes_near(unsigned long long size1, unsigned long long size2)
>> > +{
>> > + unsigned int size1_scale = ilog2(size1) / 10 * 10;
>> > + unsigned int size1_val = (10 * size1 + ((1ULL << size1_scale) >> 1)) >> size1_scale;
>> > + unsigned int size2_scale = ilog2(size2) / 10 * 10;
>> > + unsigned int size2_val = (10 * size2 + ((1ULL << size2_scale) >> 1)) >> size2_scale;
>>
>> Can you put that expression into a function with a comment, etc.? It
>> is a bit hard to understand.
>
> Ok.
>
>> > +
>> > + if (size1_val == 10240) {
>> > + size1_val = 10;
>> > + size1_scale += 10;
>> > + }
>> > +
>> > + if (size2_val == 10240) {
>> > + size2_val = 10;
>> > + size2_scale += 10;
>> > + }
>>
>> If you are doing the same thing to each, why bother? It should not
>> affect the expression below, should it? :
>
> This is interesting question, and the answer it that it is required and
> affects comparison expression below. For example for the case when size1
> is below 1GB limit, size2 is above 1GB limit and both values are near.
> Imagine that size1 is approaching value 1GB from the left and size2 from
> the right side.
>
>> > +
>> > + return size1_scale == size2_scale && size1_val == size2_val;
>> > +}
>> > +
>> > static int show_dram_config(void)
>> > {
>> > unsigned long long size;
>> > @@ -229,7 +254,11 @@ static int show_dram_config(void)
>> > }
>> > debug("\nDRAM: ");
>> >
>> > - print_size(size, "");
>> > + print_size(gd->ram_size, "");
>> > + if (!sizes_near(gd->ram_size, size)) {
>> > + printf(" (effective ");
>> > + print_size(size, ")");
>> > + }
>> > board_add_ram_info(0);
>> > putc('\n');
>> >
>> > --
>> > 2.20.1
>> >
>>
>> Can we make this testable somehow? You could put the new code into a
>> lib/ function, perhaps, and call it from a C unit test in test/lib ?
>>
>> Regards,
>> Simon
>
> Meh... I do not know how to test such code. Due to size / optimization
> requirements it is not a good idea to make function outside of board_f.c
> file.
>
You can use TEST_STATIC from test/export.h this case.
--Sean
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] board_f: show_dram_config: Print also real DRAM size
2022-09-12 21:58 ` Sean Anderson
@ 2022-09-14 17:10 ` Simon Glass
2022-09-14 17:32 ` Pali Rohár
2022-09-18 11:30 ` Pali Rohár
1 sibling, 1 reply; 11+ messages in thread
From: Simon Glass @ 2022-09-14 17:10 UTC (permalink / raw)
To: Sean Anderson; +Cc: Pali Rohár, U-Boot Mailing List
Hi Pali,
On Mon, 12 Sept 2022 at 15:58, Sean Anderson <sean.anderson@seco.com> wrote:
>
>
>
> On 9/12/22 2:56 PM, Pali Rohár wrote:
> > On Monday 12 September 2022 07:34:47 Simon Glass wrote:
> >> Hi Pali,
> >>
> >> On Sun, 11 Sept 2022 at 03:39, Pali Rohár <pali@kernel.org> wrote:
> >> >
> >> > 32-bit U-Boot builds cannot use more than around 2 GB of DDR memory. But on
> >> > some platforms/boards it is possible to connect also 4 GB SODIMM DDR memory.
> >> > U-Boot currently prints only effective size of RAM which can use, which may
> >> > be misleading as somebody would expect that this line prints total size of
> >> > connected DDR modules. So change show_dram_config code to prints both real
> >> > and effective DRAM size if they are different. If they are same then print
> >> > just one number like before. It is possible that effective size is just few
> >> > bytes smaller than the real size, so print both numbers only in case
> >> > function print_size() prints formats them differently.
> >> >
> >> > Signed-off-by: Pali Rohár <pali@kernel.org>
> >> > ---
> >> > common/board_f.c | 31 ++++++++++++++++++++++++++++++-
> >> > 1 file changed, 30 insertions(+), 1 deletion(-)
> >> >
> >> > diff --git a/common/board_f.c b/common/board_f.c
> >> > index 9e34fbee147e..3131a06db940 100644
> >> > --- a/common/board_f.c
> >> > +++ b/common/board_f.c
> >> > @@ -54,6 +54,7 @@
> >> > #include <asm/sections.h>
> >> > #include <dm/root.h>
> >> > #include <linux/errno.h>
> >> > +#include <linux/log2.h>
> >> >
> >> > /*
> >> > * Pointer to initial global data area
> >> > @@ -213,6 +214,30 @@ static int announce_dram_init(void)
> >> > return 0;
> >> > }
> >> >
> >> > +/*
> >> > + * Check if the sizes in their natural units written in decimal format with
> >> > + * one fraction number are same.
> >> > + */
> >> > +static int sizes_near(unsigned long long size1, unsigned long long size2)
> >> > +{
> >> > + unsigned int size1_scale = ilog2(size1) / 10 * 10;
> >> > + unsigned int size1_val = (10 * size1 + ((1ULL << size1_scale) >> 1)) >> size1_scale;
> >> > + unsigned int size2_scale = ilog2(size2) / 10 * 10;
> >> > + unsigned int size2_val = (10 * size2 + ((1ULL << size2_scale) >> 1)) >> size2_scale;
> >>
> >> Can you put that expression into a function with a comment, etc.? It
> >> is a bit hard to understand.
> >
> > Ok.
> >
> >> > +
> >> > + if (size1_val == 10240) {
> >> > + size1_val = 10;
> >> > + size1_scale += 10;
> >> > + }
> >> > +
> >> > + if (size2_val == 10240) {
> >> > + size2_val = 10;
> >> > + size2_scale += 10;
> >> > + }
> >>
> >> If you are doing the same thing to each, why bother? It should not
> >> affect the expression below, should it? :
> >
> > This is interesting question, and the answer it that it is required and
> > affects comparison expression below. For example for the case when size1
> > is below 1GB limit, size2 is above 1GB limit and both values are near.
> > Imagine that size1 is approaching value 1GB from the left and size2 from
> > the right side.
> >
> >> > +
> >> > + return size1_scale == size2_scale && size1_val == size2_val;
> >> > +}
> >> > +
> >> > static int show_dram_config(void)
> >> > {
> >> > unsigned long long size;
> >> > @@ -229,7 +254,11 @@ static int show_dram_config(void)
> >> > }
> >> > debug("\nDRAM: ");
> >> >
> >> > - print_size(size, "");
> >> > + print_size(gd->ram_size, "");
> >> > + if (!sizes_near(gd->ram_size, size)) {
> >> > + printf(" (effective ");
> >> > + print_size(size, ")");
> >> > + }
> >> > board_add_ram_info(0);
> >> > putc('\n');
> >> >
> >> > --
> >> > 2.20.1
> >> >
> >>
> >> Can we make this testable somehow? You could put the new code into a
> >> lib/ function, perhaps, and call it from a C unit test in test/lib ?
> >>
> >> Regards,
> >> Simon
> >
> > Meh... I do not know how to test such code. Due to size / optimization
> > requirements it is not a good idea to make function outside of board_f.c
> > file.
> >
>
> You can use TEST_STATIC from test/export.h this case.
Good point, Sean.
Also LTO take away most//all of the cost of making a static function global.
Regards,
Simon
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] board_f: show_dram_config: Print also real DRAM size
2022-09-14 17:10 ` Simon Glass
@ 2022-09-14 17:32 ` Pali Rohár
2022-09-14 18:31 ` Simon Glass
0 siblings, 1 reply; 11+ messages in thread
From: Pali Rohár @ 2022-09-14 17:32 UTC (permalink / raw)
To: Simon Glass; +Cc: Sean Anderson, U-Boot Mailing List
On Wednesday 14 September 2022 11:10:20 Simon Glass wrote:
> LTO take away most//all of the cost of making a static function global.
This is not solution as LTO does not work on powerpc and as I figured
today, it is broken also on some 32-bit ARM platforms (e.g. mvebu).
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] board_f: show_dram_config: Print also real DRAM size
2022-09-14 17:32 ` Pali Rohár
@ 2022-09-14 18:31 ` Simon Glass
0 siblings, 0 replies; 11+ messages in thread
From: Simon Glass @ 2022-09-14 18:31 UTC (permalink / raw)
To: Pali Rohár; +Cc: Sean Anderson, U-Boot Mailing List
Hi Pali,
On Wed, 14 Sept 2022 at 11:32, Pali Rohár <pali@kernel.org> wrote:
>
> On Wednesday 14 September 2022 11:10:20 Simon Glass wrote:
> > LTO take away most//all of the cost of making a static function global.
>
> This is not solution as LTO does not work on powerpc and as I figured
> today, it is broken also on some 32-bit ARM platforms (e.g. mvebu).
Yes, various platforms need debugging of particular issues, as I understand it.
Anyway, Sean's solution should work fine and doesn't need LTO.
- Simon
^ permalink raw reply [flat|nested] 11+ messages in thread
* [PATCH v2] board_f: show_dram_config: Print also real DRAM size
2022-09-11 9:39 [PATCH] board_f: show_dram_config: Print also real DRAM size Pali Rohár
2022-09-12 13:34 ` Simon Glass
@ 2022-09-18 11:23 ` Pali Rohár
2022-09-23 22:48 ` Tom Rini
1 sibling, 1 reply; 11+ messages in thread
From: Pali Rohár @ 2022-09-18 11:23 UTC (permalink / raw)
To: Simon Glass, Sean Anderson; +Cc: u-boot
32-bit U-Boot builds cannot use more than around 2 GB of DDR memory. But on
some platforms/boards it is possible to connect also 4 GB SODIMM DDR memory.
U-Boot currently prints only effective size of RAM which can use, which may
be misleading as somebody would expect that this line prints total size of
connected DDR modules. So change show_dram_config code to prints both real
and effective DRAM size if they are different. If they are same then print
just one number like before. It is possible that effective size is just few
bytes smaller than the real size, so print both numbers only in case
function print_size() prints formats them differently.
Signed-off-by: Pali Rohár <pali@kernel.org>
---
Changes in v2:
* Move calculation code into separate macro and add description of it
---
common/board_f.c | 37 ++++++++++++++++++++++++++++++++++++-
1 file changed, 36 insertions(+), 1 deletion(-)
diff --git a/common/board_f.c b/common/board_f.c
index 9e34fbee147e..88a6dfff03fc 100644
--- a/common/board_f.c
+++ b/common/board_f.c
@@ -54,6 +54,7 @@
#include <asm/sections.h>
#include <dm/root.h>
#include <linux/errno.h>
+#include <linux/log2.h>
/*
* Pointer to initial global data area
@@ -213,6 +214,36 @@ static int announce_dram_init(void)
return 0;
}
+/*
+ * From input size calculate its nearest rounded unit scale (multiply of 2^10)
+ * and value in calculated unit scale multiplied by 10 (as fractional fixed
+ * point number with one decimal digit), which is human natural format,
+ * same what uses print_size() function for displaying. Mathematically it is:
+ * round_nearest(val * 2^scale) = size * 10; where: 10 <= val < 10240.
+ *
+ * For example for size=87654321 we calculate scale=20 and val=836 which means
+ * that input has natural human format 83.6 M (mega = 2^20).
+ */
+#define compute_size_scale_val(size, scale, val) do { \
+ scale = ilog2(size) / 10 * 10; \
+ val = (10 * size + ((1ULL << scale) >> 1)) >> scale; \
+ if (val == 10240) { val = 10; scale += 10; } \
+} while (0)
+
+/*
+ * Check if the sizes in their natural units written in decimal format with
+ * one fraction number are same.
+ */
+static int sizes_near(unsigned long long size1, unsigned long long size2)
+{
+ unsigned int size1_scale, size1_val, size2_scale, size2_val;
+
+ compute_size_scale_val(size1, size1_scale, size1_val);
+ compute_size_scale_val(size2, size2_scale, size2_val);
+
+ return size1_scale == size2_scale && size1_val == size2_val;
+}
+
static int show_dram_config(void)
{
unsigned long long size;
@@ -229,7 +260,11 @@ static int show_dram_config(void)
}
debug("\nDRAM: ");
- print_size(size, "");
+ print_size(gd->ram_size, "");
+ if (!sizes_near(gd->ram_size, size)) {
+ printf(" (effective ");
+ print_size(size, ")");
+ }
board_add_ram_info(0);
putc('\n');
--
2.20.1
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH] board_f: show_dram_config: Print also real DRAM size
2022-09-12 21:58 ` Sean Anderson
2022-09-14 17:10 ` Simon Glass
@ 2022-09-18 11:30 ` Pali Rohár
2022-09-19 21:10 ` Simon Glass
1 sibling, 1 reply; 11+ messages in thread
From: Pali Rohár @ 2022-09-18 11:30 UTC (permalink / raw)
To: Sean Anderson; +Cc: Simon Glass, U-Boot Mailing List
On Monday 12 September 2022 17:58:09 Sean Anderson wrote:
> You can use TEST_STATIC from test/export.h this case.
Now I sent new patch version where I extended comment.
I have still issue with test framework. Could you help me how to write
that unit test and run it? Scenario should really simple, just check
like this:
sizes_near(87654321, 87654320) == true;
sizes_near(87654321, 1000) == false;
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] board_f: show_dram_config: Print also real DRAM size
2022-09-18 11:30 ` Pali Rohár
@ 2022-09-19 21:10 ` Simon Glass
0 siblings, 0 replies; 11+ messages in thread
From: Simon Glass @ 2022-09-19 21:10 UTC (permalink / raw)
To: Pali Rohár; +Cc: Sean Anderson, U-Boot Mailing List
Hi Pali,
On Sun, 18 Sept 2022 at 13:30, Pali Rohár <pali@kernel.org> wrote:
>
> On Monday 12 September 2022 17:58:09 Sean Anderson wrote:
> > You can use TEST_STATIC from test/export.h this case.
>
> Now I sent new patch version where I extended comment.
>
> I have still issue with test framework. Could you help me how to write
> that unit test and run it? Scenario should really simple, just check
> like this:
>
> sizes_near(87654321, 87654320) == true;
> sizes_near(87654321, 1000) == false;
I suggest copying something like test/lib/sscanf.c and then you can
run it in sandbox with 'ut lib <testname>'
Regards,
Simon
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v2] board_f: show_dram_config: Print also real DRAM size
2022-09-18 11:23 ` [PATCH v2] " Pali Rohár
@ 2022-09-23 22:48 ` Tom Rini
0 siblings, 0 replies; 11+ messages in thread
From: Tom Rini @ 2022-09-23 22:48 UTC (permalink / raw)
To: Pali Rohár; +Cc: Simon Glass, Sean Anderson, u-boot
[-- Attachment #1: Type: text/plain, Size: 842 bytes --]
On Sun, Sep 18, 2022 at 01:23:27PM +0200, Pali Rohár wrote:
> 32-bit U-Boot builds cannot use more than around 2 GB of DDR memory. But on
> some platforms/boards it is possible to connect also 4 GB SODIMM DDR memory.
> U-Boot currently prints only effective size of RAM which can use, which may
> be misleading as somebody would expect that this line prints total size of
> connected DDR modules. So change show_dram_config code to prints both real
> and effective DRAM size if they are different. If they are same then print
> just one number like before. It is possible that effective size is just few
> bytes smaller than the real size, so print both numbers only in case
> function print_size() prints formats them differently.
>
> Signed-off-by: Pali Rohár <pali@kernel.org>
Applied to u-boot/next, thanks!
--
Tom
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 659 bytes --]
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2022-09-23 22:49 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-09-11 9:39 [PATCH] board_f: show_dram_config: Print also real DRAM size Pali Rohár
2022-09-12 13:34 ` Simon Glass
2022-09-12 18:56 ` Pali Rohár
2022-09-12 21:58 ` Sean Anderson
2022-09-14 17:10 ` Simon Glass
2022-09-14 17:32 ` Pali Rohár
2022-09-14 18:31 ` Simon Glass
2022-09-18 11:30 ` Pali Rohár
2022-09-19 21:10 ` Simon Glass
2022-09-18 11:23 ` [PATCH v2] " Pali Rohár
2022-09-23 22:48 ` Tom Rini
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.