* [PATCH] imx8m, imx9: Fix DRAM size calculation in SPL/dram_init_banksize()
@ 2023-08-07 10:56 Elena Popa
2023-08-07 15:58 ` [PATCH] imx8m,imx9: " Marek Vasut
0 siblings, 1 reply; 4+ messages in thread
From: Elena Popa @ 2023-08-07 10:56 UTC (permalink / raw)
To: u-boot; +Cc: marex, Elena Popa
If dram_init_banksize() is called from SPL, the rom_pointer, at that
point, is not correctly initialized. This causes wrong calculation of
DRAM start and size in dram_init_banksize(). The issue became apparent
only in Falcon Mode. Added an extra condition to prevent using
rom_pointer in SPL.
Signed-off-by: Elena Popa <elena.popa@nxp.com>
---
arch/arm/mach-imx/imx8m/soc.c | 2 +-
arch/arm/mach-imx/imx9/soc.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/arm/mach-imx/imx8m/soc.c b/arch/arm/mach-imx/imx8m/soc.c
index d5254886be..dbb2154b08 100644
--- a/arch/arm/mach-imx/imx8m/soc.c
+++ b/arch/arm/mach-imx/imx8m/soc.c
@@ -273,7 +273,7 @@ int dram_init_banksize(void)
}
gd->bd->bi_dram[bank].start = PHYS_SDRAM;
- if (!IS_ENABLED(CONFIG_ARMV8_PSCI) && rom_pointer[1]) {
+ if (!IS_ENABLED(CONFIG_ARMV8_PSCI) && !IS_ENABLED(CONFIG_SPL_BUILD) && rom_pointer[1]) {
phys_addr_t optee_start = (phys_addr_t)rom_pointer[0];
phys_size_t optee_size = (size_t)rom_pointer[1];
diff --git a/arch/arm/mach-imx/imx9/soc.c b/arch/arm/mach-imx/imx9/soc.c
index f43b73a6c2..ca91f20fe7 100644
--- a/arch/arm/mach-imx/imx9/soc.c
+++ b/arch/arm/mach-imx/imx9/soc.c
@@ -390,7 +390,7 @@ int dram_init_banksize(void)
}
gd->bd->bi_dram[bank].start = PHYS_SDRAM;
- if (rom_pointer[1]) {
+ if (!IS_ENABLED(CONFIG_SPL_BUILD) && rom_pointer[1]) {
phys_addr_t optee_start = (phys_addr_t)rom_pointer[0];
phys_size_t optee_size = (size_t)rom_pointer[1];
--
2.25.1
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] imx8m,imx9: Fix DRAM size calculation in SPL/dram_init_banksize()
2023-08-07 10:56 [PATCH] imx8m, imx9: Fix DRAM size calculation in SPL/dram_init_banksize() Elena Popa
@ 2023-08-07 15:58 ` Marek Vasut
2023-08-07 16:43 ` [EXT] " Elena Popa
0 siblings, 1 reply; 4+ messages in thread
From: Marek Vasut @ 2023-08-07 15:58 UTC (permalink / raw)
To: Elena Popa, u-boot
Cc: Stefano Babic, Fabio Estevam, Peng Fan, NXP i.MX U-Boot Team
On 8/7/23 12:56, Elena Popa wrote:
Subject tags should be (per git log arch/arm/mach-imx/imx8m/soc.c):
arm: imx: imx8m: imx9: Fix DRAM .....
> If dram_init_banksize() is called from SPL, the rom_pointer, at that
> point, is not correctly initialized. This causes wrong calculation of
> DRAM start and size in dram_init_banksize(). The issue became apparent
> only in Falcon Mode. Added an extra condition to prevent using
> rom_pointer in SPL.
+CC Stefano, Fabio, Peng.
> Signed-off-by: Elena Popa <elena.popa@nxp.com>
> ---
> arch/arm/mach-imx/imx8m/soc.c | 2 +-
> arch/arm/mach-imx/imx9/soc.c | 2 +-
> 2 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/arch/arm/mach-imx/imx8m/soc.c b/arch/arm/mach-imx/imx8m/soc.c
> index d5254886be..dbb2154b08 100644
> --- a/arch/arm/mach-imx/imx8m/soc.c
> +++ b/arch/arm/mach-imx/imx8m/soc.c
> @@ -273,7 +273,7 @@ int dram_init_banksize(void)
> }
>
> gd->bd->bi_dram[bank].start = PHYS_SDRAM;
> - if (!IS_ENABLED(CONFIG_ARMV8_PSCI) && rom_pointer[1]) {
> + if (!IS_ENABLED(CONFIG_ARMV8_PSCI) && !IS_ENABLED(CONFIG_SPL_BUILD) && rom_pointer[1]) {
Use ./scripts/checkpatch.pl , it should indicate line too long here.
There seem to be other places which likely also need similar fix:
$ git grep CONFIG_ARMV8_PSCI.*rom_pointer arch/arm/mach-imx/
arch/arm/mach-imx/imx8m/soc.c: if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
rom_pointer[1])
arch/arm/mach-imx/imx8m/soc.c: if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
rom_pointer[1]) {
arch/arm/mach-imx/imx8m/soc.c: if
(!IS_ENABLED(CONFIG_ARMV8_PSCI) && rom_pointer[1]) {
arch/arm/mach-imx/imx8m/soc.c: if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
rom_pointer[0] &&
> phys_addr_t optee_start = (phys_addr_t)rom_pointer[0];
> phys_size_t optee_size = (size_t)rom_pointer[1];
>
> diff --git a/arch/arm/mach-imx/imx9/soc.c b/arch/arm/mach-imx/imx9/soc.c
> index f43b73a6c2..ca91f20fe7 100644
> --- a/arch/arm/mach-imx/imx9/soc.c
> +++ b/arch/arm/mach-imx/imx9/soc.c
> @@ -390,7 +390,7 @@ int dram_init_banksize(void)
> }
>
> gd->bd->bi_dram[bank].start = PHYS_SDRAM;
> - if (rom_pointer[1]) {
> + if (!IS_ENABLED(CONFIG_SPL_BUILD) && rom_pointer[1]) {
> phys_addr_t optee_start = (phys_addr_t)rom_pointer[0];
> phys_size_t optee_size = (size_t)rom_pointer[1];
[...]
^ permalink raw reply [flat|nested] 4+ messages in thread
* RE: [EXT] Re: [PATCH] imx8m,imx9: Fix DRAM size calculation in SPL/dram_init_banksize()
2023-08-07 15:58 ` [PATCH] imx8m,imx9: " Marek Vasut
@ 2023-08-07 16:43 ` Elena Popa
2023-08-07 21:44 ` Marek Vasut
0 siblings, 1 reply; 4+ messages in thread
From: Elena Popa @ 2023-08-07 16:43 UTC (permalink / raw)
To: Marek Vasut, u-boot; +Cc: Stefano Babic, Fabio Estevam, Peng Fan, dl-uboot-imx
> Subject tags should be (per git log arch/arm/mach-imx/imx8m/soc.c):
>
> arm: imx: imx8m: imx9: Fix DRAM .....
Will do!
> > If dram_init_banksize() is called from SPL, the rom_pointer, at that
> > point, is not correctly initialized. This causes wrong calculation of
> > DRAM start and size in dram_init_banksize(). The issue became apparent
> > only in Falcon Mode. Added an extra condition to prevent using
> > rom_pointer in SPL.
>
> +CC Stefano, Fabio, Peng.
>
> > Signed-off-by: Elena Popa <elena.popa@nxp.com>
> > ---
> > arch/arm/mach-imx/imx8m/soc.c | 2 +-
> > arch/arm/mach-imx/imx9/soc.c | 2 +-
> > 2 files changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/arch/arm/mach-imx/imx8m/soc.c
> > b/arch/arm/mach-imx/imx8m/soc.c index d5254886be..dbb2154b08
> 100644
> > --- a/arch/arm/mach-imx/imx8m/soc.c
> > +++ b/arch/arm/mach-imx/imx8m/soc.c
> > @@ -273,7 +273,7 @@ int dram_init_banksize(void)
> > }
> >
> > gd->bd->bi_dram[bank].start = PHYS_SDRAM;
> > - if (!IS_ENABLED(CONFIG_ARMV8_PSCI) && rom_pointer[1]) {
> > + if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
> > + !IS_ENABLED(CONFIG_SPL_BUILD) && rom_pointer[1]) {
>
> Use ./scripts/checkpatch.pl , it should indicate line too long here.
I did that, but got no warning. For text, the script limit is 75 and code limit is 100.
Should I break the line?
>
> There seem to be other places which likely also need similar fix:
>
> $ git grep CONFIG_ARMV8_PSCI.*rom_pointer arch/arm/mach-imx/
> arch/arm/mach-imx/imx8m/soc.c: if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
> rom_pointer[1])
> arch/arm/mach-imx/imx8m/soc.c: if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
> rom_pointer[1]) {
> arch/arm/mach-imx/imx8m/soc.c: if
> (!IS_ENABLED(CONFIG_ARMV8_PSCI) && rom_pointer[1]) {
> arch/arm/mach-imx/imx8m/soc.c: if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
> rom_pointer[0] &&
The other 2 places with similar conditions are in get_effective_memsize() and dram_init().
At least for i.MX8M and i.MX93 are not called from within SPL. Should we add nonetheless
the extra condition in those places as well? (for some of the other architectures,
dram_init() is called in SPL).
>
> > phys_addr_t optee_start = (phys_addr_t)rom_pointer[0];
> > phys_size_t optee_size = (size_t)rom_pointer[1];
> >
> > diff --git a/arch/arm/mach-imx/imx9/soc.c
> > b/arch/arm/mach-imx/imx9/soc.c index f43b73a6c2..ca91f20fe7 100644
> > --- a/arch/arm/mach-imx/imx9/soc.c
> > +++ b/arch/arm/mach-imx/imx9/soc.c
> > @@ -390,7 +390,7 @@ int dram_init_banksize(void)
> > }
> >
> > gd->bd->bi_dram[bank].start = PHYS_SDRAM;
> > - if (rom_pointer[1]) {
> > + if (!IS_ENABLED(CONFIG_SPL_BUILD) && rom_pointer[1]) {
> > phys_addr_t optee_start = (phys_addr_t)rom_pointer[0];
> > phys_size_t optee_size = (size_t)rom_pointer[1];
>
> [...]
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [EXT] Re: [PATCH] imx8m,imx9: Fix DRAM size calculation in SPL/dram_init_banksize()
2023-08-07 16:43 ` [EXT] " Elena Popa
@ 2023-08-07 21:44 ` Marek Vasut
0 siblings, 0 replies; 4+ messages in thread
From: Marek Vasut @ 2023-08-07 21:44 UTC (permalink / raw)
To: Elena Popa, u-boot; +Cc: Stefano Babic, Fabio Estevam, Peng Fan, dl-uboot-imx
On 8/7/23 18:43, Elena Popa wrote:
>
>> Subject tags should be (per git log arch/arm/mach-imx/imx8m/soc.c):
>>
>> arm: imx: imx8m: imx9: Fix DRAM .....
>
> Will do!
>
>>> If dram_init_banksize() is called from SPL, the rom_pointer, at that
>>> point, is not correctly initialized. This causes wrong calculation of
>>> DRAM start and size in dram_init_banksize(). The issue became apparent
>>> only in Falcon Mode. Added an extra condition to prevent using
>>> rom_pointer in SPL.
>>
>> +CC Stefano, Fabio, Peng.
>>
>>> Signed-off-by: Elena Popa <elena.popa@nxp.com>
>>> ---
>>> arch/arm/mach-imx/imx8m/soc.c | 2 +-
>>> arch/arm/mach-imx/imx9/soc.c | 2 +-
>>> 2 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/arch/arm/mach-imx/imx8m/soc.c
>>> b/arch/arm/mach-imx/imx8m/soc.c index d5254886be..dbb2154b08
>> 100644
>>> --- a/arch/arm/mach-imx/imx8m/soc.c
>>> +++ b/arch/arm/mach-imx/imx8m/soc.c
>>> @@ -273,7 +273,7 @@ int dram_init_banksize(void)
>>> }
>>>
>>> gd->bd->bi_dram[bank].start = PHYS_SDRAM;
>>> - if (!IS_ENABLED(CONFIG_ARMV8_PSCI) && rom_pointer[1]) {
>>> + if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
>>> + !IS_ENABLED(CONFIG_SPL_BUILD) && rom_pointer[1]) {
>>
>> Use ./scripts/checkpatch.pl , it should indicate line too long here.
>
> I did that, but got no warning. For text, the script limit is 75 and code limit is 100.
I wasn't aware of that, thanks for pointing it out.
> Should I break the line?
No
>> There seem to be other places which likely also need similar fix:
>>
>> $ git grep CONFIG_ARMV8_PSCI.*rom_pointer arch/arm/mach-imx/
>> arch/arm/mach-imx/imx8m/soc.c: if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
>> rom_pointer[1])
>> arch/arm/mach-imx/imx8m/soc.c: if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
>> rom_pointer[1]) {
>> arch/arm/mach-imx/imx8m/soc.c: if
>> (!IS_ENABLED(CONFIG_ARMV8_PSCI) && rom_pointer[1]) {
>> arch/arm/mach-imx/imx8m/soc.c: if (!IS_ENABLED(CONFIG_ARMV8_PSCI) &&
>> rom_pointer[0] &&
>
> The other 2 places with similar conditions are in get_effective_memsize() and dram_init().
> At least for i.MX8M and i.MX93 are not called from within SPL. Should we add nonetheless
> the extra condition in those places as well? (for some of the other architectures,
> dram_init() is called in SPL).
Better be consistent and fix them all, yes.
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2023-08-07 21:45 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-08-07 10:56 [PATCH] imx8m, imx9: Fix DRAM size calculation in SPL/dram_init_banksize() Elena Popa
2023-08-07 15:58 ` [PATCH] imx8m,imx9: " Marek Vasut
2023-08-07 16:43 ` [EXT] " Elena Popa
2023-08-07 21:44 ` Marek Vasut
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.