[U-Boot] [PATCH 0/3] i.MX refactoring patchset
This patchset include minor changes for i.MX; mainly:
1/3: improves the rare case of unkown booting machine 2/3: generalize the code so it works for i.MX23 and i.MX28
The patch 3/3 is what we've been using in Yocto and has been change to affect only mx28evk.
Otavio Salvador (3): imx: Use a clear identification of an unidentified CPU type mxs: generalize code for print_cpuinfo() mx28evk: extend default environment
arch/arm/cpu/arm926ejs/mx28/mx28.c | 29 +++++++++++- arch/arm/cpu/armv7/imx-common/cpu.c | 4 +- include/configs/mx28evk.h | 83 +++++++++++++++++++++++++++++++---- 3 files changed, 104 insertions(+), 12 deletions(-)
In case an unidentified CPU type is detected it now returns i.MX<unidentified>, in a const char.
Signed-off-by: Otavio Salvador otavio@ossystems.com.br Cc: Marek Vasut marex@denx.de Cc: Stefano Babic sbabic@denx.de Cc: Fabio Estevam fabio.estevam@freescale.com --- arch/arm/cpu/armv7/imx-common/cpu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/armv7/imx-common/cpu.c b/arch/arm/cpu/armv7/imx-common/cpu.c index b3195dd..77aac7d 100644 --- a/arch/arm/cpu/armv7/imx-common/cpu.c +++ b/arch/arm/cpu/armv7/imx-common/cpu.c @@ -66,7 +66,7 @@ char *get_reset_cause(void)
#if defined(CONFIG_DISPLAY_CPUINFO)
-static char *get_imx_type(u32 imxtype) +static const char *get_imx_type(u32 imxtype) { switch (imxtype) { case 0x63: @@ -80,7 +80,7 @@ static char *get_imx_type(u32 imxtype) case 0x53: return "53"; default: - return "unknown"; + return "<unidentified>"; } }
On Sun, Jun 17, 2012 at 9:58 AM, Otavio Salvador otavio@ossystems.com.br wrote:
-static char *get_imx_type(u32 imxtype) +static const char *get_imx_type(u32 imxtype)
This is OK, but ...
{ switch (imxtype) { case 0x63: @@ -80,7 +80,7 @@ static char *get_imx_type(u32 imxtype) case 0x53: return "53"; default:
- return "unknown";
- return "<unidentified>";
I really don't see any improvement here.
On Sun, Jun 17, 2012 at 10:05 AM, Fabio Estevam festevam@gmail.com wrote:
default:
- return "unknown";
- return "<unidentified>";
I really don't see any improvement here.
It is just to avoid a i.MXunkown return; besides this is how I am doing in the mx23 code too.
On Sun, Jun 17, 2012 at 10:09 AM, Otavio Salvador otavio@ossystems.com.br wrote:
On Sun, Jun 17, 2012 at 10:05 AM, Fabio Estevam festevam@gmail.com wrote:
default:
- return "unknown";
- return "<unidentified>";
I really don't see any improvement here.
It is just to avoid a i.MXunkown return; besides this is how I am doing in the mx23 code too.
This is OK. If you read "i.MXunkown" things are too broken already because the chip version could not be read.
Reading "i.MX<unidentified>" will make things no better.
On Sun, Jun 17, 2012 at 10:13 AM, Fabio Estevam festevam@gmail.com wrote:
Reading "i.MX<unidentified>" will make things no better.
OK; I'll drop it from the patch queue then.
Dear Fabio Estevam,
On Sun, Jun 17, 2012 at 10:09 AM, Otavio Salvador
otavio@ossystems.com.br wrote:
On Sun, Jun 17, 2012 at 10:05 AM, Fabio Estevam festevam@gmail.com wrote:
default:
return "unknown";
return "<unidentified>";I really don't see any improvement here.
It is just to avoid a i.MXunkown return; besides this is how I am doing in the mx23 code too.
This is OK. If you read "i.MXunkown" things are too broken already because the chip version could not be read.
Reading "i.MX<unidentified>" will make things no better.
This was from my head actually ... and I think this is better. The text is at least not mangled so badly. So let's keep it.
Best regards, Marek Vasut
On 17/06/2012 14:58, Otavio Salvador wrote:
In case an unidentified CPU type is detected it now returns i.MX<unidentified>, in a const char.
Signed-off-by: Otavio Salvador otavio@ossystems.com.br Cc: Marek Vasut marex@denx.de Cc: Stefano Babic sbabic@denx.de Cc: Fabio Estevam fabio.estevam@freescale.com
Hi Otavio,
arch/arm/cpu/armv7/imx-common/cpu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/armv7/imx-common/cpu.c b/arch/arm/cpu/armv7/imx-common/cpu.c index b3195dd..77aac7d 100644 --- a/arch/arm/cpu/armv7/imx-common/cpu.c +++ b/arch/arm/cpu/armv7/imx-common/cpu.c @@ -66,7 +66,7 @@ char *get_reset_cause(void)
#if defined(CONFIG_DISPLAY_CPUINFO)
-static char *get_imx_type(u32 imxtype) +static const char *get_imx_type(u32 imxtype) { switch (imxtype) { case 0x63: @@ -80,7 +80,7 @@ static char *get_imx_type(u32 imxtype) case 0x53: return "53"; default:
return "unknown";
return "<unidentified>";
Taste ist taste... but why <unidentified> is better that unknown ?
Stefano
On Sun, Jun 17, 2012 at 12:43 PM, Stefano Babic sbabic@denx.de wrote:
- return "unknown";
- return "<unidentified>";
Taste ist taste... but why <unidentified> is better that unknown ?
For me it gives a visual indication that we didn't meant to support a i.MXunknown processor ;-) i.MX<unidentified> make it easy to spot this.
But really, I don't mind it being included or not. I did it because Marek has comment about it when looking at my initial mx23 patches.
Dear Otavio Salvador,
On Sun, Jun 17, 2012 at 12:43 PM, Stefano Babic sbabic@denx.de wrote:
return "unknown";
return "<unidentified>";Taste ist taste... but why <unidentified> is better that unknown ?
For me it gives a visual indication that we didn't meant to support a i.MXunknown processor ;-) i.MX<unidentified> make it easy to spot this.
But really, I don't mind it being included or not. I did it because Marek has comment about it when looking at my initial mx23 patches.
And I like it more, it looks less like a typo.
Best regards, Marek Vasut
On Sun, Jun 17, 2012 at 1:03 PM, Marek Vasut marex@denx.de wrote:
But really, I don't mind it being included or not. I did it because Marek has comment about it when looking at my initial mx23 patches.
And I like it more, it looks less like a typo.
Sorry; I didn't get what you meant ... should it go or not?
On 17/06/2012 19:50, Otavio Salvador wrote:
On Sun, Jun 17, 2012 at 1:03 PM, Marek Vasut marex@denx.de wrote:
But really, I don't mind it being included or not. I did it because Marek has comment about it when looking at my initial mx23 patches.
And I like it more, it looks less like a typo.
Sorry; I didn't get what you meant ... should it go or not?
Hi Otavio,
I make also my point clearer for Marek: my point is that the behavior for all i.MX must be consistent. If you change here, please change also for i.MX31 (that returns "unknown in print_cpuinfo) and for other SOCs that returns "unknown" in print_cpuinfo() - or let the code as now.
Best regards, Stefano
Dear Stefano Babic,
On 17/06/2012 19:50, Otavio Salvador wrote:
On Sun, Jun 17, 2012 at 1:03 PM, Marek Vasut marex@denx.de wrote:
But really, I don't mind it being included or not. I did it because Marek has comment about it when looking at my initial mx23 patches.
And I like it more, it looks less like a typo.
Sorry; I didn't get what you meant ... should it go or not?
Hi Otavio,
I make also my point clearer for Marek: my point is that the behavior for all i.MX must be consistent. If you change here, please change also for i.MX31 (that returns "unknown in print_cpuinfo) and for other SOCs that returns "unknown" in print_cpuinfo() - or let the code as now.
Ah, I thought it was about the revision. Ok, whatever way is fine by me.
Best regards, Stefano
Best regards, Marek Vasut
Dear Otavio Salvador,
On Sun, Jun 17, 2012 at 1:03 PM, Marek Vasut marex@denx.de wrote:
But really, I don't mind it being included or not. I did it because Marek has comment about it when looking at my initial mx23 patches.
And I like it more, it looks less like a typo.
Sorry; I didn't get what you meant ... should it go or not?
It doesn't look like a typing mistake (i.MX<unknown> vs. i.MXunknown) => I like it more with "<" ">".
Best regards, Marek Vasut
The information now is gathered from HW_DIGCTL_CHIPID register and includes the revision of the chip on the output.
Signed-off-by: Otavio Salvador otavio@ossystems.com.br Cc: Marek Vasut marex@denx.de Cc: Stefano Babic sbabic@denx.de Cc: Fabio Estevam fabio.estevam@freescale.com --- arch/arm/cpu/arm926ejs/mx28/mx28.c | 29 +++++++++++++++++++++++++++-- 1 file changed, 27 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/arm926ejs/mx28/mx28.c b/arch/arm/cpu/arm926ejs/mx28/mx28.c index a82ff25..ac2f2e0 100644 --- a/arch/arm/cpu/arm926ejs/mx28/mx28.c +++ b/arch/arm/cpu/arm926ejs/mx28/mx28.c @@ -190,13 +190,38 @@ int arch_cpu_init(void) #endif
#if defined(CONFIG_DISPLAY_CPUINFO) +static const char *get_cpu_type(void) +{ + struct mx28_digctl_regs *digctl_regs = + (struct mx28_digctl_regs *)MXS_DIGCTL_BASE; + + switch (readl(&digctl_regs->hw_digctl_chipid) >> 16) { + case 0x2800: + return "28"; + case 0x3728: + return "23"; + default: + return "<unidentified>"; + } +} + +static u8 get_cpu_rev(void) +{ + struct mx28_digctl_regs *digctl_regs = + (struct mx28_digctl_regs *)MXS_DIGCTL_BASE; + + return readl(&digctl_regs->hw_digctl_chipid) & 0x0000F; +} + int print_cpuinfo(void) { struct mx28_spl_data *data = (struct mx28_spl_data *) ((CONFIG_SYS_TEXT_BASE - sizeof(struct mx28_spl_data)) & ~0xf);
- printf("Freescale i.MX28 family at %d MHz\n", - mxc_get_clock(MXC_ARM_CLK) / 1000000); + printf("CPU: Freescale i.MX%s rev%d at %d MHz\n", + get_cpu_type(), + get_cpu_rev(), + mxc_get_clock(MXC_ARM_CLK) / 1000000); printf("BOOT: %s\n", mx28_boot_modes[data->boot_mode_idx].mode); return 0; }
On Sun, Jun 17, 2012 at 9:58 AM, Otavio Salvador otavio@ossystems.com.br wrote:
The information now is gathered from HW_DIGCTL_CHIPID register and includes the revision of the chip on the output.
Signed-off-by: Otavio Salvador otavio@ossystems.com.br Cc: Marek Vasut marex@denx.de Cc: Stefano Babic sbabic@denx.de Cc: Fabio Estevam fabio.estevam@freescale.com
arch/arm/cpu/arm926ejs/mx28/mx28.c | 29 +++++++++++++++++++++++++++-- 1 file changed, 27 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/arm926ejs/mx28/mx28.c b/arch/arm/cpu/arm926ejs/mx28/mx28.c index a82ff25..ac2f2e0 100644 --- a/arch/arm/cpu/arm926ejs/mx28/mx28.c +++ b/arch/arm/cpu/arm926ejs/mx28/mx28.c @@ -190,13 +190,38 @@ int arch_cpu_init(void) #endif
#if defined(CONFIG_DISPLAY_CPUINFO) +static const char *get_cpu_type(void) +{
- struct mx28_digctl_regs *digctl_regs =
- (struct mx28_digctl_regs *)MXS_DIGCTL_BASE;
- switch (readl(&digctl_regs->hw_digctl_chipid) >> 16) {
- case 0x2800:
- return "28";
- case 0x3728:
- return "23";
- default:
- return "<unidentified>";
- }
+}
+static u8 get_cpu_rev(void) +{
- struct mx28_digctl_regs *digctl_regs =
- (struct mx28_digctl_regs *)MXS_DIGCTL_BASE;
- return readl(&digctl_regs->hw_digctl_chipid) & 0x0000F;
+}
int print_cpuinfo(void) { struct mx28_spl_data *data = (struct mx28_spl_data *) ((CONFIG_SYS_TEXT_BASE - sizeof(struct mx28_spl_data)) & ~0xf);
- printf("Freescale i.MX28 family at %d MHz\n",
- mxc_get_clock(MXC_ARM_CLK) / 1000000);
- printf("CPU: Freescale i.MX%s rev%d at %d MHz\n",
In all other i.MX SoCs we display the revision as rev%d.%d, for example:
rev1.0 , rev1.1 , etc.
I think we should print it in the same format here (x.y)
On Sun, Jun 17, 2012 at 9:58 AM, Otavio Salvador otavio@ossystems.com.br wrote:
#if defined(CONFIG_DISPLAY_CPUINFO) +static const char *get_cpu_type(void) +{
- struct mx28_digctl_regs *digctl_regs =
- (struct mx28_digctl_regs *)MXS_DIGCTL_BASE;
- switch (readl(&digctl_regs->hw_digctl_chipid) >> 16) {
- case 0x2800:
- return "28";
- case 0x3728:
Reading the Reference Manual I read 0x3780 instead.
Dear Fabio Estevam,
On Sun, Jun 17, 2012 at 9:58 AM, Otavio Salvador
otavio@ossystems.com.br wrote:
#if defined(CONFIG_DISPLAY_CPUINFO) +static const char *get_cpu_type(void) +{
struct mx28_digctl_regs *digctl_regs =(struct mx28_digctl_regs *)MXS_DIGCTL_BASE;switch (readl(&digctl_regs->hw_digctl_chipid) >> 16) {case 0x2800:return "28";
case 0x3728:Reading the Reference Manual I read 0x3780 instead.
Good catch ;-)
What's the code for mx6q? They dropped this register from it it seems ;-)
Best regards, Marek Vasut
On Sun, Jun 17, 2012 at 10:35 AM, Marek Vasut marex@denx.de wrote:
What's the code for mx6q? They dropped this register from it it seems ;-)
mx6 code does read the SoC type and revision correctly.
On Sun, Jun 17, 2012 at 10:25 AM, Fabio Estevam festevam@gmail.com wrote:
Reading the Reference Manual I read 0x3780 instead.
Indeed; I must have typoed it when writting it back.
I fixed it locally; Fabio will check the masking for mx28 revision and when I get it I update the patch according.
Dear Otavio Salvador,
On Sun, Jun 17, 2012 at 10:25 AM, Fabio Estevam festevam@gmail.com wrote:
Reading the Reference Manual I read 0x3780 instead.
Indeed; I must have typoed it when writting it back.
I fixed it locally; Fabio will check the masking for mx28 revision and when I get it I update the patch according.
It's 0x2800 ... roll out a new version ;-)
Best regards, Marek Vasut
On Sun, Jun 17, 2012 at 12:10 PM, Marek Vasut marex@denx.de wrote:
I fixed it locally; Fabio will check the masking for mx28 revision and when I get it I update the patch according.
It's 0x2800 ... roll out a new version ;-)
We are in doubt about the revision, not product code ;-)
Dear Otavio Salvador,
The information now is gathered from HW_DIGCTL_CHIPID register and includes the revision of the chip on the output.
Signed-off-by: Otavio Salvador otavio@ossystems.com.br Cc: Marek Vasut marex@denx.de Cc: Stefano Babic sbabic@denx.de Cc: Fabio Estevam fabio.estevam@freescale.com
arch/arm/cpu/arm926ejs/mx28/mx28.c | 29 +++++++++++++++++++++++++++-- 1 file changed, 27 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/arm926ejs/mx28/mx28.c b/arch/arm/cpu/arm926ejs/mx28/mx28.c index a82ff25..ac2f2e0 100644 --- a/arch/arm/cpu/arm926ejs/mx28/mx28.c +++ b/arch/arm/cpu/arm926ejs/mx28/mx28.c @@ -190,13 +190,38 @@ int arch_cpu_init(void) #endif
#if defined(CONFIG_DISPLAY_CPUINFO) +static const char *get_cpu_type(void) +{
- struct mx28_digctl_regs *digctl_regs =
(struct mx28_digctl_regs *)MXS_DIGCTL_BASE;- switch (readl(&digctl_regs->hw_digctl_chipid) >> 16) {
#define that >> 16 ... maybe offset? I'd define a mask though, see below.
#define PRODUCT_CODE_MASK (0xffff << 16)
- case 0x2800:
#define this as something ... #define PRODUCT_CODE_MX28 (0x2800 << 16)
return "28";- case 0x3728:
return "23";- default:
return "<unidentified>";
So X-Files-ish ;-)
- }
+}
+static u8 get_cpu_rev(void)
uint8_t ... maybe? it's used throughout the code, so let's keep consistent.
+{
- struct mx28_digctl_regs *digctl_regs =
(struct mx28_digctl_regs *)MXS_DIGCTL_BASE;- return readl(&digctl_regs->hw_digctl_chipid) & 0x0000F;
+}
int print_cpuinfo(void) { struct mx28_spl_data *data = (struct mx28_spl_data *) ((CONFIG_SYS_TEXT_BASE - sizeof(struct mx28_spl_data)) & ~0xf);
- printf("Freescale i.MX28 family at %d MHz\n",
mxc_get_clock(MXC_ARM_CLK) / 1000000);
- printf("CPU: Freescale i.MX%s rev%d at %d MHz\n",
get_cpu_type(),get_cpu_rev(),mxc_get_clock(MXC_ARM_CLK) / 1000000);}
Best regards, Marek Vasut
On 17/06/2012 14:58, Otavio Salvador wrote:
The information now is gathered from HW_DIGCTL_CHIPID register and includes the revision of the chip on the output.
Signed-off-by: Otavio Salvador otavio@ossystems.com.br Cc: Marek Vasut marex@denx.de Cc: Stefano Babic sbabic@denx.de Cc: Fabio Estevam fabio.estevam@freescale.com
Hi Otavio,
arch/arm/cpu/arm926ejs/mx28/mx28.c | 29 +++++++++++++++++++++++++++-- 1 file changed, 27 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/arm926ejs/mx28/mx28.c b/arch/arm/cpu/arm926ejs/mx28/mx28.c index a82ff25..ac2f2e0 100644 --- a/arch/arm/cpu/arm926ejs/mx28/mx28.c +++ b/arch/arm/cpu/arm926ejs/mx28/mx28.c @@ -190,13 +190,38 @@ int arch_cpu_init(void) #endif
#if defined(CONFIG_DISPLAY_CPUINFO) +static const char *get_cpu_type(void) +{
- struct mx28_digctl_regs *digctl_regs =
(struct mx28_digctl_regs *)MXS_DIGCTL_BASE;- switch (readl(&digctl_regs->hw_digctl_chipid) >> 16) {
- case 0x2800:
return "28";- case 0x3728:
return "23";- default:
return "<unidentified>";- }
+}
I like that there will be support for i.MX23, too. But I dislike that everything takes the name "MX28". As you suggest in your subject, maybe it is time to rename directories, and use "mxs" (as in kernel) instead of mx28.
+static u8 get_cpu_rev(void) +{
- struct mx28_digctl_regs *digctl_regs =
(struct mx28_digctl_regs *)MXS_DIGCTL_BASE;- return readl(&digctl_regs->hw_digctl_chipid) & 0x0000F;
+}
Everywhere (i.MX, omap, ...) get_cpu_rev returns u32. The function is currently exported, too.
Best regards, Stefano Babic
Dear Stefano Babic,
On 17/06/2012 14:58, Otavio Salvador wrote:
The information now is gathered from HW_DIGCTL_CHIPID register and includes the revision of the chip on the output.
Signed-off-by: Otavio Salvador otavio@ossystems.com.br Cc: Marek Vasut marex@denx.de Cc: Stefano Babic sbabic@denx.de Cc: Fabio Estevam fabio.estevam@freescale.com
Hi Otavio,
arch/arm/cpu/arm926ejs/mx28/mx28.c | 29 +++++++++++++++++++++++++++-- 1 file changed, 27 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/arm926ejs/mx28/mx28.c b/arch/arm/cpu/arm926ejs/mx28/mx28.c index a82ff25..ac2f2e0 100644 --- a/arch/arm/cpu/arm926ejs/mx28/mx28.c +++ b/arch/arm/cpu/arm926ejs/mx28/mx28.c @@ -190,13 +190,38 @@ int arch_cpu_init(void)
#endif
#if defined(CONFIG_DISPLAY_CPUINFO)
+static const char *get_cpu_type(void) +{
- struct mx28_digctl_regs *digctl_regs =
(struct mx28_digctl_regs *)MXS_DIGCTL_BASE;- switch (readl(&digctl_regs->hw_digctl_chipid) >> 16) {
- case 0x2800:
return "28";- case 0x3728:
return "23";- default:
return "<unidentified>";- }
+}
I like that there will be support for i.MX23, too. But I dislike that everything takes the name "MX28". As you suggest in your subject, maybe it is time to rename directories, and use "mxs" (as in kernel) instead of mx28.
We can do that eventually, later ... it depends on the ordering of Otavio's patches, I'm fine either way.
+static u8 get_cpu_rev(void) +{
- struct mx28_digctl_regs *digctl_regs =
(struct mx28_digctl_regs *)MXS_DIGCTL_BASE;- return readl(&digctl_regs->hw_digctl_chipid) & 0x0000F;
+}
Everywhere (i.MX, omap, ...) get_cpu_rev returns u32. The function is currently exported, too.
Correct, but isn't the return value mangled somehow (like having major rev. << 16 and minor rev. << 0 )? Or that's only IMX?
Best regards, Stefano Babic
Best regards, Marek Vasut
On 17/06/2012 18:02, Marek Vasut wrote:
I like that there will be support for i.MX23, too. But I dislike that everything takes the name "MX28". As you suggest in your subject, maybe it is time to rename directories, and use "mxs" (as in kernel) instead of mx28.
We can do that eventually, later ... it depends on the ordering of Otavio's patches, I'm fine either way.
Fine with me - anyway, it is something must be done.
+static u8 get_cpu_rev(void) +{
- struct mx28_digctl_regs *digctl_regs =
(struct mx28_digctl_regs *)MXS_DIGCTL_BASE;- return readl(&digctl_regs->hw_digctl_chipid) & 0x0000F;
+}
Everywhere (i.MX, omap, ...) get_cpu_rev returns u32. The function is currently exported, too.
Correct, but isn't the return value mangled somehow (like having major rev. << 16 and minor rev. << 0 )? Or that's only IMX?
Checking in current implementations for get_cpu_rev() (i.MX / OMAP24 / OMAP3 / AM33, and even board/apollon/sys_info.c: I do not know why the cpu detection is inseide this file), none of them requires strictly 32 bit. However, we should be coherent in all code - for this reason I dislike if one of the i.MX is implemented differently as for other SOCs.
Best regards, Stefano Babic
On Sun, Jun 17, 2012 at 2:56 PM, Stefano Babic sbabic@denx.de wrote:
Correct, but isn't the return value mangled somehow (like having major rev. << 16 and minor rev. << 0 )? Or that's only IMX?
Checking in current implementations for get_cpu_rev() (i.MX / OMAP24 / OMAP3 / AM33, and even board/apollon/sys_info.c: I do not know why the cpu detection is inseide this file), none of them requires strictly 32 bit. However, we should be coherent in all code - for this reason I dislike if one of the i.MX is implemented differently as for other SOCs.
So, we change this one to follow the others or fix the others?
Dear Otavio Salvador,
On Sun, Jun 17, 2012 at 2:56 PM, Stefano Babic sbabic@denx.de wrote:
Correct, but isn't the return value mangled somehow (like having major rev. << 16 and minor rev. << 0 )? Or that's only IMX?
Checking in current implementations for get_cpu_rev() (i.MX / OMAP24 / OMAP3 / AM33, and even board/apollon/sys_info.c: I do not know why the cpu detection is inseide this file), none of them requires strictly 32 bit. However, we should be coherent in all code - for this reason I dislike if one of the i.MX is implemented differently as for other SOCs.
So, we change this one to follow the others or fix the others?
Fix the others ;-)
Best regards, Marek Vasut
Dear Stefano Babic,
On 17/06/2012 18:02, Marek Vasut wrote:
I like that there will be support for i.MX23, too. But I dislike that everything takes the name "MX28". As you suggest in your subject, maybe it is time to rename directories, and use "mxs" (as in kernel) instead of mx28.
We can do that eventually, later ... it depends on the ordering of Otavio's patches, I'm fine either way.
Fine with me - anyway, it is something must be done.
+static u8 get_cpu_rev(void) +{
- struct mx28_digctl_regs *digctl_regs =
(struct mx28_digctl_regs *)MXS_DIGCTL_BASE;- return readl(&digctl_regs->hw_digctl_chipid) & 0x0000F;
+}
Everywhere (i.MX, omap, ...) get_cpu_rev returns u32. The function is currently exported, too.
Correct, but isn't the return value mangled somehow (like having major rev. << 16 and minor rev. << 0 )? Or that's only IMX?
Checking in current implementations for get_cpu_rev() (i.MX / OMAP24 / OMAP3 / AM33, and even board/apollon/sys_info.c: I do not know why the cpu detection is inseide this file), none of them requires strictly 32 bit. However, we should be coherent in all code - for this reason I dislike if one of the i.MX is implemented differently as for other SOCs.
+1
Best regards, Stefano Babic
Best regards, Marek Vasut
On Sun, Jun 17, 2012 at 12:49 PM, Stefano Babic sbabic@denx.de wrote:
I like that there will be support for i.MX23, too. But I dislike that everything takes the name "MX28". As you suggest in your subject, maybe it is time to rename directories, and use "mxs" (as in kernel) instead of mx28.
I have it in my wip branch; I'd like to hold it until I have mx23 working so I have freedom to rework the patches.
I can split it on mx28 part and leave mx23 on my branch; this does looks nicer for me.
+static u8 get_cpu_rev(void) +{
- struct mx28_digctl_regs *digctl_regs =
- (struct mx28_digctl_regs *)MXS_DIGCTL_BASE;
- return readl(&digctl_regs->hw_digctl_chipid) & 0x0000F;
+}
Everywhere (i.MX, omap, ...) get_cpu_rev returns u32. The function is currently exported, too.
I change it then.
Will fix it.
The environment has been based on mx53loco and m28evk but keeping the possibility to easy change the default console device as Freescale and mainline kernels differ on the device name.
Signed-off-by: Otavio Salvador otavio@ossystems.com.br Cc: Marek Vasut marex@denx.de Cc: Stefano Babic sbabic@denx.de Cc: Wolfgang Denk wd@denx.de Cc: Fabio Estevam fabio.estevam@freescale.com --- include/configs/mx28evk.h | 83 ++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 75 insertions(+), 8 deletions(-)
diff --git a/include/configs/mx28evk.h b/include/configs/mx28evk.h index e98a746..889d58a 100644 --- a/include/configs/mx28evk.h +++ b/include/configs/mx28evk.h @@ -234,7 +234,6 @@ #define CONFIG_SETUP_MEMORY_TAGS #define CONFIG_BOOTDELAY 3 #define CONFIG_BOOTFILE "uImage" -#define CONFIG_BOOTCOMMAND "run bootcmd_net" #define CONFIG_LOADADDR 0x42000000 #define CONFIG_SYS_LOAD_ADDR CONFIG_LOADADDR #define CONFIG_OF_LIBFDT @@ -243,13 +242,81 @@ * Extra Environments */ #define CONFIG_EXTRA_ENV_SETTINGS \ - "console_fsl=console=ttyAM0" \ - "console_mainline=console=ttyAMA0" \ - "netargs=setenv bootargs console=${console_mainline}" \ + "update_nand_full_filename=u-boot.nand\0" \ + "update_nand_firmware_filename=u-boot.sb\0" \ + "update_sd_firmware_filename=u-boot.sd\0" \ + "update_nand_firmware_maxsz=0x100000\0" \ + "update_nand_stride=0x40\0" /* MX28 datasheet ch. 12.12 */ \ + "update_nand_count=0x4\0" /* MX28 datasheet ch. 12.12 */ \ + "update_nand_get_fcb_size=" /* Get size of FCB blocks */ \ + "nand device 0 ; " \ + "nand info ; " \ + "setexpr fcb_sz ${update_nand_stride} * ${update_nand_count};" \ + "setexpr update_nand_fcb ${fcb_sz} * ${nand_writesize}\0" \ + "update_nand_full=" /* Update FCB, DBBT and FW */ \ + "if tftp ${update_nand_full_filename} ; then " \ + "run update_nand_get_fcb_size ; " \ + "nand scrub -y 0x0 ${filesize} ; " \ + "nand write.raw ${loadaddr} 0x0 ${update_nand_fcb} ; " \ + "setexpr update_off ${loadaddr} + ${update_nand_fcb} ; " \ + "setexpr update_sz ${filesize} - ${update_nand_fcb} ; " \ + "nand write ${update_off} ${update_nand_fcb} ${update_sz} ; " \ + "fi\0" \ + "update_nand_firmware=" /* Update only firmware */ \ + "if tftp ${update_nand_firmware_filename} ; then " \ + "run update_nand_get_fcb_size ; " \ + "setexpr fcb_sz ${update_nand_fcb} * 2 ; " /* FCB + DBBT */ \ + "setexpr fw_sz ${update_nand_firmware_maxsz} * 2 ; " \ + "setexpr fw_off ${fcb_sz} + ${update_nand_firmware_maxsz};" \ + "nand erase ${fcb_sz} ${fw_sz} ; " \ + "nand write ${loadaddr} ${fcb_sz} ${filesize} ; " \ + "nand write ${loadaddr} ${fw_off} ${filesize} ; " \ + "fi\0" \ + "update_sd_firmware=" /* Update the SD firmware partition */ \ + "if mmc rescan ; then " \ + "if tftp ${update_sd_firmware_filename} ; then " \ + "setexpr fw_sz ${filesize} / 0x200 ; " /* SD block size */ \ + "setexpr fw_sz ${fw_sz} + 1 ; " \ + "mmc write ${loadaddr} 0x800 ${fw_sz} ; " \ + "fi ; " \ + "fi\0" \ + "script=boot.scr\0" \ + "uimage=uImage\0" \ + "console_fsl=ttyAM0\0" \ + "console_mainline=ttyAMA0\0" \ + "console=${console_mainline}\0" \ + "mmcdev=0\0" \ + "mmcpart=2\0" \ + "mmcroot=/dev/mmcblk0p3 rw\0" \ + "mmcrootfstype=ext3 rootwait\0" \ + "mmcargs=setenv bootargs console=${console},${baudrate} " \ + "root=${mmcroot} " \ + "rootfstype=${mmcrootfstype}\0" \ + "loadbootscript=" \ + "fatload mmc ${mmcdev}:${mmcpart} ${loadaddr} ${script};\0" \ + "bootscript=echo Running bootscript from mmc ...; " \ + "source\0" \ + "loaduimage=fatload mmc ${mmcdev}:${mmcpart} ${loadaddr} ${uimage}\0" \ + "mmcboot=echo Booting from mmc ...; " \ + "run mmcargs; " \ + "bootm\0" \ + "netargs=setenv bootargs console=${console},${baudrate} " \ "root=/dev/nfs " \ - "ip=dhcp nfsroot=${serverip}:${nfsroot}\0" \ - "bootcmd_net=echo Booting from net ...; " \ - "run netargs; " \ - "dhcp ${uimage}; bootm\0" \ + "ip=dhcp nfsroot=${serverip}:${nfsroot},v3,tcp\0" \ + "netboot=echo Booting from net ...; " \ + "run netargs; " \ + "dhcp ${uimage}; bootm\0" + +#define CONFIG_BOOTCOMMAND \ + "if mmc rescan ${mmcdev}; then " \ + "if run loadbootscript; then " \ + "run bootscript; " \ + "else " \ + "if run loaduimage; then " \ + "run mmcboot; " \ + "else run netboot; " \ + "fi; " \ + "fi; " \ + "else run netboot; fi"
#endif /* __MX28EVK_CONFIG_H__ */
Dear Otavio Salvador,
This patchset include minor changes for i.MX; mainly:
1/3: improves the rare case of unkown booting machine
<trolling> The word's actually "unknown" ;-D </trolling>
2/3: generalize the code so it works for i.MX23 and i.MX28
The patch 3/3 is what we've been using in Yocto and has been change to affect only mx28evk.
Otavio Salvador (3): imx: Use a clear identification of an unidentified CPU type mxs: generalize code for print_cpuinfo() mx28evk: extend default environment
It's good, really good :-)
arch/arm/cpu/arm926ejs/mx28/mx28.c | 29 +++++++++++- arch/arm/cpu/armv7/imx-common/cpu.c | 4 +- include/configs/mx28evk.h | 83 +++++++++++++++++++++++++++++++---- 3 files changed, 104 insertions(+), 12 deletions(-)
Best regards, Marek Vasut
participants (5)
-
Fabio Estevam -
Marek Vasut -
Marek Vasut -
Otavio Salvador -
Stefano Babic