On Thu, 16 Oct 2014 10:50:42 +0200 Hans de Goede hdegoede@redhat.com wrote:
Hi,
On 10/15/2014 12:48 PM, Siarhei Siamashka wrote:
On Mon, 13 Oct 2014 14:55:35 +0200 Hans de Goede hdegoede@redhat.com wrote:
Many people are still using old linux-sunxi-3.4 kernels on sunxi devices, adding the proper MACH_TYPE defines for this allows people to switch to upstream u-boot, so that we can stop maintaining the linux-sunxi u-boot fork.
These machine-ids are all properly registered at:
http://www.arm.linux.org.uk/developer/machines/
Signed-off-by: Hans de Goede hdegoede@redhat.com
include/configs/sun4i.h | 1 + include/configs/sun5i.h | 1 + include/configs/sun7i.h | 1 + 3 files changed, 3 insertions(+)
diff --git a/include/configs/sun4i.h b/include/configs/sun4i.h index 5611ecc..d0191a3 100644 --- a/include/configs/sun4i.h +++ b/include/configs/sun4i.h @@ -15,6 +15,7 @@ #define CONFIG_CLK_FULL_SPEED 1008000000
#define CONFIG_SYS_PROMPT "sun4i# " +#define CONFIG_MACH_TYPE 4104
#ifdef CONFIG_USB_EHCI #define CONFIG_USB_EHCI_SUNXI diff --git a/include/configs/sun5i.h b/include/configs/sun5i.h index 6066371..7b683e9 100644 --- a/include/configs/sun5i.h +++ b/include/configs/sun5i.h @@ -15,6 +15,7 @@ #define CONFIG_CLK_FULL_SPEED 1008000000
#define CONFIG_SYS_PROMPT "sun5i# " +#define CONFIG_MACH_TYPE 4138
#ifdef CONFIG_USB_EHCI #define CONFIG_USB_EHCI_SUNXI diff --git a/include/configs/sun7i.h b/include/configs/sun7i.h index a902b84..966cbd8 100644 --- a/include/configs/sun7i.h +++ b/include/configs/sun7i.h @@ -16,6 +16,7 @@ #define CONFIG_CLK_FULL_SPEED 912000000
#define CONFIG_SYS_PROMPT "sun7i# " +#define CONFIG_MACH_TYPE 4283
#ifdef CONFIG_USB_EHCI #define CONFIG_USB_EHCI_SUNXI
This patch is just trying to remove the safety guards (which exist there for a reason!) without addressing the real compatibility issues.
Erm, no the mach_type-s are not "safety-guards", they are absolutely necessary to get old non devicetree kernels to work at all.
If it looks like a safety-guard, swims like a safety-guard, and quacks like a safety-guard, then it probably is a safety-guard.
The users of the u-boot from the v2014.10 tag can't easily hurt themselves, because the old non-mainline 3.4 kernel will just refuse to boot.
Yes, the users can easily patch the sources to add the machine id or provide it via the environment variable. But it requires some action from the user. This is kind of like a "no trespassing" tape, which can be easily stepped over. But a certain level of protection is still there.
With just this single patch, sunxi-3.4 kernels with fixed PLL5 support, will happily boot on sun4i and sun5i.
And with the bootm_boot_mode patch + env setting sun7i will boot fine too.
Yes, it's a very simple problem. And multiple simple solutions are available. You just hastily picked one of them. And it has some unfortunate practical implications related to DRAM configuration opportunities, which I have already explained to you.
As a custodian, you take full responsibility for your actions. If you think that this was the best choice, then so be it.