On 07/25/2018 10:27 PM, Ran Wang wrote:
Hi York,
-----Original Message----- From: York Sun Sent: Tuesday, July 24, 2018 23:40 To: Ran Wang ran.wang_1@nxp.com; Albert Aribaud albert.u.boot@aribaud.net Cc: u-boot@lists.denx.de Subject: Re: [PATCH] armv8: layerscape: Enable EHCI access for LS1012A
On 07/10/2018 07:11 PM, Ran Wang wrote:
Hi York,
-----Original Message----- From: York Sun Sent: Wednesday, July 11, 2018 05:06 To: Ran Wang ran.wang_1@nxp.com; Albert Aribaud albert.u.boot@aribaud.net Cc: u-boot@lists.denx.de Subject: Re: [PATCH] armv8: layerscape: Enable EHCI access for LS1012A
On 07/02/2018 10:34 PM, Ran Wang wrote:
Program Central Security Unit (CSU) to grant access permission for USB 2.0 controller, otherwiase EHCI funciton will down.
Signed-off-by: Ran Wang ran.wang_1@nxp.com
arch/arm/cpu/armv8/fsl-layerscape/soc.c | 8 ++++++++ arch/arm/include/asm/arch-fsl-layerscape/ns_access.h | 1 + 2 files changed, 9 insertions(+)
diff --git a/arch/arm/cpu/armv8/fsl-layerscape/soc.c b/arch/arm/cpu/armv8/fsl-layerscape/soc.c index 6a56269..2c4cf7f 100644 --- a/arch/arm/cpu/armv8/fsl-layerscape/soc.c +++ b/arch/arm/cpu/armv8/fsl-layerscape/soc.c @@ -14,6 +14,7 @@ #include <asm/io.h> #include <asm/global_data.h> #include <asm/arch-fsl-layerscape/config.h> +#include <asm/arch-fsl-layerscape/ns_access.h> #ifdef CONFIG_LAYERSCAPE_NS_ACCESS #include <fsl_csu.h> #endif @@ -668,6 +669,13 @@ void fsl_lsch2_early_init_f(void) CCI400_DVM_MESSAGE_REQ_EN |
CCI400_SNOOP_REQ_EN);
}
- /*
* Program Central Security Unit (CSU) to grant access* permission for USB 2.0 controller*/+#if defined(CONFIG_ARCH_LS1012A) &&
defined(CONFIG_USB_EHCI_FSL)
- set_devices_ns_access(CSU_CSLX_USB_2, CSU_ALL_RW); #endif
Is this LS1012A specific?
For Layerscape platforms, only LS1012A and LS1021A have USB2.0(EHCI) controller, Others have USB3.0 controller only. For now I can only verify on LS1012A, so didn't cover LS1021A yet.
Ran,
I think calling function set_devices_ns_access() may have an issue. It is not EL2 safe, is it? Please check enable_layerscape_ns_access(). It detects exception level before accessing EL3-only registers.
Yes, I re-checked the code and the experiments prove your point, I will send v2 patch soon. Thanks for reminding.
BTW, I observed that in function board_late_init(), the return value of current_el() will be different: Case 1: Boot with PPA: return 2 ... Using SERDES1 Protocol: 13576 (0x3508) PPA Firmware: Version LSDK-1712-TC1-dirty SEC Firmware: 'loadables' present in config loadables: 'trustedOS@1' ... FPGA: v4 (LS1012QDS_2016_0405_1443), build 100 on Tue Apr 05 09:13:13 2016 wangran!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! current_el():2
Case 2: without PPA: return 3 ... Using SERDES1 Protocol: 13576 (0x3508) SEC Firmware: config-1: no such config SEC Firmware: error (-2) SEC Firmware: Failed to load image ... FPGA: v4 (LS1012QDS_2016_0405_1443), build 100 on Tue Apr 05 09:13:13 2016 wangran!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! current_el():3
Do you know the reason?
That's exactly what PPA does -- drop to EL2.
York