On 08/19/2015 09:29 AM, Itaru Kitayama wrote:
Hi Al,
Below is a dmesg excerpt corresponding to the acpi_boot_table_init().
[ 0.000000] ACPI: RSDP 0x00000047FA397014 000024 (v02 APM ) [ 0.000000] ACPI: XSDT 0x00000047FA3960E8 000064 (v01 APM XGENE 00000003 01000013) [ 0.000000] ACPI: FACP 0x00000047FA38E000 00010C (v05 APM XGENE 00000003 INTL 20141107) [ 0.000000] ACPI: DSDT 0x00000047FA38F000 003D82 (v05 APM APM88xxx 00000001 INTL 20141107) [ 0.000000] ACPI: DBG2 0x00000047FA393000 0000AA (v00 APMC0D XGENEDBG 00000000 INTL 20141107) [ 0.000000] ACPI: GTDT 0x00000047FA38C000 0000E0 (v02 APM XGENE 00000001 INTL 20141107) [ 0.000000] ACPI: MCFG 0x00000047FA38B000 00003C (v01 APM XGENE 00000002 INTL 20141107) [ 0.000000] ACPI: SPCR 0x00000047FA38A000 000050 (v02 APMC0D XGENESPC 00000000 INTL 20141107) [ 0.000000] ACPI: SSDT 0x00000047FA389000 00002D (v02 APM XGENE 00000001 INTL 20141107) [ 0.000000] ACPI: APIC 0x00000047FA388000 0002A4 (v03 APM XGENE 00000003 01000013) [ 0.000000] ACPI: SSDT 0x00000047FA387000 000063 (v02 REDHAT MACADDRS 00000001 01000013)
I've been testing the Mustang topic branch on the platform with the Rev 1 CPUs with 32 GB memory.
Okay. That's interesting output, but it unfortunately tells me nothing. Those messages only indicate that the kernel has found ACPI tables with those names, at those addresses, and that they have valid signatures in them. The content of the tables could be completely bogus, unless the kernel command acpi_force_table_verification was used, which will at least verify the checksums in the tables are correct, even if the content is not examined. I can deduce that acpi=force has been used, but that's about it.
What would be more useful is the remainder of the dmesg/console output so we can tell how far the kernel got in using the ACPI tables that it found, all the way up to the very last message out.
On 8/18/15 8:02 AM, Itaru Kitayama wrote:
Hi Al,
Even with your .config the HEAD of acpi-topic-mustang does not boot with acpi=force. Here are the messages appear on the console:
EFI stub: Booting Linux Kernel... EFI stub: Using DTB from configuration table EFI stub: Exiting boot services and installing virtual address map... XhcClearBiosOwnership: called to clear BIOS ownership XhcClearBiosOwnership: called to clear BIOS ownership L3c Cache: 8MB
On 8/18/15 6:58 AM, Itaru Kitayama wrote:
Hi Al,
Where does it "stall"? What's the last message you see? And if you're not using earlycon, please do so; that will help determine how far the boot is getting.
If I boot it with acpi=force, no output. Does my usage below look correct?
GRUB_CMDLINE_LINUX="earlycon=uart8250,mmio32,0x1c021000 acpi=force console=ttyS0,115200 loglevel=7"
How were those checked?
In setup_arch() I was calling disable_acpi() right after the>>> Itaru
Linaro-acpi mailing list Linaro-acpi@lists.linaro.org https://lists.linaro.org/mailman/listinfo/linaro-acpi
acpi_boot_table_init() call as that makes the system get to the prompt.
I am about to build the kernel with your .config and test it on one of Mustangs in the Linaro Austin lab. The Boot firmware version is 1.1.0-rh-0.15.
I may have mentioned this before (or not), but I think we need to get this Mustang updated to newer firmware; while that version is okay, the preferred version is the latest 1.15.x from APM (x = 18 or greater, I believe). We may have to figure out how to get the Austin lab updated somehow.