Re: [tip:x86/apic] x86: read apic ID in the !acpi_lapic case

[Ingo Molnar <mingo@elte.hu>]

* Cyrill Gorcunov <gorcunov@gmail.com> wrote:

> [Ingo Molnar - Mon, May 11, 2009 at 01:02:34PM +0200]
> | 
> | * tip-bot for Yinghai Lu <yinghai@kernel.org> wrote:
> | 
> | > Commit-ID:  4401da6111ac58f94234417427d06a72c4048c74
> | > Gitweb:     http://git.kernel.org/tip/4401da6111ac58f94234417427d06a72c4048c74
> | > Author:     Yinghai Lu <yinghai@kernel.org>
> | > AuthorDate: Sat, 2 May 2009 10:40:57 -0700
> | > Committer:  Ingo Molnar <mingo@elte.hu>
> | > CommitDate: Mon, 11 May 2009 11:29:23 +0200
> | > 
> | > x86: read apic ID in the !acpi_lapic case
> | > 
> | > Ed found that on 32-bit, boot_cpu_physical_apicid is not read right,
> | > when the mptable is broken.
> | 
> | -tip testing found that this patch causes a kernel crash:
> | 
> | [    1.208000] initcall pci_iommu_init+0x0/0x11 returned 0 after 0 usecs
> | [    1.212000] calling  print_all_ICs+0x0/0x62a @ 1
> | [    1.216000] 
> | [    1.216000] printing PIC contents
> | [    1.220000] ... PIC  IMR: fffa
> | [    1.220000] ... PIC  IRR: 0000
> | [    1.224000] ... PIC  ISR: 0000
> | [    1.228000] ... PIC ELCR: 0828
> | [    1.232000] BUG: unable to handle kernel paging request at ffffb020
> | [    1.232000] IP: [<78417d5c>] native_apic_mem_read+0x15/0x19
> | [    1.232000] *pde = 010ef067 *pte = 00000000 
> | [    1.232000] Oops: 0000 [#1] PREEMPT 
> | [    1.232000] last sysfs file: 
> | [    1.232000] Modules linked in:
> | [    1.232000] 
> | [    1.232000] Pid: 1, comm: swapper Not tainted (2.6.30-rc5-tip #40847) System Product Name
> | [    1.232000] EIP: 0060:[<78417d5c>] EFLAGS: 00010046 CPU: 0
> | [    1.232000] EIP is at native_apic_mem_read+0x15/0x19
> | [    1.232000] EAX: ffffb020 EBX: 00000000 ECX: 00000000 EDX: fffff000
> | [    1.232000] ESI: 00000000 EDI: 7908c52e EBP: 9786fd38 ESP: 9786fd38
> | [    1.232000]  DS: 007b ES: 007b FS: 0000 GS: 00e0 SS: 0068
> | [    1.232000] Process swapper (pid: 1, ti=9786e000 task=97870000 task.ti=9786e000)
> | [    1.232000] Stack:
> | [    1.232000]  9786fd40 78416e0d 9786fd48 78416fad 9786fd64 7908c55f 00000001 00000000
> | [    1.232000]  00000000 00000000 7908c52e 9786fd78 784671bf 00000001 00000000 00000000
> | [    1.232000]  9786fdac 7908ca42 00000001 9786fda0 784551ad 00000001 00000000 487ab0fc
> | [    1.232000] Call Trace:
> | [    1.232000]  [<78416e0d>] ? read_apic_id+0x19/0x24
> | [    1.232000]  [<78416fad>] ? hard_smp_processor_id+0xd/0xf
> | [    1.232000]  [<7908c55f>] ? print_local_APIC+0x31/0x3c7
> | [    1.232000]  [<7908c52e>] ? print_local_APIC+0x0/0x3c7
> | [    1.232000]  [<784671bf>] ? smp_call_function_single+0x5b/0x64
> | [    1.232000]  [<7908ca42>] ? print_all_ICs+0x14d/0x62a
> | [    1.232000]  [<784551ad>] ? ktime_get+0x16/0x34
> | [    1.232000]  [<7908c8f5>] ? print_all_ICs+0x0/0x62a
> | [    1.232000]  [<7908c8f5>] ? print_all_ICs+0x0/0x62a
> | [    1.232000]  [<7840107f>] ? _stext+0x7f/0x1f2
> | [    1.232000]  [<786c2658>] ? put_dec+0x44/0x137
> | [    1.232000]  [<784afae4>] ? page_count+0x22/0x3f
> | [    1.232000]  [<784b1c0a>] ? get_page_from_freelist+0x51f/0x6b0
> | [    1.232000]  [<786c2658>] ? put_dec+0x44/0x137
> | [    1.232000]  [<784dcbee>] ? check_object+0x207/0x252
> | [    1.232000]  [<786bb54f>] ? ida_get_new_above+0x2b0/0x2ca
> | [    1.232000]  [<784dd5c5>] ? __slab_free+0x11a/0x4af
> | [    1.232000]  [<784df012>] ? kmem_cache_free+0xce/0x163
> | [    1.232000]  [<786bb54f>] ? ida_get_new_above+0x2b0/0x2ca
> | [    1.232000]  [<786bb54f>] ? ida_get_new_above+0x2b0/0x2ca
> | [    1.232000]  [<786bb75f>] ? ida_pre_get+0x13/0x18b
> | [    1.232000]  [<7853f384>] ? proc_register+0x32a/0x34e
> | [    1.232000]  [<7853f5bd>] ? create_proc_entry+0xf0/0x119
> | [    1.232000]  [<7847e0bc>] ? register_irq_proc+0x87/0xbb
> | [    1.232000]  [<7908251b>] ? kernel_init+0x8d/0x10a
> | [    1.232000]  [<7908248e>] ? kernel_init+0x0/0x10a
> | [    1.232000]  [<78403ee7>] ? kernel_thread_helper+0x7/0x10
> | [    1.232000] Code: 44 00 00 89 d1 8b 15 08 8b de 78 8d 84 10 00 c0 ff ff 89 08 5d c3 55 89 e5 0f 1f 44 00 00 8b 15 08 8b de 78 8d 84 10 00 c0 ff ff <8b> 00 5d c3 55 89 e5 0f 1f 44 00 00 89 c2 eb 02 f3 90 8b 02 85 
> | [    1.232000] EIP: [<78417d5c>] native_apic_mem_read+0x15/0x19 SS:ESP 0068:9786fd38
> | [    1.232000] CR2: 00000000ffffb020
> | [    1.232000] ---[ end trace a7919e7f17c0a725 ]---
> | [    1.232000] Kernel panic - not syncing: Fatal exception
> | 
> | 	Ingo
> 
> Ingo, could check if the following do the trick?
> (on top of 
> 
> 	commit f6271becbae641903a9f703b7b75cccb202df82a
> 	Merge: 486cb6c 0498625
> 	Author: Ingo Molnar <mingo@elte.hu>
> 	Date:   Mon May 11 12:26:00 2009 +0200
> 
> 	    Merge branch 'tracing/core'
> )
> 
> 
> 	-- Cyrill
> ---
> 
> x86: apic - fixmap apic address even if apic disabled
> 
> In case if apic were disabled by boot option
> we still need read_apic operation. So fixmap
> a fake apic area if needed.
> 
> Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
> ---
>  arch/x86/kernel/apic/apic.c |   17 +++++++++--------
>  1 file changed, 9 insertions(+), 8 deletions(-)
> 
> Index: linux-2.6.git/arch/x86/kernel/apic/apic.c
> =====================================================================
> --- linux-2.6.git.orig/arch/x86/kernel/apic/apic.c
> +++ linux-2.6.git/arch/x86/kernel/apic/apic.c
> @@ -1591,13 +1591,6 @@ void __init init_apic_mappings(void)
>  	} else
>  		apic_phys = mp_lapic_addr;
>  
> -	/* lets check if we may NOP'ify apic operations */
> -	if (!cpu_has_apic) {
> -		pr_info("APIC: disable apic facility\n");
> -		apic_disable();
> -		return;
> -	}
> -
>  	/*
>  	 * acpi lapic path already maps that address in
>  	 * acpi_register_lapic_address()
> @@ -1606,7 +1599,15 @@ void __init init_apic_mappings(void)
>  		set_fixmap_nocache(FIX_APIC_BASE, apic_phys);
>  
>  	apic_printk(APIC_VERBOSE, "mapped APIC to %08lx (%08lx)\n",
> -				APIC_BASE, apic_phys);
> +			APIC_BASE, apic_phys);
> +
> +	/* lets check if we may NOP'ify apic operations */
> +	if (!cpu_has_apic) {
> +		pr_info("APIC: disable apic facility\n");
> +		apic_disable();
> +		return;
> +	}
> +

Will check this - thanks.

	Ingo