* Re: [PATCH v1] x86/apic: Fix kernel panic when "intremap=off" and "x2apic_phys" are set
2023-06-16 21:22 [PATCH v1] x86/apic: Fix kernel panic when "intremap=off" and "x2apic_phys" are set Dheeraj Kumar Srivastava
@ 2023-06-18 22:24 ` Cyrill Gorcunov
2023-06-19 15:25 ` Thomas Gleixner
2023-06-19 14:50 ` Borislav Petkov
` (2 subsequent siblings)
3 siblings, 1 reply; 6+ messages in thread
From: Cyrill Gorcunov @ 2023-06-18 22:24 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava
Cc: linux-kernel, tglx, mingo, bp, dave.hansen, x86, hpa,
suresh.b.siddha, Kishon Vijay Abraham I, Vasant Hegde
On Sat, Jun 17, 2023 at 02:52:36AM +0530, Dheeraj Kumar Srivastava wrote:
> x2APIC mode requires "Interrupt Remapping" to be enabled and the
> "physical x2apic" driver can be used only when x2APIC mode is enabled.
> However when "intremap=off" and "x2apic_phys" kernel command line
> parameters are passed, "physical x2apic" driver is being used even when
> x2APIC mode is disabled ("intremap=off" disables x2APIC mode).
> This results in the below kernel panic:
...
Hi! Good catch! In long term I think we could switch to use x2apic_state variable
instead since at the moment the code is somehow hard to read and remember which exactly
deps are to be satisfied to enable x2apic mode.
Reviewed-by: Cyrill Gorcunov <gorcunov@gmail.com>
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH v1] x86/apic: Fix kernel panic when "intremap=off" and "x2apic_phys" are set
2023-06-18 22:24 ` Cyrill Gorcunov
@ 2023-06-19 15:25 ` Thomas Gleixner
0 siblings, 0 replies; 6+ messages in thread
From: Thomas Gleixner @ 2023-06-19 15:25 UTC (permalink / raw)
To: Cyrill Gorcunov, Dheeraj Kumar Srivastava
Cc: linux-kernel, mingo, bp, dave.hansen, x86, hpa, suresh.b.siddha,
Kishon Vijay Abraham I, Vasant Hegde
On Mon, Jun 19 2023 at 01:24, Cyrill Gorcunov wrote:
>
> Hi! Good catch! In long term I think we could switch to use
> x2apic_state variable instead since at the moment the code is somehow
> hard to read and remember which exactly deps are to be satisfied to
> enable x2apic mode.
It's on my todo list for a very long time. Unfortunately that list is
growing only into one direction...
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v1] x86/apic: Fix kernel panic when "intremap=off" and "x2apic_phys" are set
2023-06-16 21:22 [PATCH v1] x86/apic: Fix kernel panic when "intremap=off" and "x2apic_phys" are set Dheeraj Kumar Srivastava
2023-06-18 22:24 ` Cyrill Gorcunov
@ 2023-06-19 14:50 ` Borislav Petkov
2023-06-19 16:31 ` Thomas Gleixner
2023-06-19 19:10 ` [tip: x86/urgent] x86/apic: Fix kernel panic when booting with intremap=off and x2apic_phys tip-bot2 for Dheeraj Kumar Srivastava
3 siblings, 0 replies; 6+ messages in thread
From: Borislav Petkov @ 2023-06-19 14:50 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava
Cc: linux-kernel, tglx, mingo, dave.hansen, x86, hpa, gorcunov,
suresh.b.siddha, Kishon Vijay Abraham I, Vasant Hegde
On Sat, Jun 17, 2023 at 02:52:36AM +0530, Dheeraj Kumar Srivastava wrote:
> x2APIC mode requires "Interrupt Remapping" to be enabled and the
> "physical x2apic" driver can be used only when x2APIC mode is enabled.
> However when "intremap=off" and "x2apic_phys" kernel command line
> parameters are passed, "physical x2apic" driver is being used even when
> x2APIC mode is disabled ("intremap=off" disables x2APIC mode).
> This results in the below kernel panic:
>
> unchecked MSR access error: RDMSR from 0x80f at rIP: 0xffffffff87eab4ec
Not a kernel panic - that's a warning about accessing a non-existent
MSR.
Can you send full dmesg? Privately is fine too.
How exactly do you trigger this? What hw?
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH v1] x86/apic: Fix kernel panic when "intremap=off" and "x2apic_phys" are set
2023-06-16 21:22 [PATCH v1] x86/apic: Fix kernel panic when "intremap=off" and "x2apic_phys" are set Dheeraj Kumar Srivastava
2023-06-18 22:24 ` Cyrill Gorcunov
2023-06-19 14:50 ` Borislav Petkov
@ 2023-06-19 16:31 ` Thomas Gleixner
2023-06-19 19:10 ` [tip: x86/urgent] x86/apic: Fix kernel panic when booting with intremap=off and x2apic_phys tip-bot2 for Dheeraj Kumar Srivastava
3 siblings, 0 replies; 6+ messages in thread
From: Thomas Gleixner @ 2023-06-19 16:31 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava, linux-kernel, mingo, bp, dave.hansen,
x86
Cc: hpa, gorcunov, suresh.b.siddha, Dheeraj Kumar Srivastava,
Kishon Vijay Abraham I, Vasant Hegde
On Sat, Jun 17 2023 at 02:52, Dheeraj Kumar Srivastava wrote:
> x2APIC mode requires "Interrupt Remapping" to be enabled and the
> "physical x2apic" driver can be used only when x2APIC mode is enabled.
> However when "intremap=off" and "x2apic_phys" kernel command line
> parameters are passed, "physical x2apic" driver is being used even when
> x2APIC mode is disabled ("intremap=off" disables x2APIC mode).
> This results in the below kernel panic:
>
> unchecked MSR access error: RDMSR from 0x80f at rIP: 0xffffffff87eab4ec
> (native_read_msr+0xc/0x40)
> Call Trace:
> <TASK>
> native_apic_msr_read+0x1f/0x30
> setup_local_APIC+0x4e/0x380
> ? zen_untrain_ret+0x1/0x1
> ? enable_IR_x2apic+0xe8/0x250
> apic_intr_mode_init+0x4c/0x120
> x86_late_time_init+0x28/0x40
> start_kernel+0x626/0xa80
> x86_64_start_reservations+0x1c/0x30
> x86_64_start_kernel+0xbf/0x110
> secondary_startup_64_no_verify+0x10b/0x10b
> </TASK>
>
> This is due to an incorrect conditional check in x2apic_phys_probe().
> Fix it here by returning 0 when "x2apic_mode" is not set in
> x2apic_phys_probe(). This now prevents default_setup_apic_routing() from
> selecting "physical x2apic" driver.
>
> Fixes: 9ebd680bd029 ("x86, apic: Use probe routines to simplify apic selection")
> Reviewed-by: Kishon Vijay Abraham I <kvijayab@amd.com>
> Reviewed-by: Vasant Hegde <vasant.hegde@amd.com>
> Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
^ permalink raw reply [flat|nested] 6+ messages in thread* [tip: x86/urgent] x86/apic: Fix kernel panic when booting with intremap=off and x2apic_phys
2023-06-16 21:22 [PATCH v1] x86/apic: Fix kernel panic when "intremap=off" and "x2apic_phys" are set Dheeraj Kumar Srivastava
` (2 preceding siblings ...)
2023-06-19 16:31 ` Thomas Gleixner
@ 2023-06-19 19:10 ` tip-bot2 for Dheeraj Kumar Srivastava
3 siblings, 0 replies; 6+ messages in thread
From: tip-bot2 for Dheeraj Kumar Srivastava @ 2023-06-19 19:10 UTC (permalink / raw)
To: linux-tip-commits
Cc: Dheeraj Kumar Srivastava, Borislav Petkov (AMD),
Kishon Vijay Abraham I, Vasant Hegde, Cyrill Gorcunov,
Thomas Gleixner, x86, linux-kernel
The following commit has been merged into the x86/urgent branch of tip:
Commit-ID: 85d38d5810e285d5aec7fb5283107d1da70c12a9
Gitweb: https://git.kernel.org/tip/85d38d5810e285d5aec7fb5283107d1da70c12a9
Author: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
AuthorDate: Sat, 17 Jun 2023 02:52:36 +05:30
Committer: Borislav Petkov (AMD) <bp@alien8.de>
CommitterDate: Mon, 19 Jun 2023 20:59:40 +02:00
x86/apic: Fix kernel panic when booting with intremap=off and x2apic_phys
When booting with "intremap=off" and "x2apic_phys" on the kernel command
line, the physical x2APIC driver ends up being used even when x2APIC
mode is disabled ("intremap=off" disables x2APIC mode). This happens
because the first compound condition check in x2apic_phys_probe() is
false due to x2apic_mode == 0 and so the following one returns true
after default_acpi_madt_oem_check() having already selected the physical
x2APIC driver.
This results in the following panic:
kernel BUG at arch/x86/kernel/apic/io_apic.c:2409!
invalid opcode: 0000 [#1] PREEMPT SMP NOPTI
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.4.0-rc2-ver4.1rc2 #2
Hardware name: Dell Inc. PowerEdge R6515/07PXPY, BIOS 2.3.6 07/06/2021
RIP: 0010:setup_IO_APIC+0x9c/0xaf0
Call Trace:
<TASK>
? native_read_msr
apic_intr_mode_init
x86_late_time_init
start_kernel
x86_64_start_reservations
x86_64_start_kernel
secondary_startup_64_no_verify
</TASK>
which is:
setup_IO_APIC:
apic_printk(APIC_VERBOSE, "ENABLING IO-APIC IRQs\n");
for_each_ioapic(ioapic)
BUG_ON(mp_irqdomain_create(ioapic));
Return 0 to denote that x2APIC has not been enabled when probing the
physical x2APIC driver.
[ bp: Massage commit message heavily. ]
Fixes: 9ebd680bd029 ("x86, apic: Use probe routines to simplify apic selection")
Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Kishon Vijay Abraham I <kvijayab@amd.com>
Reviewed-by: Vasant Hegde <vasant.hegde@amd.com>
Reviewed-by: Cyrill Gorcunov <gorcunov@gmail.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20230616212236.1389-1-dheerajkumar.srivastava@amd.com
---
arch/x86/kernel/apic/x2apic_phys.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kernel/apic/x2apic_phys.c b/arch/x86/kernel/apic/x2apic_phys.c
index 6bde05a..896bc41 100644
--- a/arch/x86/kernel/apic/x2apic_phys.c
+++ b/arch/x86/kernel/apic/x2apic_phys.c
@@ -97,7 +97,10 @@ static void init_x2apic_ldr(void)
static int x2apic_phys_probe(void)
{
- if (x2apic_mode && (x2apic_phys || x2apic_fadt_phys()))
+ if (!x2apic_mode)
+ return 0;
+
+ if (x2apic_phys || x2apic_fadt_phys())
return 1;
return apic == &apic_x2apic_phys;
^ permalink raw reply related [flat|nested] 6+ messages in thread