arm/arm64: xen: Fix to convert percpu address to gfn correctly
authorMasami Hiramatsu <mhiramat@kernel.org>
Tue, 6 Oct 2020 06:49:31 +0000 (15:49 +0900)
committerJuergen Gross <jgross@suse.com>
Wed, 7 Oct 2020 05:08:43 +0000 (07:08 +0200)
commit5a0677110b73dd3e1766f89159701bfe8ac06808
tree65880f64f5a68f436dced52db8e0a9893989cd00
parent0891fb39ba67bd7ae023ea0d367297ffff010781
arm/arm64: xen: Fix to convert percpu address to gfn correctly

Use per_cpu_ptr_to_phys() instead of virt_to_phys() for per-cpu
address conversion.

In xen_starting_cpu(), per-cpu xen_vcpu_info address is converted
to gfn by virt_to_gfn() macro. However, since the virt_to_gfn(v)
assumes the given virtual address is in linear mapped kernel memory
area, it can not convert the per-cpu memory if it is allocated on
vmalloc area.

This depends on CONFIG_NEED_PER_CPU_EMBED_FIRST_CHUNK.
If it is enabled, the first chunk of percpu memory is linear mapped.
In the other case, that is allocated from vmalloc area. Moreover,
if the first chunk of percpu has run out until allocating
xen_vcpu_info, it will be allocated on the 2nd chunk, which is
based on kernel memory or vmalloc memory (depends on
CONFIG_NEED_PER_CPU_KM).

Without this fix and kernel configured to use vmalloc area for
the percpu memory, the Dom0 kernel will fail to boot with following
errors.

[    0.466172] Xen: initializing cpu0
[    0.469601] ------------[ cut here ]------------
[    0.474295] WARNING: CPU: 0 PID: 1 at arch/arm64/xen/../../arm/xen/enlighten.c:153 xen_starting_cpu+0x160/0x180
[    0.484435] Modules linked in:
[    0.487565] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.9.0-rc4+ #4
[    0.493895] Hardware name: Socionext Developer Box (DT)
[    0.499194] pstate: 00000005 (nzcv daif -PAN -UAO BTYPE=--)
[    0.504836] pc : xen_starting_cpu+0x160/0x180
[    0.509263] lr : xen_starting_cpu+0xb0/0x180
[    0.513599] sp : ffff8000116cbb60
[    0.516984] x29: ffff8000116cbb60 x28: ffff80000abec000
[    0.522366] x27: 0000000000000000 x26: 0000000000000000
[    0.527754] x25: ffff80001156c000 x24: fffffdffbfcdb600
[    0.533129] x23: 0000000000000000 x22: 0000000000000000
[    0.538511] x21: ffff8000113a99c8 x20: ffff800010fe4f68
[    0.543892] x19: ffff8000113a9988 x18: 0000000000000010
[    0.549274] x17: 0000000094fe0f81 x16: 00000000deadbeef
[    0.554655] x15: ffffffffffffffff x14: 0720072007200720
[    0.560037] x13: 0720072007200720 x12: 0720072007200720
[    0.565418] x11: 0720072007200720 x10: 0720072007200720
[    0.570801] x9 : ffff8000100fbdc0 x8 : ffff800010715208
[    0.576182] x7 : 0000000000000054 x6 : ffff00001b790f00
[    0.581564] x5 : ffff800010bbf880 x4 : 0000000000000000
[    0.586945] x3 : 0000000000000000 x2 : ffff80000abec000
[    0.592327] x1 : 000000000000002f x0 : 0000800000000000
[    0.597716] Call trace:
[    0.600232]  xen_starting_cpu+0x160/0x180
[    0.604309]  cpuhp_invoke_callback+0xac/0x640
[    0.608736]  cpuhp_issue_call+0xf4/0x150
[    0.612728]  __cpuhp_setup_state_cpuslocked+0x128/0x2c8
[    0.618030]  __cpuhp_setup_state+0x84/0xf8
[    0.622192]  xen_guest_init+0x324/0x364
[    0.626097]  do_one_initcall+0x54/0x250
[    0.630003]  kernel_init_freeable+0x12c/0x2c8
[    0.634428]  kernel_init+0x1c/0x128
[    0.637988]  ret_from_fork+0x10/0x18
[    0.641635] ---[ end trace d95b5309a33f8b27 ]---
[    0.646337] ------------[ cut here ]------------
[    0.651005] kernel BUG at arch/arm64/xen/../../arm/xen/enlighten.c:158!
[    0.657697] Internal error: Oops - BUG: 0 [#1] SMP
[    0.662548] Modules linked in:
[    0.665676] CPU: 0 PID: 1 Comm: swapper/0 Tainted: G        W         5.9.0-rc4+ #4
[    0.673398] Hardware name: Socionext Developer Box (DT)
[    0.678695] pstate: 00000005 (nzcv daif -PAN -UAO BTYPE=--)
[    0.684338] pc : xen_starting_cpu+0x178/0x180
[    0.688765] lr : xen_starting_cpu+0x144/0x180
[    0.693188] sp : ffff8000116cbb60
[    0.696573] x29: ffff8000116cbb60 x28: ffff80000abec000
[    0.701955] x27: 0000000000000000 x26: 0000000000000000
[    0.707344] x25: ffff80001156c000 x24: fffffdffbfcdb600
[    0.712718] x23: 0000000000000000 x22: 0000000000000000
[    0.718107] x21: ffff8000113a99c8 x20: ffff800010fe4f68
[    0.723481] x19: ffff8000113a9988 x18: 0000000000000010
[    0.728863] x17: 0000000094fe0f81 x16: 00000000deadbeef
[    0.734245] x15: ffffffffffffffff x14: 0720072007200720
[    0.739626] x13: 0720072007200720 x12: 0720072007200720
[    0.745008] x11: 0720072007200720 x10: 0720072007200720
[    0.750390] x9 : ffff8000100fbdc0 x8 : ffff800010715208
[    0.755771] x7 : 0000000000000054 x6 : ffff00001b790f00
[    0.761153] x5 : ffff800010bbf880 x4 : 0000000000000000
[    0.766534] x3 : 0000000000000000 x2 : 00000000deadbeef
[    0.771916] x1 : 00000000deadbeef x0 : ffffffffffffffea
[    0.777304] Call trace:
[    0.779819]  xen_starting_cpu+0x178/0x180
[    0.783898]  cpuhp_invoke_callback+0xac/0x640
[    0.788325]  cpuhp_issue_call+0xf4/0x150
[    0.792317]  __cpuhp_setup_state_cpuslocked+0x128/0x2c8
[    0.797619]  __cpuhp_setup_state+0x84/0xf8
[    0.801779]  xen_guest_init+0x324/0x364
[    0.805683]  do_one_initcall+0x54/0x250
[    0.809590]  kernel_init_freeable+0x12c/0x2c8
[    0.814016]  kernel_init+0x1c/0x128
[    0.817583]  ret_from_fork+0x10/0x18
[    0.821226] Code: d0006980 f9427c00 cb000300 17ffffea (d4210000)
[    0.827415] ---[ end trace d95b5309a33f8b28 ]---
[    0.832076] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
[    0.839815] ---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b ]---

Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
Link: https://lore.kernel.org/r/160196697165.60224.17470743378683334995.stgit@devnote2
Signed-off-by: Juergen Gross <jgross@suse.com>
arch/arm/xen/enlighten.c
include/xen/arm/page.h