Commit ea6eac90 authored by Shanker Donthineni's avatar Shanker Donthineni Committed by Will Deacon

arm64: Avoid clobbering mm in erratum workaround on QDF2400

Commit 38fd94b0 ("arm64: Work around Falkor erratum 1003") tried to
work around a hardware erratum, but actually caused a system crash of
its own during switch_mm:

 cpu_do_switch_mm+0x20/0x40
 efi_virtmap_load+0x34/0x40
 virt_efi_get_next_variable+0x64/0xc8
 efivar_init+0x8c/0x348
 efisubsys_init+0xd4/0x270
 do_one_initcall+0x80/0x110
 kernel_init_freeable+0x19c/0x240
 kernel_init+0x10/0x100
 ret_from_fork+0x10/0x50

 Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b

In cpu_do_switch_mm, x1 contains the mm_struct pointer, which needs to
be preserved by the pre_ttbr0_update_workaround macro rather than passed
as a temporary.

This patch clobbers x2 and x3 instead, keeping the mm_struct intact
after the workaround has run.

Fixes: 38fd94b0 ("arm64: Work around Falkor erratum 1003")
Tested-by: default avatarManoj Iyer <manoj.iyer@canonical.com>
Signed-off-by: default avatarShanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: default avatarWill Deacon <will.deacon@arm.com>
parent ffe7afd1
...@@ -138,7 +138,7 @@ ENDPROC(cpu_do_resume) ...@@ -138,7 +138,7 @@ ENDPROC(cpu_do_resume)
* - pgd_phys - physical address of new TTB * - pgd_phys - physical address of new TTB
*/ */
ENTRY(cpu_do_switch_mm) ENTRY(cpu_do_switch_mm)
pre_ttbr0_update_workaround x0, x1, x2 pre_ttbr0_update_workaround x0, x2, x3
mmid x1, x1 // get mm->context.id mmid x1, x1 // get mm->context.id
bfi x0, x1, #48, #16 // set the ASID bfi x0, x1, #48, #16 // set the ASID
msr ttbr0_el1, x0 // set TTBR0 msr ttbr0_el1, x0 // set TTBR0
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment