[PATCH] arch: arm64: fix KERNEL_SEGMENT_COUNT error

[Zhaoyang Huang]

On Thu, Apr 30, 2026 at 9:29 AM Zhaoyang Huang <huangzhaoyang at gmail.com> wrote:
>
> On Wed, Apr 29, 2026 at 5:28 PM Ard Biesheuvel <ardb at kernel.org> wrote:
> >
> > Hello Zhaoyang,
> >
> > On Wed, 29 Apr 2026, at 08:17, zhaoyang.huang wrote:
> > > From: Zhaoyang Huang <zhaoyang.huang at unisoc.com>
> > >
> > > During debug of a kernel panic, we find that the pte_t of the final
> > > part of [data, end] segment got overflow to the next page of
> > > init_pg_end[1] which is the gap page before early_init_stack[2].
> > > This should be introduced by the KERNEL_SEGMENT_COUNT's value is 5
> > > which should be 6 as map_segment are called 6 times for the segments
> > > of (text, stext, rodata, inittext, initdata, data+bss)
> > >
> > > [1]
> > > crash_arm64_v9.0.1> vtop ffffffed00601000
> > > VIRTUAL           PHYSICAL
> > > ffffffed00601000  83401000
> > >
> > > PAGE DIRECTORY: ffffffecffd62000
> > >    PGD: ffffffecffd62da0 => 10000000833fb003
> > >    PMD: ffffff80033fb018 => 10000000833fe003
> > >    PTE: ffffff80033fe008 => 68000083401f03
> > >   PAGE: 83401000
> > >
> > >      PTE        PHYSICAL  FLAGS
> > > 68000083401f03  83401000  (VALID|SHARED|AF|NG|PXN|UXN)
> > >
> > >       PAGE       PHYSICAL      MAPPING       INDEX CNT FLAGS
> > > fffffffec00d0040 83401000                0        0  1 4000 reserved
> > >
> > > [2]
> > > ffffffed002c8000 (r) __pi__data
> > > ffffffed0054e000 (d) __pi___bss_start
> > > ffffffed005f5000 (b) __pi_init_pg_dir
> > > ffffffed005fe000 (b) __pi_init_pg_end
> > > ffffffed005ff000 (B) early_init_stack
> > > ffffffed00608000 (b) __pi__end
> > >
> > > Signed-off-by: Zhaoyang Huang <zhaoyang.huang at unisoc.com>
> > > ---
> > >  arch/arm64/include/asm/kernel-pgtable.h | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/arch/arm64/include/asm/kernel-pgtable.h
> > > b/arch/arm64/include/asm/kernel-pgtable.h
> > > index 74a4f738c5f5..14dd1c52552a 100644
> > > --- a/arch/arm64/include/asm/kernel-pgtable.h
> > > +++ b/arch/arm64/include/asm/kernel-pgtable.h
> > > @@ -64,8 +64,8 @@
> > >  #define INIT_IDMAP_FDT_PAGES (EARLY_PAGES(INIT_IDMAP_PGTABLE_LEVELS,
> > > 0UL, UL(MAX_FDT_SIZE), 1) - 1)
> > >  #define INIT_IDMAP_FDT_SIZE  ((INIT_IDMAP_FDT_PAGES +
> > > EARLY_IDMAP_EXTRA_FDT_PAGES) * PAGE_SIZE)
> > >
> > > -/* The number of segments in the kernel image (text, rodata, inittext,
> > > initdata, data+bss) */
> > > -#define KERNEL_SEGMENT_COUNT 5
> > > +/* The number of segments in the kernel image (text, stext, rodata,
> > > inittext, initdata, data+bss) */
> > > +#define KERNEL_SEGMENT_COUNT 6
> > >
> > >  #if SWAPPER_BLOCK_SIZE > SEGMENT_ALIGN
> > >  #define EARLY_SEGMENT_EXTRA_PAGES (KERNEL_SEGMENT_COUNT + 1)
> >
> > I think the better fix would be
> >
> > --- a/arch/arm64/include/asm/kernel-pgtable.h
> > +++ b/arch/arm64/include/asm/kernel-pgtable.h
> > @@ -68,7 +68,7 @@
> >  #define KERNEL_SEGMENT_COUNT   5
> >
> >  #if SWAPPER_BLOCK_SIZE > SEGMENT_ALIGN
> > -#define EARLY_SEGMENT_EXTRA_PAGES (KERNEL_SEGMENT_COUNT + 1)
> > +#define EARLY_SEGMENT_EXTRA_PAGES (KERNEL_SEGMENT_COUNT + 2)
> Thanks for the comment. This makes sense since KERNEL_SEGMENT_COUNT is
> also used for defining vmlinux_seg[] in declare_kernel_vmas.
Is the following fix more reasonable, since it makes both map_segment
and declare_kernel_vmas compatible with the value 6?

-/* The number of segments in the kernel image (text, rodata,
inittext, initdata, data+bss) */
-#define KERNEL_SEGMENT_COUNT   5
+/* The number of segments in the kernel image (text, stext, rodata,
inittext, initdata, data+bss) */
+#define KERNEL_SEGMENT_COUNT   6

 static void __init declare_kernel_vmas(void)
 {
-       static struct vm_struct vmlinux_seg[KERNEL_SEGMENT_COUNT];
+       /*
+        * [_text, _stext) is mapped in init_pg_dir seperatly while they got
+        * merged into one here
+        */
+       static struct vm_struct vmlinux_seg[KERNEL_SEGMENT_COUNT - 1];


> >  /*
> >   * The initial ID map consists of the kernel image, mapped as two separate
> >   * segments, and may appear misaligned wrt the swapper block size. This means
> >
> > and add a line
> >
> > Fixes: 5973a62efa34 ("arm64: map [_text, _stext) virtual address range non-executable+read-only")
> >
> >