Skip to contentSkip to conversation

xen/riscv: introduce identity mapping

PATCH Superseded Oleksii Kurochko 4 participants 25 messages Jun 19, 2023 – Jul 10, 2023 RISC-V

v2 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
#0 lore raw

v2 cover xen/riscv: introduce identity mapping

The patch series introduces things necessary to implement identity mapping:
1. Make identity mapping for the entire Xen.
2. Enable MMU.
3. Jump to the virtual address world
4. Remove identity mapping.

Also current patch series introduces the calculation of physical offset before MMU is enabled as access to physical offset will be calculated wrong after MMU will be enabled because access to phys_off variable is PC-relative and in the case when linker address != load address, it will cause MMU fault.

One more thing that was done is:
* Added SPDX tags.
* move extern of cpu0_boot_stack to a header.

The reason for this patch series can be found here:
https://lore.kernel.org/xen-devel/[email protected]/

---
Changes in V2:
- update the patch series message.
- drop patches from the previous version of the patch series:
* xen/riscv: add __ASSEMBLY__ guards". ( merged )
* xen/riscv: make sure that identity mapping isn't bigger then page size ( entire Xen is 1:1 mapped so there is no need for the checks from the patch )
- add .sbss.* and put it befor .bss* .
- move out reset_stack() to .text section.
- add '__ro_after_init' for phys_offset variable.
- add '__init' for calc_phys_offset().
- declaring variable phys_off as non static as it will be used in head.S.
- update definition of PGTBL_INITIAL_COUNT and the comment above.
- code style fixes.
- remove id_addrs array becase entire Xen is mapped.
- reverse condition for cycle inside remove_identity_mapping().
- fix page table walk in remove_identity_mapping().
- save hart_id and dtb_addr before call MMU related C functions
- use phys_offset variable instead of doing calcultations to get phys offset in head.S file. ( it can be easily done as entire Xen is 1:1 mapped now )
- declare enable_muu() as __init.
- Update SPDX tags.
- Add Review-By/Suggested-By for some patches.
- code style fixes.

Oleksii Kurochko (6):
xen/riscv: add .sbss section to .bss
xen/riscv: introduce reset_stack() function xen/riscv: introduce function for physical offset calculation xen/riscv: introduce identity mapping
xen/riscv: add SPDX tags
xen/riscv: move extern of cpu0_boot_stack to header

 xen/arch/riscv/include/asm/config.h       |   2 +
 xen/arch/riscv/include/asm/current.h      |   2 +
 xen/arch/riscv/include/asm/early_printk.h |   2 +
 xen/arch/riscv/include/asm/mm.h           |   9 +-
 xen/arch/riscv/include/asm/page-bits.h    |   2 +
 xen/arch/riscv/include/asm/page.h         |   2 +
 xen/arch/riscv/include/asm/traps.h        |   2 +
 xen/arch/riscv/include/asm/types.h        |   2 +
 xen/arch/riscv/mm.c                       | 104 +++++++++++++---------
 xen/arch/riscv/riscv64/head.S             |  46 +++++++++-
 xen/arch/riscv/setup.c                    |  16 +---
 xen/arch/riscv/xen.lds.S                  |   2 +-
 12 files changed, 136 insertions(+), 55 deletions(-)
Signature · 1 line
2.40.1
#1 lore raw

PATCH v2 1/6 xen/riscv: add .sbss section to .bss

Sometimes variables are located in .sbss section but it won't be mapped after MMU will be enabled.
To avoid MMU failures .sbss should be mapped

---
Changes in V2:
- add .sbss.*.
- move .sbss* ahead of .bss*.
- add Acked-by: Alistair Francis <[email protected]>

---
 xen/arch/riscv/xen.lds.S | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
xen/arch/riscv/xen.lds.S +1 -1
index 878130f313..9064852173 100644
@@ -149,7 +149,7 @@ SECTIONS
         *(.bss.percpu.read_mostly)
         . = ALIGN(SMP_CACHE_BYTES);
         __per_cpu_data_end = .;
-        *(.bss .bss.*)
+        *(.sbss .sbss.* .bss .bss.*)
         . = ALIGN(POINTER_ALIGN);
         __bss_end = .;
     } :text
Signature · 1 line
2.40.1
#2 lore raw

PATCH v2 2/6 xen/riscv: introduce reset_stack() function

In reply to Oleksii Kurochko

The reason for reset_stack() introduction is that stack should be reset twice:
1. Before jumping to C world at the start of _start() function.
2. After jumping from 1:1 mapping world.

---
Changes in V2:
- update the commit message.
- move out reset_stack() from .text.header to .text.
- add Reviewed-by: Alistair Francis <[email protected]>.

---
 xen/arch/riscv/riscv64/head.S | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
xen/arch/riscv/riscv64/head.S +9 -1
index 8887f0cbd4..2c0304646a 100644
@@ -27,8 +27,16 @@ ENTRY(start)
         add     t3, t3, __SIZEOF_POINTER__
         bltu    t3, t4, .L_clear_bss
 
+        jal     reset_stack
+
+        tail    start_xen
+
+        .section .text, "ax", %progbits
+
+ENTRY(reset_stack)
         la      sp, cpu0_boot_stack
         li      t0, STACK_SIZE
         add     sp, sp, t0
 
-        tail    start_xen
+        ret
+
Signature · 1 line
2.40.1
#3 lore raw

Re: [PATCH v2 2/6] xen/riscv: introduce reset_stack() function

Quoted 10 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +4 -0
@@ -27,8 +27,16 @@ ENTRY(start)
         add     t3, t3, __SIZEOF_POINTER__
         bltu    t3, t4, .L_clear_bss
 
+        jal     reset_stack
+
+        tail    start_xen
+

xen/arch/riscv/riscv64/head.S ENTRY(start) · line 42

Quoted 10 lines
xen/arch/riscv/riscv64/head.S +5 -1
+        .section .text, "ax", %progbits
+
+ENTRY(reset_stack)
         la      sp, cpu0_boot_stack
         li      t0, STACK_SIZE
         add     sp, sp, t0
 
-        tail    start_xen
+        ret
+

Looking at patch 4 you will want to add a comment here to emphasize that a0 and a1 have to remain unclobbered.

Jan

author
#4 lore raw

Re: [PATCH v2 2/6] xen/riscv: introduce reset_stack() function

Quoted 16 lines · Jan Beulich
Quoted 15 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +0 -0
@@ -27,8 +27,16 @@ ENTRY(start)

         add     t3, t3, __SIZEOF_POINTER__
         bltu    t3, t4, .L_clear_bss

+        jal     reset_stack

xen/arch/riscv/riscv64/head.S +6 -0
+
+        tail    start_xen
+
+        .section .text, "ax", %progbits
+
+ENTRY(reset_stack)

         la      sp, cpu0_boot_stack
         li      t0, STACK_SIZE

Quoted 8 lines
Quoted 5 lines

         add     sp, sp, t0

-        tail    start_xen
+        ret
+

Looking at patch 4 you will want to add a comment here to emphasize that a0 and a1 have to remain unclobbered.

Thanks for a note. I'll add it in the new patch version

~ Oleksii

#5 lore raw

Re: [PATCH v2 2/6] xen/riscv: introduce reset_stack() function

Quoted 25 lines · Oleksii
Quoted 16 lines · Jan Beulich
Quoted 15 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +0 -0
@@ -27,8 +27,16 @@ ENTRY(start)

         add     t3, t3, __SIZEOF_POINTER__
         bltu    t3, t4, .L_clear_bss

+        jal     reset_stack

xen/arch/riscv/riscv64/head.S +6 -0
+
+        tail    start_xen
+
+        .section .text, "ax", %progbits
+
+ENTRY(reset_stack)

         la      sp, cpu0_boot_stack
         li      t0, STACK_SIZE

Quoted 8 lines
Quoted 5 lines

         add     sp, sp, t0

-        tail    start_xen
+        ret
+

Looking at patch 4 you will want to add a comment here to emphasize that a0 and a1 have to remain unclobbered.

Quoted 1 line

Thanks for a note. I'll add it in the new patch version

Having seen how things end up by the end of the series, there's an
alternative: You could save a0 and a1 ahead of the 1st call to reset_stack, rather than immediately afterwards.

Jan

author
#6 lore raw

Re: [PATCH v2 2/6] xen/riscv: introduce reset_stack() function

Quoted 27 lines · Jan Beulich
Quoted 26 lines · Oleksii
Quoted 25 lines · Jan Beulich
Quoted 13 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +0 -0
@@ -27,8 +27,16 @@ ENTRY(start)

         add     t3, t3, __SIZEOF_POINTER__
         bltu    t3, t4, .L_clear_bss

+        jal     reset_stack

xen/arch/riscv/riscv64/head.S +6 -0
+
+        tail    start_xen
+
+        .section .text, "ax", %progbits
+
+ENTRY(reset_stack)
Quoted 7 lines

         la      sp, cpu0_boot_stack
         li      t0, STACK_SIZE
         add     sp, sp, t0

-        tail    start_xen
+        ret
+

Looking at patch 4 you will want to add a comment here to emphasize
that a0 and a1 have to remain unclobbered.

Quoted 5 lines
Quoted 1 line

Thanks for a note. I'll add it in the new patch version

Having seen how things end up by the end of the series, there's an
alternative: You could save a0 and a1 ahead of the 1st call to reset_stack, rather than immediately afterwards.

It makes sense. So lets stick to saving of a0 and a1 before 1st call of reset_stack().

~ Oleksii

#7 lore raw

PATCH v2 3/6 xen/riscv: introduce function for physical offset calculation

In reply to Oleksii Kurochko

The function was introduced to calculate and save physical offset before MMU is enabled because access to start() is PC-relative and in case of linker_addr != load_addr it will result in incorrect value in phys_offset.

---
Changes in V2:
- add __ro_after_init for phys_offset variable.
- remove double blank lines.
- add __init for calc_phys_offset().
- update the commit message.
- declaring variable phys_off as non static as it will be used in head.S.

---
 xen/arch/riscv/include/asm/mm.h |  2 ++
 xen/arch/riscv/mm.c             | 18 +++++++++++++++---
 xen/arch/riscv/riscv64/head.S   |  2 ++
 3 files changed, 19 insertions(+), 3 deletions(-)
xen/arch/riscv/include/asm/mm.h +2 -0
index 64293eacee..996041ce81 100644
@@ -11,4 +11,6 @@ void setup_initial_pagetables(void);
 void enable_mmu(void);
 void cont_after_mmu_is_enabled(void);
 
+void calc_phys_offset(void);
+
 #endif /* _ASM_RISCV_MM_H */
xen/arch/riscv/mm.c +15 -3
index 8ceed445cf..570033f17f 100644
@@ -1,3 +1,4 @@
+#include <xen/cache.h>
 #include <xen/compiler.h>
 #include <xen/init.h>
 #include <xen/kernel.h>
@@ -19,9 +20,10 @@ struct mmu_desc {
 
 extern unsigned char cpu0_boot_stack[STACK_SIZE];
 
-#define PHYS_OFFSET ((unsigned long)_start - XEN_VIRT_START)
-#define LOAD_TO_LINK(addr) ((addr) - PHYS_OFFSET)
-#define LINK_TO_LOAD(addr) ((addr) + PHYS_OFFSET)
+unsigned long __ro_after_init phys_offset;
+
+#define LOAD_TO_LINK(addr) ((unsigned long)(addr) - phys_offset)
+#define LINK_TO_LOAD(addr) ((unsigned long)(addr) + phys_offset)
 
 /*
  * It is expected that Xen won't be more then 2 MB.
@@ -273,3 +275,13 @@ void __init noreturn noinline enable_mmu()
     switch_stack_and_jump((unsigned long)cpu0_boot_stack + STACK_SIZE,
                           cont_after_mmu_is_enabled);
 }
+
+/*
+ * calc_phys_offset() should be used before MMU is enabled because access to
+ * start() is PC-relative and in case when load_addr != linker_addr phys_offset
+ * will have an incorrect value
+ */
+void __init calc_phys_offset(void)
+{
+    phys_offset = (unsigned long)start - XEN_VIRT_START;
+}
xen/arch/riscv/riscv64/head.S +2 -0
index 2c0304646a..850fbb58a7 100644
@@ -29,6 +29,8 @@ ENTRY(start)
 
         jal     reset_stack
 
+        jal     calc_phys_offset
+
         tail    start_xen
 
         .section .text, "ax", %progbits
Signature · 1 line
2.40.1
#8 lore raw

Re: [PATCH v2 3/6] xen/riscv: introduce function for physical offset calculation

xen/arch/riscv/riscv64/head.S ENTRY(start) · line 36

Quoted 11 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +2 -0
@@ -29,6 +29,8 @@ ENTRY(start)
 
         jal     reset_stack
 
+        jal     calc_phys_offset
+
         tail    start_xen
 
         .section .text, "ax", %progbits

Since you call a C function, the code to save/restore a0/a1 needs to move here (from patch 4).

Jan

author
#9 lore raw

Re: [PATCH v2 3/6] xen/riscv: introduce function for physical offset calculation

Quoted 15 lines · Jan Beulich
Quoted 11 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +0 -0
@@ -29,6 +29,8 @@ ENTRY(start)


         jal     reset_stack

+        jal     calc_phys_offset
+
         tail    start_xen

         .section .text, "ax", %progbits

Since you call a C function, the code to save/restore a0/a1 needs to move here (from patch 4).

Thanks. It makes sense.
It would be better to move save/restore a0/a1 ( from patch 4 )code here.

The only one reason I didn't do that before that calc_phys_offset doesn't touch that and it is guaranteed that it will not ( as it doesn't have arguments )

~ Oleksii

#10 lore raw

Re: [PATCH v2 3/6] xen/riscv: introduce function for physical offset calculation

Quoted 16 lines · Oleksii
Quoted 7 lines · Jan Beulich
Quoted 6 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +0 -0
@@ -29,6 +29,8 @@ ENTRY(start)
 
          jal     reset_stack
 
Quoted 8 lines

xen/arch/riscv/riscv64/head.S ENTRY(start) · line 36

Quoted 5 lines
xen/arch/riscv/riscv64/head.S +2 -0
+        jal     calc_phys_offset
+
          tail    start_xen
 
          .section .text, "ax", %progbits

Since you call a C function, the code to save/restore a0/a1 needs to move here (from patch 4).

Quoted 7 lines

Thanks. It makes sense.
It would be better to move save/restore a0/a1 ( from patch 4 )code here.

The only one reason I didn't do that before that calc_phys_offset doesn't touch that and it is guaranteed that it will not ( as it doesn't have arguments )

IIUC, the calling convention requires a0/a1 to be caller saved. So even if they are not used for arguments, such callee is still free to use them for internal purpose.

Cheers,

Signature · 1 line
Julien Grall
author
#11 lore raw

Re: [PATCH v2 3/6] xen/riscv: introduce function for physical offset calculation

Quoted 24 lines · Julien Grall
Quoted 21 lines · Oleksii
Quoted 16 lines · Jan Beulich
Quoted 11 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +0 -0
@@ -29,6 +29,8 @@ ENTRY(start)


          jal     reset_stack

+        jal     calc_phys_offset
+
          tail    start_xen

          .section .text, "ax", %progbits

Since you call a C function, the code to save/restore a0/a1 needs to
move here (from patch 4).

Thanks. It makes sense.
It would be better to move save/restore a0/a1 ( from patch 4 )code here.

Quoted 8 lines
Quoted 3 lines

The only one reason I didn't do that before that calc_phys_offset doesn't touch that and it is guaranteed that it will not ( as it doesn't have arguments )

IIUC, the calling convention requires a0/a1 to be caller saved. So even
if they are not used for arguments, such callee is still free to use them for internal purpose.

You are right.

I haven't seen that compiler use them if 'void' is passed to function as an argument. But I agree that we have to follow the calling convention to be sure that all is fine.

~ Oleksii

#12 lore raw

Re: [PATCH v2 3/6] xen/riscv: introduce function for physical offset calculation

In reply to Oleksii
Quoted 16 lines · Oleksii
Quoted 15 lines · Jan Beulich
Quoted 11 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +0 -0
@@ -29,6 +29,8 @@ ENTRY(start)


         jal     reset_stack

+        jal     calc_phys_offset
+
         tail    start_xen

         .section .text, "ax", %progbits

Since you call a C function, the code to save/restore a0/a1 needs to move here (from patch 4).

Quoted 7 lines

Thanks. It makes sense.
It would be better to move save/restore a0/a1 ( from patch 4 )code here.

The only one reason I didn't do that before that calc_phys_offset doesn't touch that and it is guaranteed that it will not ( as it doesn't have arguments )

How does a function not having parameters guarantee that registers used for parameter passing aren't touched? Inside a function, the compiler is free to use argument-passing registers just like other temporary ones; their values don't need preserving, from all I know (otherwise the RISC-V ABI would be different to all other ABIs I know of).

Jan

author
#13 lore raw

Re: [PATCH v2 3/6] xen/riscv: introduce function for physical offset calculation

Quoted 26 lines · Jan Beulich
Quoted 17 lines · Oleksii
Quoted 16 lines · Jan Beulich
Quoted 11 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +0 -0
@@ -29,6 +29,8 @@ ENTRY(start)


         jal     reset_stack

+        jal     calc_phys_offset
+
         tail    start_xen

         .section .text, "ax", %progbits

Since you call a C function, the code to save/restore a0/a1 needs to
move here (from patch 4).

Quoted 7 lines

Thanks. It makes sense.
It would be better to move save/restore a0/a1 ( from patch 4 )code here.

The only one reason I didn't do that before that calc_phys_offset doesn't touch that and it is guaranteed that it will not ( as it doesn't have arguments )

Quoted 6 lines

How does a function not having parameters guarantee that registers used for parameter passing aren't touched? Inside a function, the compiler is free to use argument-passing registers just like other temporary ones; their values don't need preserving, from all I know (otherwise the RISC-V ABI would be different to all other ABIs I know of).

Well, you are right that it doesn't guarantee and the calling convention tells that arg registers should be saved/restored before/after function call.

But I haven't seen yet that compiler touch arg registers if function accepts 'void' as an function argument. So 'guarantee' isn't correct word.

Thanks for the note.

~ Oleksii

#14 lore raw

PATCH v2 4/6 xen/riscv: introduce identity mapping

In reply to Oleksii Kurochko

The way how switch to virtual address was implemented in the commit e66003e7be ("xen/riscv: introduce setup_initial_pages") isn't safe enough as:
* enable_mmu() depends on hooking all exceptions
and pagefault.
* Any exception other than pagefault, or not taking a pagefault causes it to malfunction, which means you will fail to boot depending on where Xen was loaded into memory.

Instead of the proposed way of switching to virtual addresses was decided to use identity mapping of the entrire Xen and after switching to virtual addresses identity mapping is removed from page-tables.
Since it is not easy to keep track where the identity map was mapped, so we will look for the top-level entry exclusive to the identity map and remove it.

Fixes e66003e7be ("xen/riscv: introduce setup_initial_pages")

---
Changes in V2:
- update definition of PGTBL_INITIAL_COUNT and the comment above.
- code style fixes.
- 1:1 mapping for entire Xen.
- remove id_addrs array becase entire Xen is mapped.
- reverse condition for cycle inside remove_identity_mapping().
- fix page table walk in remove_identity_mapping().
- update the commit message.
- add Suggested-by: Andrew Cooper <[email protected]>
- save hart_id and dtb_addr before call MMU related C functions.
- use phys_offset variable instead of doing calcultations to get phys offset in head.S file. ( it can be easily done as entire Xen is 1:1 mapped )
- declare enable_muu() as __init.

---
 xen/arch/riscv/include/asm/config.h |  2 +
 xen/arch/riscv/include/asm/mm.h     |  3 +-
 xen/arch/riscv/mm.c                 | 84 ++++++++++++++++-------------
 xen/arch/riscv/riscv64/head.S       | 34 ++++++++++++
 xen/arch/riscv/setup.c              | 14 +----
 5 files changed, 88 insertions(+), 49 deletions(-)
xen/arch/riscv/include/asm/config.h +2 -0
index 38862df0b8..fa90ae0898 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __RISCV_CONFIG_H__
 #define __RISCV_CONFIG_H__
 
xen/arch/riscv/include/asm/mm.h +2 -1
index 996041ce81..500fdc9c5a 100644
@@ -9,7 +9,8 @@
 void setup_initial_pagetables(void);
 
 void enable_mmu(void);
-void cont_after_mmu_is_enabled(void);
+
+void remove_identity_mapping(void);
 
 void calc_phys_offset(void);
 
xen/arch/riscv/mm.c +48 -36

131 lines not on this page · raw message

xen/arch/riscv/riscv64/head.S +34 -0
index 850fbb58a7..41983ffe63 100644
@@ -29,8 +29,42 @@ ENTRY(start)
 
         jal     reset_stack
 
+        /*
+         * save hart_id and dtb_base as a0 and a1 register can be used
+         * by C code ( f.e. setup_initial_pagetables will update a0 and
+         * a1 )
+         */
+        mv      s0, a0
+        mv      s1, a1
+
         jal     calc_phys_offset
 
+        jal     setup_initial_pagetables
+
+        jal     enable_mmu
+
+        la      t1, phys_offset
+        REG_L   t1, (t1)
+
+        /* Calculate proper VA after jump from 1:1 mapping */
+        la      t0, .L_primary_switched
+        sub     t0, t0, t1
+
+        /* Jump from 1:1 mapping world */
+        jr      t0
+
+.L_primary_switched:
+        /*
+         * cpu0_boot_stack address is 1:1 mapping related so it should be
+         * recalculated after jump from 1:1 mapping world as 1:1 mapping
+         * will be removed soon in start_xen().
+         */
+        jal     reset_stack
+
+        /* restore bootcpu_id and dtb address */
+        mv      a0, s0
+        mv      a1, s1
+
         tail    start_xen
 
         .section .text, "ax", %progbits
xen/arch/riscv/setup.c +2 -12
index 845d18d86f..c4ef0b3165 100644
@@ -11,20 +11,10 @@ unsigned char __initdata cpu0_boot_stack[STACK_SIZE]
 void __init noreturn start_xen(unsigned long bootcpu_id,
                                paddr_t dtb_addr)
 {
-    early_printk("Hello from C env\n");
-
-    setup_initial_pagetables();
-
-    enable_mmu();
-
-    for ( ;; )
-        asm volatile ("wfi");
+    remove_identity_mapping();
 
-    unreachable();
-}
+    early_printk("Hello from C env\n");
 
-void __init noreturn cont_after_mmu_is_enabled(void)
-{
     early_printk("All set up\n");
 
     for ( ;; )
Signature · 1 line
2.40.1
#15 lore raw
Quoted 3 lines · Oleksii Kurochko

Since it is not easy to keep track where the identity map was mapped, so we will look for the top-level entry exclusive to the identity map and remove it.

I think you mean "top-most" or some such, as it's not necessarily the top-level entry you zap.

xen/arch/riscv/include/asm/config.h line 5

Quoted 8 lines
xen/arch/riscv/include/asm/config.h +2 -0
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __RISCV_CONFIG_H__
 #define __RISCV_CONFIG_H__
 

Unrelated change?

Quoted 16 lines
xen/arch/riscv/mm.c +6 -0
@@ -25,6 +25,12 @@ unsigned long __ro_after_init phys_offset;
 #define LOAD_TO_LINK(addr) ((unsigned long)(addr) - phys_offset)
 #define LINK_TO_LOAD(addr) ((unsigned long)(addr) + phys_offset)
 
+/*
+ * Should be removed as soon as enough headers will be merged for inclusion of
+ * <xen/lib.h>.
+ */
+#define ARRAY_SIZE(arr)		(sizeof(arr) / sizeof((arr)[0]))
+
 /*
  * It is expected that Xen won't be more then 2 MB.
  * The check in xen.lds.S guarantees that.

xen/arch/riscv/mm.c unsigned long __ro_after_init phys_offset; · line 47

Quoted 8 lines
xen/arch/riscv/mm.c +3 -1
@@ -35,8 +41,10 @@ unsigned long __ro_after_init phys_offset;
  *
  * It might be needed one more page table in case when Xen load address
  * isn't 2 MB aligned.
+ *
+ * (CONFIG_PAGING_LEVELS - 1) page tables are needed for identity mapping.
  */
-#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) + 1)
+#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) * 2 + 1)

How come the extra page (see the comment sentence in context) isn't needed for the identity-mapping case?

Quoted 22 lines
xen/arch/riscv/mm.c +2 -15
@@ -255,25 +262,30 @@ void __init noreturn noinline enable_mmu()
     csr_write(CSR_SATP,
               PFN_DOWN((unsigned long)stage1_pgtbl_root) |
               RV_STAGE1_MODE << SATP_MODE_SHIFT);
+}
 
-    asm volatile ( ".p2align 2" );
- mmu_is_enabled:
-    /*
-     * Stack should be re-inited as:
-     * 1. Right now an address of the stack is relative to load time
-     *    addresses what will cause an issue in case of load start address
-     *    isn't equal to linker start address.
-     * 2. Addresses in stack are all load time relative which can be an
-     *    issue in case when load start address isn't equal to linker
-     *    start address.
-     *
-     * We can't return to the caller because the stack was reseted
-     * and it may have stash some variable on the stack.
-     * Jump to a brand new function as the stack was reseted
-     */
+void __init remove_identity_mapping(void)

xen/arch/riscv/mm.c void __init noreturn noinline enable_mmu() · line 275

Quoted 10 lines
xen/arch/riscv/mm.c +7 -2
+{
+    unsigned int i;
+    pte_t *pgtbl;
+    unsigned int index, xen_index;
+    unsigned long load_addr = LINK_TO_LOAD(_start);
 
-    switch_stack_and_jump((unsigned long)cpu0_boot_stack + STACK_SIZE,
-                          cont_after_mmu_is_enabled);
+    for ( pgtbl = stage1_pgtbl_root, i = 0;
+          i <= (CONFIG_PAGING_LEVELS - 1);

i < CONFIG_PAGING_LEVELS ? But maybe it would be easier for i to start at CONFIG_PAGING_LEVELS and be decremented, simplifying ...

xen/arch/riscv/mm.c void __init noreturn noinline enable_mmu() · line 279

Quoted 4 lines
xen/arch/riscv/mm.c +4 -0
+          i++ )
+    {
+        index = pt_index(CONFIG_PAGING_LEVELS - 1 - i, load_addr);
+        xen_index = pt_index(CONFIG_PAGING_LEVELS - 1 - i, XEN_VIRT_START);

... these two expressions?

xen/arch/riscv/mm.c void __init noreturn noinline enable_mmu() · line 285

Quoted 5 lines
xen/arch/riscv/mm.c +5 -0
+        if ( index != xen_index )
+        {
+            pgtbl[index].pte = 0;
+            break;
+        }

Is this enough? When load and link address are pretty close (but not overlapping), can't they share a leaf table, in which case you need to clear more than just a single entry? The present overlap check looks to be 4k-granular, not 2M (in which case this couldn't happen; IOW adjusting the overlap check may also be a way out).

Jan

author
#16 lore raw
Quoted 8 lines · Jan Beulich
Quoted 4 lines · Oleksii Kurochko

Since it is not easy to keep track where the identity map was mapped,
so we will look for the top-level entry exclusive to the identity map and remove it.

I think you mean "top-most" or some such, as it's not necessarily the top-level entry you zap.

Yeah, 'top-most' is more appropriate in this context.

Quoted 11 lines
Quoted 9 lines
xen/arch/riscv/include/asm/config.h +2 -0
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
  #ifndef __RISCV_CONFIG_H__
  #define __RISCV_CONFIG_H__

Unrelated change?

It should be part of [PATCH v2 5/6] xen/riscv: introduce identity mapping.

Quoted 24 lines
Quoted 24 lines
xen/arch/riscv/mm.c +0 -0
@@ -25,6 +25,12 @@ unsigned long __ro_after_init phys_offset;
  #define LOAD_TO_LINK(addr) ((unsigned long)(addr) - phys_offset)
  #define LINK_TO_LOAD(addr) ((unsigned long)(addr) + phys_offset)


+/*
+ * Should be removed as soon as enough headers will be merged for inclusion of
+ * <xen/lib.h>.
+ */
+#define ARRAY_SIZE(arr)                (sizeof(arr) / sizeof((arr)[0]))
+
 /*
  * It is expected that Xen won't be more then 2 MB.
  * The check in xen.lds.S guarantees that.

xen/arch/riscv/mm.c +0 -0
@@ -35,8 +41,10 @@ unsigned long __ro_after_init phys_offset;
   *
   * It might be needed one more page table in case when Xen load

address
  * isn't 2 MB aligned.
+ *

Quoted 8 lines

xen/arch/riscv/mm.c unsigned long __ro_after_init phys_offset; · line 47

Quoted 5 lines

+ * (CONFIG_PAGING_LEVELS - 1) page tables are needed for identity mapping.
  */

xen/arch/riscv/mm.c +1 -1
-#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) + 1)
+#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) * 2 + 1)

How come the extra page (see the comment sentence in context) isn't needed for the identity-mapping case?

It is needed to allocate no more than two 'nonroot' page tables (L0 and L1 in case of Sv39 ) as page 'root' table ( L2 in case of Sv39 ) is always re-used.

The same ( only 'nonroot' page tables might be needed to allocate ) works for any MMU mode.

Quoted 43 lines
Quoted 32 lines
xen/arch/riscv/mm.c +0 -0
@@ -255,25 +262,30 @@ void __init noreturn noinline enable_mmu()
      csr_write(CSR_SATP,

               PFN_DOWN((unsigned long)stage1_pgtbl_root) | RV_STAGE1_MODE << SATP_MODE_SHIFT);
+}

-    asm volatile ( ".p2align 2" );
- mmu_is_enabled:
-    /*
-     * Stack should be re-inited as:
-     * 1. Right now an address of the stack is relative to load time
-     *    addresses what will cause an issue in case of load start address
-     *    isn't equal to linker start address.
-     * 2. Addresses in stack are all load time relative which can be an
-     *    issue in case when load start address isn't equal to linker
-     *    start address.
-     *
-     * We can't return to the caller because the stack was reseted -     * and it may have stash some variable on the stack. -     * Jump to a brand new function as the stack was reseted -     */

xen/arch/riscv/mm.c +6 -0
+void __init remove_identity_mapping(void)
+{
+    unsigned int i;
+    pte_t *pgtbl;
+    unsigned int index, xen_index;
+    unsigned long load_addr = LINK_TO_LOAD(_start);
Quoted 6 lines


-    switch_stack_and_jump((unsigned long)cpu0_boot_stack + STACK_SIZE,
-                          cont_after_mmu_is_enabled); +    for ( pgtbl = stage1_pgtbl_root, i = 0;
+          i <= (CONFIG_PAGING_LEVELS - 1);

i < CONFIG_PAGING_LEVELS ? But maybe it would be easier for i to start
at CONFIG_PAGING_LEVELS and be decremented, simplifying ...

Quoted 7 lines
Quoted 5 lines

+          i++ )
+    {
+        index = pt_index(CONFIG_PAGING_LEVELS - 1 - i, load_addr); +        xen_index = pt_index(CONFIG_PAGING_LEVELS - 1 - i, XEN_VIRT_START);

... these two expressions?

It makes sense. I'll update this part of the code.

Quoted 7 lines
Quoted 6 lines

+        if ( index != xen_index )
+        {
+            pgtbl[index].pte = 0;
+            break;
+        }

Quoted 5 lines

Is this enough? When load and link address are pretty close (but not overlapping), can't they share a leaf table, in which case you need to clear more than just a single entry? The present overlap check looks to be 4k-granular, not 2M (in which case this couldn't happen; IOW adjusting the overlap check may also be a way out).

At the start of setup_initial_pagetables() there is a code which checks that load and link address don't overlap:

    if ( (linker_start != load_start) &&
         (linker_start <= load_end) && (load_start <= linker_end) )
    {
        early_printk("(XEN) linker and load address ranges overlap\n");
        die();
    }

So the closest difference between load and link address can be 4kb. Otherwise load and link address ranges are equal ( as we can't map less then 4kb).

Let's take concrete examples:
Load address range is 0x8020_0000 - 0x8020_0FFF
Linker address range is 0x8020_1000 - 0x8020_1FFF MMU mode: Sv39 ( so we have 3 page tables )

So we have:
* L2 index = 2, L1 index = 1, L0 index = 0 for load address
* L2 index = 2, L1 index = 1, L0 index = 1 for linker address
Thereby we have two different L0 tables for load and linker address
ranges.
And it looks like it is pretty safe to remove only one L0 index that
was used for identity mapping.

Is it possible that I missed something?

~ Oleksii

#17 lore raw
Quoted 12 lines · Oleksii
Quoted 11 lines · Jan Beulich
Quoted 8 lines · Oleksii Kurochko
xen/arch/riscv/include/asm/config.h +2 -0
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
  #ifndef __RISCV_CONFIG_H__
  #define __RISCV_CONFIG_H__

Unrelated change?

Quoted 2 lines

It should be part of [PATCH v2 5/6] xen/riscv: introduce identity mapping.

Hmm, here we're discussing "[PATCH v2 4/6] xen/riscv: introduce identity mapping". I'm confused, I guess.

Quoted 31 lines
Quoted 23 lines
Quoted 23 lines
xen/arch/riscv/mm.c +0 -0
@@ -25,6 +25,12 @@ unsigned long __ro_after_init phys_offset;
  #define LOAD_TO_LINK(addr) ((unsigned long)(addr) - phys_offset)
  #define LINK_TO_LOAD(addr) ((unsigned long)(addr) + phys_offset)


+/*
+ * Should be removed as soon as enough headers will be merged for inclusion of
+ * <xen/lib.h>.
+ */
+#define ARRAY_SIZE(arr)                (sizeof(arr) / sizeof((arr)[0]))
+
 /*
  * It is expected that Xen won't be more then 2 MB.
  * The check in xen.lds.S guarantees that.

xen/arch/riscv/mm.c +0 -0
@@ -35,8 +41,10 @@ unsigned long __ro_after_init phys_offset;
   *
   * It might be needed one more page table in case when Xen load

address
  * isn't 2 MB aligned.
+ *

Quoted 8 lines

xen/arch/riscv/mm.c unsigned long __ro_after_init phys_offset; · line 47

Quoted 5 lines

+ * (CONFIG_PAGING_LEVELS - 1) page tables are needed for identity mapping.
  */

xen/arch/riscv/mm.c +1 -1
-#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) + 1)
+#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) * 2 + 1)

How come the extra page (see the comment sentence in context) isn't needed for the identity-mapping case?

Quoted 6 lines

It is needed to allocate no more than two 'nonroot' page tables (L0 and L1 in case of Sv39 ) as page 'root' table ( L2 in case of Sv39 ) is always re-used.

The same ( only 'nonroot' page tables might be needed to allocate ) works for any MMU mode.

Of course, but if you cross a 2Mb boundary you'll need 2 L0 tables.

Quoted 90 lines
Quoted 42 lines
Quoted 31 lines
xen/arch/riscv/mm.c +0 -0
@@ -255,25 +262,30 @@ void __init noreturn noinline enable_mmu()
      csr_write(CSR_SATP,

               PFN_DOWN((unsigned long)stage1_pgtbl_root) | RV_STAGE1_MODE << SATP_MODE_SHIFT);
+}

-    asm volatile ( ".p2align 2" );
- mmu_is_enabled:
-    /*
-     * Stack should be re-inited as:
-     * 1. Right now an address of the stack is relative to load time
-     *    addresses what will cause an issue in case of load start address
-     *    isn't equal to linker start address.
-     * 2. Addresses in stack are all load time relative which can be an
-     *    issue in case when load start address isn't equal to linker
-     *    start address.
-     *
-     * We can't return to the caller because the stack was reseted -     * and it may have stash some variable on the stack. -     * Jump to a brand new function as the stack was reseted -     */

xen/arch/riscv/mm.c +6 -0
+void __init remove_identity_mapping(void)
+{
+    unsigned int i;
+    pte_t *pgtbl;
+    unsigned int index, xen_index;
+    unsigned long load_addr = LINK_TO_LOAD(_start);
Quoted 6 lines


-    switch_stack_and_jump((unsigned long)cpu0_boot_stack + STACK_SIZE,
-                          cont_after_mmu_is_enabled); +    for ( pgtbl = stage1_pgtbl_root, i = 0;
+          i <= (CONFIG_PAGING_LEVELS - 1);

i < CONFIG_PAGING_LEVELS ? But maybe it would be easier for i to start
at CONFIG_PAGING_LEVELS and be decremented, simplifying ...

Quoted 7 lines
Quoted 5 lines

+          i++ )
+    {
+        index = pt_index(CONFIG_PAGING_LEVELS - 1 - i, load_addr); +        xen_index = pt_index(CONFIG_PAGING_LEVELS - 1 - i, XEN_VIRT_START);

... these two expressions?

It makes sense. I'll update this part of the code.

Quoted 7 lines
Quoted 6 lines

+        if ( index != xen_index )
+        {
+            pgtbl[index].pte = 0;
+            break;
+        }

Quoted 5 lines

Is this enough? When load and link address are pretty close (but not overlapping), can't they share a leaf table, in which case you need to clear more than just a single entry? The present overlap check looks to be 4k-granular, not 2M (in which case this couldn't happen; IOW adjusting the overlap check may also be a way out).

At the start of setup_initial_pagetables() there is a code which checks that load and link address don't overlap:

    if ( (linker_start != load_start) &&
         (linker_start <= load_end) && (load_start <= linker_end) )
    {
        early_printk("(XEN) linker and load address ranges overlap\n");
        die();
    }

So the closest difference between load and link address can be 4kb. Otherwise load and link address ranges are equal ( as we can't map less then 4kb).

Let's take concrete examples:
Load address range is 0x8020_0000 - 0x8020_0FFF
Linker address range is 0x8020_1000 - 0x8020_1FFF MMU mode: Sv39 ( so we have 3 page tables )

So we have:
* L2 index = 2, L1 index = 1, L0 index = 0 for load address
* L2 index = 2, L1 index = 1, L0 index = 1 for linker address
Thereby we have two different L0 tables for load and linker address
ranges.
And it looks like it is pretty safe to remove only one L0 index that
was used for identity mapping.

Quoted 1 line

Is it possible that I missed something?

Looks as if you are thinking of only a Xen which fits in 4k. The code here wants to cope with Xen getting quite a bit bigger.

Jan

author
#18 lore raw
Quoted 13 lines · Jan Beulich
Quoted 12 lines · Oleksii
Quoted 11 lines · Jan Beulich
Quoted 8 lines · Oleksii Kurochko
xen/arch/riscv/include/asm/config.h +2 -0
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
  #ifndef __RISCV_CONFIG_H__
  #define __RISCV_CONFIG_H__

Unrelated change?

Quoted 6 lines
Quoted 2 lines

It  should be part of [PATCH v2 5/6] xen/riscv: introduce identity mapping.

Hmm, here we're discussing "[PATCH v2 4/6] xen/riscv: introduce identity
mapping". I'm confused, I guess.

Sorry for confusion. i meant the patch: [PATCH v2 5/6] xen/riscv: add SPDX tags.

Quoted 44 lines
Quoted 44 lines
Quoted 31 lines
Quoted 31 lines
xen/arch/riscv/mm.c +0 -0
@@ -25,6 +25,12 @@ unsigned long __ro_after_init phys_offset;
  #define LOAD_TO_LINK(addr) ((unsigned long)(addr) -
 phys_offset)
  #define LINK_TO_LOAD(addr) ((unsigned long)(addr) +
 phys_offset)


+/*
+ * Should be removed as soon as enough headers will be merged for
inclusion of
+ * <xen/lib.h>.
+ */
+#define ARRAY_SIZE(arr)                (sizeof(arr) / sizeof((arr)[0]))
+
 /*
  * It is expected that Xen won't be more then 2 MB.
  * The check in xen.lds.S guarantees that.

xen/arch/riscv/mm.c +0 -0
@@ -35,8 +41,10 @@ unsigned long __ro_after_init phys_offset;
   *
   * It might be needed one more page table in case when Xen

load
address
  * isn't 2 MB aligned.
+ *
+ * (CONFIG_PAGING_LEVELS - 1) page tables are needed for identity
mapping.

Quoted 8 lines
Quoted 4 lines

  */

xen/arch/riscv/mm.c +1 -1
-#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) + 1)
+#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) * 2 +

1)

How come the extra page (see the comment sentence in context) isn't
needed for the identity-mapping case?

It is needed to allocate no more than two 'nonroot' page tables (L0 and
L1 in case of Sv39 ) as page 'root' table ( L2 in case of Sv39 ) is always re-used.

Quoted 4 lines
Quoted 2 lines

The same ( only 'nonroot' page tables might be needed to allocate ) works for any MMU mode.

Of course, but if you cross a 2Mb boundary you'll need 2 L0 tables.

Yes, in the case of crossing a 2Mb boundary, it will require 2 L0 tables.

Then, the number of required page tables is needed depending on Xen size and load address alignment. Because for each 2Mb, we need a new L0 table.

Sure, this is not needed now ( as in xen.lds.S, we have a Xen size check ), but if someone increases Xen size binary to 4Mb, then the amount of page tables should be updated too.
Should we take into account that?

Quoted 105 lines
Quoted 105 lines
Quoted 48 lines
Quoted 37 lines
xen/arch/riscv/mm.c +0 -0
@@ -255,25 +262,30 @@ void __init noreturn noinline
 enable_mmu()
      csr_write(CSR_SATP,

               PFN_DOWN((unsigned long)stage1_pgtbl_root) | RV_STAGE1_MODE << SATP_MODE_SHIFT);
+}

-    asm volatile ( ".p2align 2" );
- mmu_is_enabled:
-    /*
-     * Stack should be re-inited as:
-     * 1. Right now an address of the stack is relative to load
time
-     *    addresses what will cause an issue in case of load start
address
-     *    isn't equal to linker start address.
-     * 2. Addresses in stack are all load time relative which can
be an
-     *    issue in case when load start address isn't equal to linker
-     *    start address.
-     *
-     * We can't return to the caller because the stack was reseted
-     * and it may have stash some variable on the stack. -     * Jump to a brand new function as the stack was reseted -     */

xen/arch/riscv/mm.c +6 -0
+void __init remove_identity_mapping(void)
+{
+    unsigned int i;
+    pte_t *pgtbl;
+    unsigned int index, xen_index;
+    unsigned long load_addr = LINK_TO_LOAD(_start);
Quoted 6 lines


-    switch_stack_and_jump((unsigned long)cpu0_boot_stack + STACK_SIZE,
-                          cont_after_mmu_is_enabled); +    for ( pgtbl = stage1_pgtbl_root, i = 0;
+          i <= (CONFIG_PAGING_LEVELS - 1);

i < CONFIG_PAGING_LEVELS ? But maybe it would be easier for i to start
at CONFIG_PAGING_LEVELS and be decremented, simplifying ...

Quoted 8 lines
Quoted 6 lines

+          i++ )
+    {
+        index = pt_index(CONFIG_PAGING_LEVELS - 1 - i, load_addr);
+        xen_index = pt_index(CONFIG_PAGING_LEVELS - 1 - i, XEN_VIRT_START);

... these two expressions?

It makes sense. I'll update this part of the code.

Quoted 7 lines
Quoted 6 lines

+        if ( index != xen_index )
+        {
+            pgtbl[index].pte = 0;
+            break;
+        }

Quoted 8 lines

Is this enough? When load and link address are pretty close (but not
overlapping), can't they share a leaf table, in which case you need
to clear more than just a single entry? The present overlap check looks to be 4k-granular, not 2M (in which case this couldn't happen;
IOW adjusting the overlap check may also be a way out).

At the start of setup_initial_pagetables() there is a code which checks
that load and link address don't overlap:

    if ( (linker_start != load_start) &&
         (linker_start <= load_end) && (load_start <= linker_end) ) {
        early_printk("(XEN) linker and load address ranges overlap\n");
        die();
    }

So the closest difference between load and link address can be 4kb. Otherwise load and link address ranges are equal ( as we can't map less
then 4kb).

Let's take concrete examples:
  Load address range is   0x8020_0000 - 0x8020_0FFF Linker address range is 0x8020_1000 - 0x8020_1FFF MMU mode: Sv39 ( so we have 3 page tables )

  So we have:
    * L2 index = 2, L1 index = 1, L0 index = 0 for load address
    * L2 index = 2, L1 index = 1, L0 index = 1 for linker address Thereby we have two different L0 tables for load and linker address
ranges.
  And it looks like it is pretty safe to remove only one L0 index that
was used for identity mapping.

Quoted 4 lines
Quoted 1 line

Is it possible that I missed something?

Looks as if you are thinking of only a Xen which fits in 4k. The code here wants to cope with Xen getting quite a bit bigger.

Yeah, I missed that when I tried to come up with an example.

So it will probably be more universal if we recursively go through the whole identity mapping and unmap each pte individually.

#19 lore raw
Quoted 54 lines · Oleksii
Quoted 46 lines · Jan Beulich
Quoted 45 lines · Oleksii
Quoted 31 lines · Jan Beulich
Quoted 30 lines · Oleksii Kurochko
xen/arch/riscv/mm.c +0 -0
@@ -25,6 +25,12 @@ unsigned long __ro_after_init phys_offset;
  #define LOAD_TO_LINK(addr) ((unsigned long)(addr) -
 phys_offset)
  #define LINK_TO_LOAD(addr) ((unsigned long)(addr) +
 phys_offset)


+/*
+ * Should be removed as soon as enough headers will be merged for
inclusion of
+ * <xen/lib.h>.
+ */
+#define ARRAY_SIZE(arr)                (sizeof(arr) / sizeof((arr)[0]))
+
 /*
  * It is expected that Xen won't be more then 2 MB.
  * The check in xen.lds.S guarantees that.

xen/arch/riscv/mm.c +0 -0
@@ -35,8 +41,10 @@ unsigned long __ro_after_init phys_offset;
   *
   * It might be needed one more page table in case when Xen

load
address
  * isn't 2 MB aligned.
+ *
+ * (CONFIG_PAGING_LEVELS - 1) page tables are needed for identity
mapping.

Quoted 8 lines
Quoted 4 lines

  */

xen/arch/riscv/mm.c +1 -1
-#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) + 1)
+#define PGTBL_INITIAL_COUNT ((CONFIG_PAGING_LEVELS - 1) * 2 +

1)

How come the extra page (see the comment sentence in context) isn't
needed for the identity-mapping case?

It is needed to allocate no more than two 'nonroot' page tables (L0 and
L1 in case of Sv39 ) as page 'root' table ( L2 in case of Sv39 ) is always re-used.

Quoted 4 lines
Quoted 2 lines

The same ( only 'nonroot' page tables might be needed to allocate ) works for any MMU mode.

Of course, but if you cross a 2Mb boundary you'll need 2 L0 tables.

Yes, in the case of crossing a 2Mb boundary, it will require 2 L0 tables.

Quoted 8 lines

Then, the number of required page tables is needed depending on Xen size and load address alignment. Because for each 2Mb, we need a new L0 table.

Sure, this is not needed now ( as in xen.lds.S, we have a Xen size check ), but if someone increases Xen size binary to 4Mb, then the amount of page tables should be updated too.
Should we take into account that?

Perhaps by way of a BUILD_BUG_ON(), yes. We want to avoid setting up a (runtime) trap for someone to fall into.

Jan

#20 lore raw

PATCH v2 5/6 xen/riscv: add SPDX tags

In reply to Oleksii Kurochko

---
Changes in V2:
- change SPDX tags from GPL-2.0-or-later to GPL-2.0-only.
- add Suggested-by: Andrew Cooper <[email protected]>.

---
 xen/arch/riscv/include/asm/current.h      | 2 ++
 xen/arch/riscv/include/asm/early_printk.h | 2 ++
 xen/arch/riscv/include/asm/mm.h           | 2 ++
 xen/arch/riscv/include/asm/page-bits.h    | 2 ++
 xen/arch/riscv/include/asm/page.h         | 2 ++
 xen/arch/riscv/include/asm/traps.h        | 2 ++
 xen/arch/riscv/include/asm/types.h        | 2 ++
 xen/arch/riscv/mm.c                       | 2 ++
 xen/arch/riscv/setup.c                    | 2 ++
 9 files changed, 18 insertions(+)
xen/arch/riscv/include/asm/current.h +2 -0
index d87e6717e0..d84f15dc50 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __ASM_CURRENT_H
 #define __ASM_CURRENT_H
 
xen/arch/riscv/include/asm/early_printk.h +2 -0
index 05106e160d..85e60df33a 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __EARLY_PRINTK_H__
 #define __EARLY_PRINTK_H__
 
xen/arch/riscv/include/asm/mm.h +2 -0
index 500fdc9c5a..3f694a43ef 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef _ASM_RISCV_MM_H
 #define _ASM_RISCV_MM_H
 
xen/arch/riscv/include/asm/page-bits.h +2 -0
index 4a3e33589a..8f1f474371 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __RISCV_PAGE_BITS_H__
 #define __RISCV_PAGE_BITS_H__
 
xen/arch/riscv/include/asm/page.h +2 -0
index 22b36e03a4..95074e29b3 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef _ASM_RISCV_PAGE_H
 #define _ASM_RISCV_PAGE_H
 
xen/arch/riscv/include/asm/traps.h +2 -0
index f3fb6b25d1..3fef318478 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __ASM_TRAPS_H__
 #define __ASM_TRAPS_H__
 
xen/arch/riscv/include/asm/types.h +2 -0
index 0c0ce78c8f..c2cdfd9caa 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __RISCV_TYPES_H__
 #define __RISCV_TYPES_H__
 
xen/arch/riscv/mm.c +2 -0
index 2693b817c5..663048c783 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #include <xen/cache.h>
 #include <xen/compiler.h>
 #include <xen/init.h>
xen/arch/riscv/setup.c +2 -0
index c4ef0b3165..6593f601c1 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #include <xen/compile.h>
 #include <xen/init.h>
 
Signature · 1 line
2.40.1
Reviewed-by on v2 5/6
#21 lore raw

Re: [PATCH v2 5/6] xen/riscv: add SPDX tags

Quoted 3 lines · Oleksii Kurochko

Alistair

Quoted 110 lines

---
Changes in V2:
- change SPDX tags from GPL-2.0-or-later to GPL-2.0-only.
- add Suggested-by: Andrew Cooper <[email protected]>.

---
 xen/arch/riscv/include/asm/current.h      | 2 ++
 xen/arch/riscv/include/asm/early_printk.h | 2 ++
 xen/arch/riscv/include/asm/mm.h           | 2 ++
 xen/arch/riscv/include/asm/page-bits.h    | 2 ++
 xen/arch/riscv/include/asm/page.h         | 2 ++
 xen/arch/riscv/include/asm/traps.h        | 2 ++
 xen/arch/riscv/include/asm/types.h        | 2 ++
 xen/arch/riscv/mm.c                       | 2 ++
 xen/arch/riscv/setup.c                    | 2 ++
 9 files changed, 18 insertions(+)
xen/arch/riscv/include/asm/current.h +2 -0
index d87e6717e0..d84f15dc50 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __ASM_CURRENT_H
 #define __ASM_CURRENT_H
 
xen/arch/riscv/include/asm/early_printk.h +2 -0
index 05106e160d..85e60df33a 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __EARLY_PRINTK_H__
 #define __EARLY_PRINTK_H__
 
xen/arch/riscv/include/asm/mm.h +2 -0
index 500fdc9c5a..3f694a43ef 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef _ASM_RISCV_MM_H
 #define _ASM_RISCV_MM_H
 
xen/arch/riscv/include/asm/page-bits.h +2 -0
index 4a3e33589a..8f1f474371 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __RISCV_PAGE_BITS_H__
 #define __RISCV_PAGE_BITS_H__
 
xen/arch/riscv/include/asm/page.h +2 -0
index 22b36e03a4..95074e29b3 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef _ASM_RISCV_PAGE_H
 #define _ASM_RISCV_PAGE_H
 
xen/arch/riscv/include/asm/traps.h +2 -0
index f3fb6b25d1..3fef318478 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __ASM_TRAPS_H__
 #define __ASM_TRAPS_H__
 
xen/arch/riscv/include/asm/types.h +2 -0
index 0c0ce78c8f..c2cdfd9caa 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #ifndef __RISCV_TYPES_H__
 #define __RISCV_TYPES_H__
 
xen/arch/riscv/mm.c +2 -0
index 2693b817c5..663048c783 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #include <xen/cache.h>
 #include <xen/compiler.h>
 #include <xen/init.h>
xen/arch/riscv/setup.c +2 -1
index c4ef0b3165..6593f601c1 100644
@@ -1,3 +1,5 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
 #include <xen/compile.h>
 #include <xen/init.h>
 
--

2.40.1

#22 lore raw

PATCH v2 6/6 xen/riscv: move extern of cpu0_boot_stack to header

In reply to Oleksii Kurochko

---
Changes in V2:
- add Suggested-by: Andrew Cooper <[email protected]>.

---
 xen/arch/riscv/include/asm/mm.h | 2 ++
 xen/arch/riscv/mm.c             | 2 --
 2 files changed, 2 insertions(+), 2 deletions(-)
xen/arch/riscv/include/asm/mm.h +2 -0
index 3f694a43ef..085eaab7fb 100644
@@ -8,6 +8,8 @@
 #define pfn_to_paddr(pfn) ((paddr_t)(pfn) << PAGE_SHIFT)
 #define paddr_to_pfn(pa)  ((unsigned long)((pa) >> PAGE_SHIFT))
 
+extern unsigned char cpu0_boot_stack[];
+
 void setup_initial_pagetables(void);
 
 void enable_mmu(void);
xen/arch/riscv/mm.c +0 -2
index 663048c783..602b89aeed 100644
@@ -20,8 +20,6 @@ struct mmu_desc {
     pte_t *pgtbl_base;
 };
 
-extern unsigned char cpu0_boot_stack[STACK_SIZE];
-
 unsigned long __ro_after_init phys_offset;
 
 #define LOAD_TO_LINK(addr) ((unsigned long)(addr) - phys_offset)
Signature · 1 line
2.40.1
Reviewed-by on v2 6/6
#23 lore raw

Re: [PATCH v2 6/6] xen/riscv: move extern of cpu0_boot_stack to header

Quoted 3 lines · Oleksii Kurochko

Alistair

Quoted 38 lines

---
Changes in V2:
- add Suggested-by: Andrew Cooper <[email protected]>.

---
 xen/arch/riscv/include/asm/mm.h | 2 ++
 xen/arch/riscv/mm.c             | 2 --
 2 files changed, 2 insertions(+), 2 deletions(-)
xen/arch/riscv/include/asm/mm.h +2 -0
index 3f694a43ef..085eaab7fb 100644
@@ -8,6 +8,8 @@
 #define pfn_to_paddr(pfn) ((paddr_t)(pfn) << PAGE_SHIFT)
 #define paddr_to_pfn(pa)  ((unsigned long)((pa) >> PAGE_SHIFT))
 
+extern unsigned char cpu0_boot_stack[];
+
 void setup_initial_pagetables(void);
 
 void enable_mmu(void);
xen/arch/riscv/mm.c +0 -3
index 663048c783..602b89aeed 100644
@@ -20,8 +20,6 @@ struct mmu_desc {
     pte_t *pgtbl_base;
 };
 
-extern unsigned char cpu0_boot_stack[STACK_SIZE];
-
 unsigned long __ro_after_init phys_offset;
 
 #define LOAD_TO_LINK(addr) ((unsigned long)(addr) - phys_offset)
--

2.40.1

author
#24 lore raw
In reply to Oleksii Kurochko

Hello Bobby and Alistair,

Some patches of the current patch series were merged, but some still require your Ack-by/Reviewed-by.

Could you please look at the following patches:
1. [PATCH v2 3/6] xen/riscv: introduce function for physical offset calculation
2. [PATCH v2 4/6] xen/riscv: introduce identity mapping

Thank you so much for your attention and participation.

~ Oleksii

Quoted 76 lines · Oleksii Kurochko

The patch series introduces things necessary to implement identity mapping:
  1. Make identity mapping for the entire Xen.
  2. Enable MMU.
  3. Jump to the virtual address world
  4. Remove identity mapping.

Also current patch series introduces the calculation of physical offset before
MMU is enabled as access to physical offset will be calculated wrong after
MMU will be enabled because access to phys_off variable is PC- relative and
in the case when linker address != load address, it will cause MMU fault.

One more thing that was done is:
  * Added SPDX tags.
  * move extern of cpu0_boot_stack to a header.

The reason for this patch series can be found here:
https://lore.kernel.org/xen-devel/[email protected]/

---
Changes in V2:
 - update the patch series message.
 - drop patches from the previous version of the patch series:
   * xen/riscv: add __ASSEMBLY__ guards". ( merged )
   * xen/riscv: make sure that identity mapping isn't bigger then page size
     ( entire Xen is 1:1 mapped so there is no need for the checks from the patch )
 - add .sbss.* and put it befor .bss* .
 - move out reset_stack() to .text section.
 - add '__ro_after_init' for phys_offset variable.
 - add '__init' for calc_phys_offset().
 - declaring variable phys_off as non static as it will be used in head.S.
 - update definition of PGTBL_INITIAL_COUNT and the comment above.
 - code style fixes.
 - remove id_addrs array becase entire Xen is mapped.
 - reverse condition for cycle inside remove_identity_mapping().
 - fix page table walk in remove_identity_mapping().
 - save hart_id and dtb_addr before call MMU related C functions
 - use phys_offset variable instead of doing calcultations to get phys offset
   in head.S file. ( it can be easily done as entire Xen is 1:1 mapped now )
 - declare enable_muu() as __init.
 - Update SPDX tags.
 - Add Review-By/Suggested-By for some patches.
 - code style fixes.

Oleksii Kurochko (6):
  xen/riscv: add .sbss section to .bss
  xen/riscv: introduce reset_stack() function xen/riscv: introduce function for physical offset calculation xen/riscv: introduce identity mapping
  xen/riscv: add SPDX tags
  xen/riscv: move extern of cpu0_boot_stack to header

 xen/arch/riscv/include/asm/config.h       |   2 + xen/arch/riscv/include/asm/current.h      |   2 + xen/arch/riscv/include/asm/early_printk.h |   2 + xen/arch/riscv/include/asm/mm.h           |   9 +- xen/arch/riscv/include/asm/page-bits.h    |   2 + xen/arch/riscv/include/asm/page.h         |   2 + xen/arch/riscv/include/asm/traps.h        |   2 + xen/arch/riscv/include/asm/types.h        |   2 + xen/arch/riscv/mm.c                       | 104 +++++++++++++------- --
 xen/arch/riscv/riscv64/head.S             |  46 +++++++++- xen/arch/riscv/setup.c                    |  16 +--- xen/arch/riscv/xen.lds.S                  |   2 +-

 12 files changed, 136 insertions(+), 55 deletions(-)