xen/riscv: introduce identity mapping
v6 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.
The reason for this patch series can be found here:
https://lore.kernel.org/xen-devel/[email protected]/
---
Changes in V6:
- Update calc_phys_offset() after rebase.
- Refactor turn_on_mmu() and a way how an argument of turn_on_mmu() is
calculated.
---
Changes in V5:
- update the algo of identity mapping removing.
- introduce IDENT_AREA_SIZE.
- introduce turn_on_mmu() function to enable and switch from 1:1 mapping.
- fix typo in PGTBL_INITIAL_COUNT define.
- update the comment above PGTBL_INITIAL_COUNT.
- update prototype of calc_phys_offset(). now it returns phys_offset.
- declare phys_offset as static.
- save returned value of calc_phys_offset to register s2.
---
Changes in V4:
- drop patch [PATCH v3 1/3] xen/riscv: add SPDX tag to config.h as it was
merged to staging
- remove definition of ARRAY_SIZE and ROUNDUP as <xen/macors.h> was introduced where these macros are located now.
- update definition of PGTBL_INITIAL_COUNT
- update the commit message for patch 'xen/riscv: introduce identity mapping'
- update the comments in head.S
- update the algo of identity mapping removing
---
Changes in V3:
- Update the patch series message.
- The following patches were merged to staging so droped from the patch series:
* xen/riscv: add .sbss section to .bss
* xen/riscv: introduce reset_stack() function
* xen/riscv: move extern of cpu0_boot_stack to header
* xen/riscv: add SPDX tags
- move save/restore of a0/a1 registers from patch 4 to patch 2 ( numbers are
from the previous patch series version )
- add SPDX tag in config.h
- update definition of PGTBL_INITIAL_COUNT taking into account identity mapping.
- refactor remove_identity_mapping() function.
- add explanatory comments in xen.lds.S and mm.c.
---
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 (2):
xen/riscv: introduce function for physical offset calculation xen/riscv: introduce identity mapping
xen/arch/riscv/include/asm/config.h | 2 + xen/arch/riscv/include/asm/mm.h | 7 +- xen/arch/riscv/mm.c | 109 +++++++++++++++++----------- xen/arch/riscv/riscv64/head.S | 44 +++++++++++ xen/arch/riscv/setup.c | 14 +--- xen/arch/riscv/xen.lds.S | 11 +++ 6 files changed, 130 insertions(+), 57 deletions(-)
Signature · 1 line
2.41.0
PATCH v6 1/2 xen/riscv: introduce function for physical offset calculation
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 V6:
- After 9380f06fe8("xen: Drop the (almost) unused extern start[]") an update of calc_phys_offset() was needed:
- change start to _start.
- add local volatile variable load_start to get properly Xens' load address. Otherwise compile thinks that (start - XEN_VIRT_START) is equal to 0.
---
Changes in V5:
- update prototype of calc_phys_offset(). now it returns phys_offset.
- declare phys_offset as static.
- save returned value of calc_phys_offset to register s2.
---
Changes in V4:
- update the comment messages in head.S related to save/restore of a0/a1 regs.
---
Changes in V3:
- save/restore of a0/a1 registers before C first function call.
---
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 | 21 ++++++++++++++++++--- xen/arch/riscv/riscv64/head.S | 14 ++++++++++++++ 3 files changed, 34 insertions(+), 3 deletions(-)
xen/arch/riscv/include/asm/mm.h +2 -0
@@ -15,4 +15,6 @@ void setup_initial_pagetables(void);void enable_mmu (void );void cont_after_mmu_is_enabled (void ); +unsigned long calc_phys_offset (void ); +#endif /* _ASM_RISCV_MM_H */
xen/arch/riscv/mm.c +18 -3
@@ -1,5 +1,6 @@/* SPDX-License-Identifier: GPL-2.0-only */ +#include <xen/cache.h> #include <xen/compiler.h> #include <xen/init.h> #include <xen/kernel.h>
@@ -19,9 +20,10 @@ struct mmu_desc {pte_t *pgtbl_base; }; -#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) +static 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 +275,16 @@ 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 +*/ +unsigned long __init calc_phys_offset (void ) +{ +volatile unsigned long load_start = (unsigned long )_start; + + phys_offset = load_start -XEN_VIRT_START ; +return phys_offset; +}
xen/arch/riscv/riscv64/head.S +14 -0
@@ -29,6 +29,20 @@ ENTRY(start) jal reset_stack + /* + * save hart_id ( bootcpu_id ) and dtb_base as a0 and a1 register can + * be used by C code + */ + mv s0, a0 + mv s1, a1 + + jal calc_phys_offset + mv s2, a0 + + /* restore hart_id ( bootcpu_id ) and dtb address */ + mv a0, s0 + mv a1, s1 + tail start_xen .section .text, "ax", %progbits
Signature · 1 line
2.41.0
PATCH v6 2/2 xen/riscv: introduce identity mapping
In reply to Oleksii KurochkoThe 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 for area which constains needed code to switch from identity mapping and after switching to virtual addresses, identity mapping is removed from page-tables in the following way:
search for top-most page table entry and remove it.
---
Changes in V6:
- t2 register was renamed to t1 as t1 isn't used anymore in
turn_on_mmu().
- fold li, sll instructions to slli.
- use a0 directly to calculate an argument of turn_on_mmu() instead of t0.
---
Changes in V5:
- update the algo of identity mapping removing.
- introduce IDENT_AREA_SIZE.
- introduce turn_on_mmu() function to enable and switch from 1:1 mapping.
- fix typo in PGTBL_INITIAL_COUNT define.
- update the comment above PGTBL_INITIAL_COUNT.
- update the commit message.
---
Changes in V4:
- remove definition of ARRAY_SIZE and ROUNDUP as <xen/macors.h> was introduced where these macros are located now.
- update definition of PGTBL_INITIAL_COUNT
- update the commit message
- update the algo of identity mapping removing
---
Changes in V3:
- remove unrelated to the patch changes ( SPDX tags in config.h ).
- update definition of PGTBL_INITIAL_COUNT taking into account identity mapping.
- refactor remove_identity_mapping() function.
- add explanatory comments in xen.lds.S and mm.c.
- update commit message.
- move save/restore of a0/a1 registers to [PATCH v2 2/3] xen/riscv: introduce
function for physical offset calculation.
---
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 | 5 +- xen/arch/riscv/mm.c | 90 +++++++++++++++-------------- xen/arch/riscv/riscv64/head.S | 30 ++++++++++ xen/arch/riscv/setup.c | 14 +---- xen/arch/riscv/xen.lds.S | 11 ++++ 6 files changed, 97 insertions(+), 55 deletions(-)
xen/arch/riscv/include/asm/config.h +2 -0
@@ -95,6 +95,8 @@#define RV_STAGE1_MODE SATP_MODE_SV32#endif +#define IDENT_AREA_SIZE 64 +#endif /* __RISCV_CONFIG_H__ */ /* * Local variables:
xen/arch/riscv/include/asm/mm.h +4 -1
@@ -13,8 +13,11 @@ extern unsigned char cpu0_boot_stack[];void setup_initial_pagetables (void );void enable_mmu (void ); -void cont_after_mmu_is_enabled (void ); + +void remove_identity_mapping (void );unsigned long calc_phys_offset (void ); +void turn_on_mmu (unsigned long ra); +#endif /* _ASM_RISCV_MM_H */
xen/arch/riscv/mm.c +48 -42
138 lines not on this page · raw message
xen/arch/riscv/riscv64/head.S +30 -0
@@ -39,6 +39,22 @@ ENTRY(start) jal calc_phys_offset mv s2, a0 + jal setup_initial_pagetables + + /* Calculate proper VA after jump from 1:1 mapping */ + la a0, .L_primary_switched + sub a0, a0, s2 + + jal turn_on_mmu + +.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 hart_id ( bootcpu_id ) and dtb address */ mv a0, s0 mv a1, s1
@@ -54,3 +70,17 @@ ENTRY(reset_stack) ret + .section .text.ident, "ax", %progbits + +ENTRY(turn_on_mmu) + sfence.vma + + li t0, RV_STAGE1_MODE + slli t0, t0, SATP_MODE_SHIFT + + la t1, stage1_pgtbl_root + srl t1, t1, PAGE_SHIFT + or t1, t1, t0 + csrw CSR_SATP, t1 + + jr a0
xen/arch/riscv/setup.c +2 -12
@@ -13,20 +13,10 @@ unsigned char __initdata cpu0_boot_stack[STACK_SIZE]void __init noreturnstart_xen (unsigned long bootcpu_id,paddr_t dtb_addr) { -early_printk ("Hello from C env\n" ); - -setup_initial_pagetables (); - -enable_mmu (); - -for ( ;; ) - asmvolatile ("wfi" ); +remove_identity_mapping (); -unreachable (); -} +early_printk ("Hello from C env\n" ); -void __init noreturncont_after_mmu_is_enabled (void ) -{early_printk ("All set up\n" );for ( ;; )
xen/arch/riscv/xen.lds.S +11 -0
@@ -31,6 +31,11 @@ SECTIONS *(.text.*) #endif + . = ALIGN(IDENT_AREA_SIZE); + _ident_start = .; + *(.text.ident) + _ident_end = .; + *(.fixup) *(.gnu.warning) . = ALIGN(POINTER_ALIGN);
@@ -173,4 +178,10 @@ ASSERT(IS_ALIGNED(__bss_end, POINTER_ALIGN), "__bss_end is misaligned") ASSERT(!SIZEOF(.got), ".got non-empty") ASSERT(!SIZEOF(.got.plt), ".got.plt non-empty") +/* + * Changing the size of Xen binary can require an update of + * PGTBL_INITIAL_COUNT. + */ ASSERT(_end - _start <= MB(2), "Xen too large for early-boot assumptions") + +ASSERT(_ident_end - _ident_start <= IDENT_AREA_SIZE, "identity region is too big");
Signature · 1 line
2.41.0
Quoted 4 lines · Oleksii Kurochko
xen/arch/riscv/riscv64/head.S +0
-0@@ -54,3 +70,17 @@ ENTRY(reset_stack) ret
xen/arch/riscv/riscv64/head.S ENTRY(reset_stack) · line 82
Quoted 10 lines
xen/arch/riscv/riscv64/head.S +10
-0+ .section .text.ident, "ax", %progbits + +ENTRY(turn_on_mmu) + sfence.vma + + li t0, RV_STAGE1_MODE + slli t0, t0, SATP_MODE_SHIFT + + la t1, stage1_pgtbl_root + srl t1, t1, PAGE_SHIFT
I think it would be good to be consistent in the use of pseudo insns:
Above you use slli, so here it would want to be srli (or the other way around).
Jan
Quoted 4 lines
xen/arch/riscv/riscv64/head.S +4
-0+ or t1, t1, t0 + csrw CSR_SATP, t1 + + jr a0
Quoted 11 lines · Jan Beulich
Quoted 10 lines · Oleksii Kurochko
(quoted diff) +0
-0@@ -54,3 +70,17 @@ ENTRY(reset_stack)
ret
+ .section .text.ident, "ax", %progbitsxen/arch/riscv/riscv64/head.S +5
-0+ +ENTRY(turn_on_mmu) + sfence.vma + + li t0, RV_STAGE1_MODE
Quoted 8 lines
Quoted 4 lines
+ slli t0, t0,
SATP_MODE_SHIFT
+
+ la t1,stage1_pgtbl_root
+ srl t1, t1,PAGE_SHIFTI think it would be good to be consistent in the use of pseudo insns:
Above you use slli, so here it would want to be srli (or the other way around).
Oops, I overlooked that.
When I examined the disassembler, it automatically transformed it to 'srli',
so I forgot to change it.
It would be more appropriate to use 'srli'.
I'll wait for any additional comments, and if there are none, I'll send a new patch series version.
Thanks.
~ Oleksii
Quoted 6 lines
Quoted 5 lines
+ or t1, t1, t0
+ csrwCSR_SATP, t1
+
+ jr a0