Skip to contentSkip to conversation

arm64: map FDT as RW for early_init_dt_scan()

PATCH In review Michael Bestas 2 participants 5 messages Aug 9, 2022 – Aug 19, 2022

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 arm64: map FDT as RW for early_init_dt_scan()

From: Hsin-Yi Wang <[email protected]>

commit e112b032a72c78f15d0c803c5dc6be444c2e6c66 upstream.

Currently in arm64, FDT is mapped to RO before it's passed to early_init_dt_scan(). However, there might be some codes (eg. commit "fdt: add support for rng-seed") that need to modify FDT during init. Map FDT to RO after early fixups are done.

[mkbestas: fixed trivial conflicts for 4.9 backport]
---
 arch/arm64/include/asm/mmu.h |  2 +-
 arch/arm64/kernel/kaslr.c    |  5 +----
 arch/arm64/kernel/setup.c    |  9 ++++++++-
 arch/arm64/mm/mmu.c          | 15 +--------------
 4 files changed, 11 insertions(+), 20 deletions(-)
arch/arm64/include/asm/mmu.h +1 -1
index f4377b005cba90..c944253b3a4b63 100644
@@ -90,7 +90,7 @@ extern void init_mem_pgprot(void);
 extern void create_pgd_mapping(struct mm_struct *mm, phys_addr_t phys,
 			       unsigned long virt, phys_addr_t size,
 			       pgprot_t prot, bool allow_block_mappings);
-extern void *fixmap_remap_fdt(phys_addr_t dt_phys);
+extern void *fixmap_remap_fdt(phys_addr_t dt_phys, int *size, pgprot_t prot);
 
 #endif	/* !__ASSEMBLY__ */
 #endif
arch/arm64/kernel/kaslr.c +1 -4
index c9ca903462a68d..6a9668f6e933f3 100644
@@ -65,9 +65,6 @@ static __init const u8 *kaslr_get_cmdline(void *fdt)
 	return default_cmdline;
 }
 
-extern void *__init __fixmap_remap_fdt(phys_addr_t dt_phys, int *size,
-				       pgprot_t prot);
-
 /*
  * This routine will be executed with the kernel mapped at its default virtual
  * address, and if it returns successfully, the kernel will be remapped, and
@@ -96,7 +93,7 @@ u64 __init kaslr_early_init(u64 dt_phys, u64 modulo_offset)
 	 * attempt at mapping the FDT in setup_machine()
 	 */
 	early_fixmap_init();
-	fdt = __fixmap_remap_fdt(dt_phys, &size, PAGE_KERNEL);
+	fdt = fixmap_remap_fdt(dt_phys, &size, PAGE_KERNEL);
 	if (!fdt)
 		return 0;
 
arch/arm64/kernel/setup.c +8 -1
index f534f492a26874..ae82d9694542cd 100644
@@ -178,7 +178,11 @@ static void __init smp_build_mpidr_hash(void)
 
 static void __init setup_machine_fdt(phys_addr_t dt_phys)
 {
-	void *dt_virt = fixmap_remap_fdt(dt_phys);
+	int size;
+	void *dt_virt = fixmap_remap_fdt(dt_phys, &size, PAGE_KERNEL);
+
+	if (dt_virt)
+		memblock_reserve(dt_phys, size);
 
 	if (!dt_virt || !early_init_dt_scan(dt_virt)) {
 		pr_crit("\n"
@@ -191,6 +195,9 @@ static void __init setup_machine_fdt(phys_addr_t dt_phys)
 			cpu_relax();
 	}
 
+	/* Early fixups are done, map the FDT as read-only now */
+	fixmap_remap_fdt(dt_phys, &size, PAGE_KERNEL_RO);
+
 	dump_stack_set_arch_desc("%s (DT)", of_flat_dt_get_machine_name());
 }
 
arch/arm64/mm/mmu.c +1 -14
index 36bd50091c4bbc..784ea7c8d99667 100644
@@ -718,7 +718,7 @@ void __set_fixmap(enum fixed_addresses idx,
 	}
 }
 
-void *__init __fixmap_remap_fdt(phys_addr_t dt_phys, int *size, pgprot_t prot)
+void *__init fixmap_remap_fdt(phys_addr_t dt_phys, int *size, pgprot_t prot)
 {
 	const u64 dt_virt_base = __fix_to_virt(FIX_FDT);
 	int offset;
@@ -771,19 +771,6 @@ void *__init __fixmap_remap_fdt(phys_addr_t dt_phys, int *size, pgprot_t prot)
 	return dt_virt;
 }
 
-void *__init fixmap_remap_fdt(phys_addr_t dt_phys)
-{
-	void *dt_virt;
-	int size;
-
-	dt_virt = __fixmap_remap_fdt(dt_phys, &size, PAGE_KERNEL_RO);
-	if (!dt_virt)
-		return NULL;
-
-	memblock_reserve(dt_phys, size);
-	return dt_virt;
-}
-
 int __init arch_ioremap_pud_supported(void)
 {
 	/*
Signature · 1 line
2.37.1
maintainer
#1 lore raw
Quoted 16 lines · Michael Bestas

From: Hsin-Yi Wang <[email protected]>

commit e112b032a72c78f15d0c803c5dc6be444c2e6c66 upstream.

Currently in arm64, FDT is mapped to RO before it's passed to early_init_dt_scan(). However, there might be some codes (eg. commit "fdt: add support for rng-seed") that need to modify FDT during init. Map FDT to RO after early fixups are done.

[mkbestas: fixed trivial conflicts for 4.9 backport]
Quoted 6 lines
---
 arch/arm64/include/asm/mmu.h |  2 +-
 arch/arm64/kernel/kaslr.c    |  5 +----
 arch/arm64/kernel/setup.c    |  9 ++++++++-
 arch/arm64/mm/mmu.c          | 15 +--------------
 4 files changed, 11 insertions(+), 20 deletions(-)

What about 4.14.y and newer?

thanks,

greg k-h

#2 lore raw
Quoted 5 lines · On Tue, 9 Aug 2022 19:12:54 +0200 Greg KH <[email protected]> wrote:

What about 4.14.y and newer?

thanks,

greg k-h

This patch should be required on all stable kernels that got commit "fdt: add support for rng-seed", however I have not tested it.

A similar backport exists in android 4.19 kernel:
https://android-review.googlesource.com/c/kernel/common/+/1238592

Without this patch, Google Pixel 3/3a fails at a very early boot stage after merging v4.9.320+ due to the random backport.

Thanks,

Michael Bestas

#3 lore raw
In reply to Greg KH
Quoted 5 lines · On Tue, 9 Aug 2022 19:12:54 +0200 Greg KH <[email protected]> wrote:

What about 4.14.y and newer?

thanks,

greg k-h

This patch should be required on all stable kernels that got commit "fdt: add support for rng-seed", however I have not tested it.

A similar backport exists in android 4.19 kernel:
https://android-review.googlesource.com/c/kernel/common/+/1238592

Without this patch, Google Pixel 3/3a fails at a very early boot stage after merging v4.9.320+ due to the random backport.

Sorry if I messed something up with the emails, this is the first time I submit something to a mailing list.

Thanks,

Michael Bestas

maintainer
#4 lore raw
Quoted 7 lines · Michael Bestas
Quoted 4 lines · On Tue, 9 Aug 2022 19:12:54 +0200 Greg KH <[email protected]> wrote:

What about 4.14.y and newer?

thanks,

Quoted 7 lines
Quoted 1 line

greg k-h

This patch should be required on all stable kernels that got commit "fdt: add support for rng-seed", however I have not tested it.

A similar backport exists in android 4.19 kernel:
https://android-review.googlesource.com/c/kernel/common/+/1238592

Great, please submit it so that we can include it.

Also for 4.14.y. I can't take a patch for an older kernel tree without the same commit being in a newer one, otherwise people would have regressions when upgrading.

thanks,

greg k-h