Re: [PATCH v2] arm64: mm: fix pass user prot to ioremap_prot in generic_access_phys
From: Jinjiang Tu <hidden>
Date: 2026-01-28 07:13:09
Also in:
linux-mm
在 2026/1/27 17:01, Jinjiang Tu 写道:
quoted hunk ↗ jump to hunk
Here is a syzkaller error log: [0000000020ffc000] pgd=080000010598d403, p4d=080000010598d403, pud=0800000125ddb403, pmd=080000007833c403, pte=01608000007fcfcf Unable to handle kernel read from unreadable memory at virtual address ffff80008ea89000 KASAN: probably user-memory-access in range [0x0000000475448000-0x0000000475448007] Mem abort info: ESR = 0x000000009600000f EC = 0x25: DABT (current EL), IL = 32 bits SET = 0, FnV = 0 EA = 0, S1PTW = 0 FSC = 0x0f: level 3 permission fault Data abort info: ISV = 0, ISS = 0x0000000f, ISS2 = 0x00000000 CM = 0, WnR = 0, TnD = 0, TagAccess = 0 GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 swapper pgtable: 4k pages, 48-bit VAs, pgdp=00000001244aa000 [ffff80008ea89000] pgd=100000013ffff403, p4d=100000013ffff403, pud=100000013fffe403, pmd=100000010a453403, pte=01608000007fcfcf Internal error: Oops: 000000009600000f [#1] SMP Modules linked in: team CPU: 1 PID: 10840 Comm: syz.9.83 Kdump: loaded Tainted: G Hardware name: QEMU KVM Virtual Machine, BIOS 0.0.0 02/06/2015 pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) pc : __memcpy_fromio+0x80/0xf8 lr : generic_access_phys+0x20c/0x2b8 sp : ffff8000a0507960 x29: ffff8000a0507960 x28: 1ffff000140a0f44 x27: ffff00003833cfe0 x26: 0000000000000000 x25: 0000000000001000 x24: 0010000000000001 x23: ffff80008ea89000 x22: ffff00004ea63000 x21: 0000000000001000 x20: ffff80008ea89000 x19: ffff00004ea62000 x18: 0000000000000000 x17: 0000000000000000 x16: 0000000000000000 x15: ffff8000806f1e3c x14: ffff8000806f1d44 x13: 0000000041b58ab3 x12: ffff7000140a0f23 x11: 1ffff000140a0f22 x10: ffff7000140a0f22 x9 : ffff800080579d24 x8 : 0000000000000004 x7 : 0000000000000003 x6 : 0000000000000001 x5 : ffff8000a0507910 x4 : ffff7000140a0f22 x3 : dfff800000000000 x2 : 0000000000001000 x1 : ffff80008ea89000 x0 : ffff00004ea62000 Call trace: __memcpy_fromio+0x80/0xf8 generic_access_phys+0x20c/0x2b8 __access_remote_vm+0x46c/0x5b8 access_remote_vm+0x18/0x30 environ_read+0x238/0x3e8 vfs_read+0xe4/0x2b0 ksys_read+0xcc/0x178 __arm64_sys_read+0x4c/0x68 invoke_syscall+0x68/0x1a0 el0_svc_common.constprop.0+0x11c/0x150 do_el0_svc+0x38/0x50 el0_svc+0x50/0x258 el0t_64_sync_handler+0xc0/0xc8 el0t_64_sync+0x1a4/0x1a8 Code: 91002339 aa1403f7 8b190276 d503201f (f94002f8) The local syzkaller first maps I/O address from /dev/mem to userspace, overiding the stack vma with MAP_FIXED flag. As a result, when reading /proc/$pid/environ, generic_access_phys() is called to access the region, which triggers a PAN permission-check fault and causes a kernel access fault. The root cause is that generic_access_phys() passes a user pte to ioremap_prot(), the user pte sets PTE_USER and PTE_NG bits. Consequently, any subsequent kernel-mode access to the remapped address raises a fault. To fix it, define arch_mk_kernel_prot() to convert user prot to kernel prot for arm64, and call arch_mk_kernel_prot() in generic_access_phys(), so that a user prot is passed to ioremap_prot(). Fixes: 893dea9ccd08 ("arm64: Add HAVE_IOREMAP_PROT support") Signed-off-by: Zeng Heng <redacted> Signed-off-by: Jinjiang Tu <redacted> --- Change since v1: * fix in ther caller generic_access_phys(), suggested by Catalin Marinas. arch/arm64/include/asm/io.h | 11 +++++++++++ mm/memory.c | 12 +++++++++++- 2 files changed, 22 insertions(+), 1 deletion(-)diff --git a/arch/arm64/include/asm/io.h b/arch/arm64/include/asm/io.h index 83e03abbb2ca..fe8607eafab6 100644 --- a/arch/arm64/include/asm/io.h +++ b/arch/arm64/include/asm/io.h@@ -267,6 +267,17 @@ int arm64_ioremap_prot_hook_register(const ioremap_prot_hook_t hook); #define ioremap_prot ioremap_prot +#define arch_mk_kernel_prot arch_mk_kernel_prot +static inline pgprot_t arch_mk_kernel_prot(pgprot_t user_prot) +{ + unsigned long kernel_prot_val; + + kernel_prot_val = _PAGE_KERNEL & ~(PTE_WRITE | PTE_ATTRINDX_MASK); + kernel_prot_val |= pgprot_val(user_prot) & (PTE_WRITE | PTE_ATTRINDX_MASK); + + return __pgprot(kernel_prot_val); +}
I found I misunderstand the READ/WRITE permisson here. If the pte is writeale, the PTE_WRITE is 1, and PTE_RDONLY means dirty or not (With DBM). If the pte is read-only, the PTE_WRITE is 0, and PTE_RDONLY is 1. Since generic_access_phys() have checked if the user can write with: if ((write & FOLL_WRITE) && !pte_write(pte)) return -EINVAL; So, maybe we could always grant writable permission, just as x86's ioremap() does? Thanks.
quoted hunk ↗ jump to hunk
+ #define _PAGE_IOREMAP PROT_DEVICE_nGnRE #define ioremap_wc(addr, size) \diff --git a/mm/memory.c b/mm/memory.c index da360a6eb8a4..68fcfc3bbad8 100644 --- a/mm/memory.c +++ b/mm/memory.c@@ -6836,6 +6836,14 @@ void follow_pfnmap_end(struct follow_pfnmap_args *args) EXPORT_SYMBOL_GPL(follow_pfnmap_end); #ifdef CONFIG_HAVE_IOREMAP_PROT +#ifndef arch_mk_kernel_prot +#define arch_mk_kernel_prot arch_mk_kernel_prot +static inline pgprot_t arch_mk_kernel_prot(pgprot_t user_prot) +{ + return user_prot; +} +#endif + /** * generic_access_phys - generic implementation for iomem mmap access * @vma: the vma to access@@ -6853,6 +6861,7 @@ int generic_access_phys(struct vm_area_struct *vma, unsigned long addr, { resource_size_t phys_addr; pgprot_t prot = __pgprot(0); + pgprot_t kernel_prot; void __iomem *maddr; int offset = offset_in_page(addr); int ret = -EINVAL;@@ -6870,7 +6879,8 @@ int generic_access_phys(struct vm_area_struct *vma, unsigned long addr, if ((write & FOLL_WRITE) && !writable) return -EINVAL; - maddr = ioremap_prot(phys_addr, PAGE_ALIGN(len + offset), prot); + kernel_prot = arch_mk_kernel_prot(prot); + maddr = ioremap_prot(phys_addr, PAGE_ALIGN(len + offset), kernel_prot); if (!maddr) return -ENOMEM;