Re: [PATCH] arm: fix page faults in do_alignment
From: Eric W. Biederman <hidden>
Date: 2019-08-30 19:46:30
Also in:
linux-mm, lkml
Subsystem:
arm port, the rest · Maintainers:
Russell King, Linus Torvalds
Possibly related (same subject, not in this thread)
- 2019-09-16 · Re: [PATCH] arm: fix page faults in do_alignment · Eric W. Biederman <hidden>
- 2019-09-15 · Re: [PATCH] arm: fix page faults in do_alignment · Russell King - ARM Linux admin <linux@armlinux.org.uk>
- 2019-09-06 · Re: [PATCH] arm: fix page faults in do_alignment · Russell King - ARM Linux admin <linux@armlinux.org.uk>
- 2019-09-04 · Re: [PATCH] arm: fix page faults in do_alignment · Jing Xiangfeng <hidden>
- 2019-09-02 · Re: [PATCH] arm: fix page faults in do_alignment · Eric W. Biederman <hidden>
Russell King - ARM Linux admin [off-list ref] writes:
On Fri, Aug 30, 2019 at 09:31:17PM +0800, Jing Xiangfeng wrote:quoted
The function do_alignment can handle misaligned address for user and kernel space. If it is a userspace access, do_alignment may fail on a low-memory situation, because page faults are disabled in probe_kernel_address. Fix this by using __copy_from_user stead of probe_kernel_address. Fixes: b255188 ("ARM: fix scheduling while atomic warning in alignment handling code") Signed-off-by: Jing Xiangfeng <redacted>NAK. The "scheduling while atomic warning in alignment handling code" is caused by fixing up the page fault while trying to handle the mis-alignment fault generated from an instruction in atomic context. Your patch re-introduces that bug.
And the patch that fixed scheduling while atomic apparently introduced a regression. Admittedly a regression that took 6 years to track down but still. So it looks like the code needs to do something like:
diff --git a/arch/arm/mm/alignment.c b/arch/arm/mm/alignment.c
index 04b36436cbc0..5e2b8623851e 100644
--- a/arch/arm/mm/alignment.c
+++ b/arch/arm/mm/alignment.c@@ -784,6 +784,9 @@ do_alignment(unsigned long addr, unsigned int fsr, struct pt_regs *regs) instrptr = instruction_pointer(regs); + if (user_mode(regs)) + goto user; + if (thumb_mode(regs)) { u16 *ptr = (u16 *)(instrptr & ~1); fault = probe_kernel_address(ptr, tinstr);
@@ -933,6 +936,34 @@ do_alignment(unsigned long addr, unsigned int fsr, struct pt_regs *regs) return 1; user: + if (thumb_mode(regs)) { + u16 *ptr = (u16 *)(instrptr & ~1); + fault = get_user(tinstr, ptr); + tinstr = __mem_to_opcode_thumb16(tinstr); + if (!fault) { + if (cpu_architecture() >= CPU_ARCH_ARMv7 && + IS_T32(tinstr)) { + /* Thumb-2 32-bit */ + u16 tinst2 = 0; + fault = get_user(ptr + 1, tinst2); + tinst2 = __mem_to_opcode_thumb16(tinst2); + instr = __opcode_thumb32_compose(tinstr, tinst2); + thumb2_32b = 1; + } else { + isize = 2; + instr = thumb2arm(tinstr); + } + } + } else { + fault = get_user(instr, (u32*)instrptr); + instr = __mem_to_opcode_arm(instr); + } + + if (fault) { + type = TYPE_FAULT; + goto bad_or_fault; + } + ai_user += 1; if (ai_usermode & UM_WARN)
Eric _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel