linux-kernel - Re: [PATCH v2] x86: Stop relying on magic jmp behavior for early_idt

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 21 May 2015 23:51:50 -0700
From:	Andy Lutomirski <luto@...capital.net>
To:	Ingo Molnar <mingo@...nel.org>
Cc:	Andy Lutomirski <luto@...nel.org>, X86 ML <x86@...nel.org>,
	"H. Peter Anvin" <hpa@...or.com>, "H.J. Lu" <hjl.tools@...il.com>,
	Borislav Petkov <bp@...en8.de>,
	Jan Beulich <JBeulich@...e.com>,
	Binutils <binutils@...rceware.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] x86: Stop relying on magic jmp behavior for early_idt_handlers

On Thu, May 21, 2015 at 11:22 PM, Ingo Molnar <mingo@...nel.org> wrote:
>
> * Andy Lutomirski <luto@...nel.org> wrote:
>
>> --- a/arch/x86/include/asm/segment.h
>> +++ b/arch/x86/include/asm/segment.h
>> @@ -231,9 +231,17 @@
>>  #define TLS_SIZE                     (GDT_ENTRY_TLS_ENTRIES* 8)
>>
>>  #ifdef __KERNEL__
>> +
>> +/*
>> + * early_idt_handlers is an array of entry points.  For simplicity, it's
>> + * a real array.  We allocate nine bytes for each entry: two one-byte
>> + * push instructions and a five-byte jump in the worst case.
>> + */
>> +#define EARLY_IDT_HANDLER_STRIDE 9
>
> So how come that two plus five equals nine? ;-)

A "one-byte push" would be pretty impressive :)

>
> It's two two-bype pushes (sometimes a 2-byte NOP), plus a 5-byte jump,
> in the worst case.
>
> I'd also mention that it's an array of 32 small trampolines that set
> up parameters and jump to a common entry point.

Will do.

>
>> +/* Build the early_idt_handlers array */
>>  ENTRY(early_idt_handlers)
>
> Please rename the function accordingly: early_idt_handlers_array, to
> make clear this is never jumped to directly, only through the IDT.

Makes sense.

>
>>       # 36(%esp) %eflags
>>       # 32(%esp) %cs
>> @@ -531,19 +532,18 @@ ENTRY(early_idt_handlers)
>>       # 24(%rsp) error code
>>       i = 0
>>       .rept NUM_EXCEPTION_VECTORS
>> -     .if (EXCEPTION_ERRCODE_MASK >> i) & 1
>> -     ASM_NOP2
>> -     .else
>> +     .fill early_idt_handlers + i * EARLY_IDT_HANDLER_STRIDE - ., 1, 0xcc
>> +     .ifeq (EXCEPTION_ERRCODE_MASK >> i) & 1
>>       pushl $0                # Dummy error code, to make stack frame uniform
>>       .endif
>>       pushl $i                # 20(%esp) Vector number
>>       jmp early_idt_handler
>>       i = i + 1
>>       .endr
>> +     .fill early_idt_handlers + i * EARLY_IDT_HANDLER_STRIDE - ., 1, 0xcc
>>  ENDPROC(early_idt_handlers)
>>
>> -     /* This is global to keep gas from relaxing the jumps */
>> -ENTRY(early_idt_handler)
>> +early_idt_handler:
>
> Please rename this as well to make it clearer what's happening,
> something like:
>
>    early_idt_handler_common:
>
> would work for me.
>
> Ditto for the 64-bit side.

Sounds good.  I'll do all this tomorrow and resend.

--Andy
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/