linux-kernel - Re: [RFC][PATCH 1/2] x86: Allow breakpoints to emulate call functions

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]

Message-Id: <2045370D-38D8-406C-9E94-C1D483E232C9@amacapital.net>
Date:   Fri, 3 May 2019 09:20:55 -0700
From:   Andy Lutomirski <luto@...capital.net>
To:     Steven Rostedt <rostedt@...dmis.org>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Linux List Kernel Mailing <linux-kernel@...r.kernel.org>,
        Ingo Molnar <mingo@...nel.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Andy Lutomirski <luto@...nel.org>,
        Nicolai Stange <nstange@...e.de>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        "H. Peter Anvin" <hpa@...or.com>,
        the arch/x86 maintainers <x86@...nel.org>,
        Josh Poimboeuf <jpoimboe@...hat.com>,
        Jiri Kosina <jikos@...nel.org>,
        Miroslav Benes <mbenes@...e.cz>,
        Petr Mladek <pmladek@...e.com>,
        Joe Lawrence <joe.lawrence@...hat.com>,
        Shuah Khan <shuah@...nel.org>,
        Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>,
        Tim Chen <tim.c.chen@...ux.intel.com>,
        Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
        Mimi Zohar <zohar@...ux.ibm.com>,
        Juergen Gross <jgross@...e.com>,
        Nick Desaulniers <ndesaulniers@...gle.com>,
        Nayna Jain <nayna@...ux.ibm.com>,
        Masahiro Yamada <yamada.masahiro@...ionext.com>,
        Joerg Roedel <jroedel@...e.de>,
        "open list:KERNEL SELFTEST FRAMEWORK" 
        <linux-kselftest@...r.kernel.org>, stable <stable@...r.kernel.org>
Subject: Re: [RFC][PATCH 1/2] x86: Allow breakpoints to emulate call functions



> On May 3, 2019, at 6:22 AM, Steven Rostedt <rostedt@...dmis.org> wrote:
> 
> On Fri, 3 May 2019 11:29:59 +0200
> Peter Zijlstra <peterz@...radead.org> wrote:
> 
> 
>> OMG, WTF, ARGH... That code is fsck'ing horrible. I'd almost argue to
>> always do the INT3 thing, just to avoid games like that.
> 
> Hehe, that's almost the exact same thoughts I had when seeing this
> code ;-)
> 
>> 
>> That said; for normal traps &regs->sp is indeed the previous context --
>> if it doesn't fall off the stack. Your hack detects the regular INT3
>> frame. Howver if regs->sp has been modified (int3_emulate_push, for
>> example) your detectoring comes unstuck.
> 
> Yep. I realized the issue as well. But wanted to make sure this did
> work when sp wasn't changed.
> 
>> 
>> Now, it is rather unlikely these two code paths interact, but just to be
>> safe, something like so might be more reliable:
>> 
>> 
>> diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
>> index 4b8ee05dd6ad..aceaad0cc9a9 100644
>> --- a/arch/x86/kernel/ptrace.c
>> +++ b/arch/x86/kernel/ptrace.c
>> @@ -163,6 +163,9 @@ static inline bool invalid_selector(u16 value)
>>  * stack pointer we fall back to regs as stack if no previous stack
>>  * exists.
>>  *
>> + * There is a special case for INT3, there we construct a full pt_regs
>> + * environment. We can detect this case by a high bit in regs->cs
>> + *
>>  * This is valid only for kernel mode traps.
>>  */
>> unsigned long kernel_stack_pointer(struct pt_regs *regs)
>> @@ -171,6 +174,9 @@ unsigned long kernel_stack_pointer(struct pt_regs *regs)
>>    unsigned long sp = (unsigned long)&regs->sp;
>>    u32 *prev_esp;
>> 
>> +    if (regs->__csh & (1 << 13)) /* test CS_FROM_INT3 */
>> +        return regs->sp;
>> +
> 
> Thanks, I was looking into doing something like this (setting a flag in
> the int3 code), but didn't have the time to see the best way to do this.
> 
> I'll add this version of the code and run it through my tests.
> 
> -- Steve
> 
>>    if (context == (sp & ~(THREAD_SIZE - 1)))
>>        return sp;
>> 
>> --- a/arch/x86/entry/entry_32.S
>> +++ b/arch/x86/entry/entry_32.S
>> @@ -388,6 +388,7 @@
>> 
>> #define CS_FROM_ENTRY_STACK    (1 << 31)
>> #define CS_FROM_USER_CR3    (1 << 30)
>> +#define CS_FROM_INT3        (1 << 29)
>> 
>> .macro SWITCH_TO_KERNEL_STACK
>> 
>> @@ -1515,6 +1516,9 @@ ENTRY(int3)
>> 
>>    add    $16, 12(%esp) # point sp back at the previous context
>> 
>> +    andl    $0x0000ffff, 4(%esp)
>> +    orl    $CS_FROM_INT3, 4(%esp)
>> +
>>    pushl    $-1                # orig_eax; mark as interrupt
>> 
>>    SAVE_ALL
> 

So here’s a somewhat nutty suggestion: how about we tweak the 32-bit entry code to emulate the sane 64-bit frame, not just for int3 but always?  Basically, the entry asm for entries from kernel mode would do, roughly:

push $0 ;dummy for call emulation
push %ss
push $0 ;a dummy for ESP
push 3*4(%esp) ;EFLAGS
push 3*4(%esp) ;CS
push 3*4(%esp) ;EIP
push %rax
lea 7*4(%esp), %rax
mov %rax, 4*4(%esp) ;ESP

And the exit asm would do a little dance to write EFLAGS, CS, and EIP to the right spot, then load ESP-3*4 into %esp and do IRET.

Now the annoying kernel_stack_pointer() hack can just go away, since regs->sp is always correct!

I probably screwed up some arithmetic there, but it’s the idea that counts :)