linux-kernel - Re: [PATCH v4 2/5] x86/alternative: add indirect call patching

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: <9ccdcff2-90f9-47da-9aea-46cffe7fae85@suse.com>
Date:   Tue, 14 Nov 2023 13:56:37 +0100
From:   Juergen Gross <jgross@...e.com>
To:     Borislav Petkov <bp@...en8.de>
Cc:     linux-kernel@...r.kernel.org, x86@...nel.org,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        "H. Peter Anvin" <hpa@...or.com>,
        Peter Zijlstra <peterz@...radead.org>
Subject: Re: [PATCH v4 2/5] x86/alternative: add indirect call patching

On 14.11.23 13:37, Borislav Petkov wrote:
> On Mon, Oct 30, 2023 at 03:25:05PM +0100, Juergen Gross wrote:
>> + * Rewrite the "call BUG_func" replacement to point to the target of the
>> + * indirect pv_ops call "call *disp(%ip)".
>> + */
>> +static int alt_replace_call(u8 *instr, u8 *insn_buff, struct alt_instr *a)
>> +{
>> +	void *target, *bug = &BUG_func;
>> +
>> +	if (a->replacementlen != 5 || insn_buff[0] != CALL_INSN_OPCODE) {
>> +		pr_err("Alternative: ALT_FLAG_CALL set for a non-call replacement instruction\n");
> 
> No need for the printk prefix.

Okay.

> 
>> +		pr_err("  Ignoring the flag for the instruction at %pS (%px)\n", instr, instr);
> 
> No ignoring - BUG

Okay.

> 
>> +		return 5;
>> +	}
>> +
>> +	if (a->instrlen != 6 || instr[0] != 0xff || instr[1] != 0x15) {
>> +		pr_err("Alternative: ALT_FLAG_CALL set for unrecognized indirect call\n");
>> +		pr_err("  Not replacing the instruction at %pS (%px)\n", instr, instr);
>> +		return -1;
> 
> Ditto.

Okay.

> 
>> +	}
>> +
>> +#ifdef CONFIG_X86_64
>> +	/* ff 15 00 00 00 00   call   *0x0(%rip) */
>> +	target = *(void **)(instr + a->instrlen + *(s32 *)(instr + 2));
>> +#else
>> +	/* ff 15 00 00 00 00   call   *0x0 */
>> +	target = *(void **)(*(s32 *)(instr + 2));
>> +#endif
>> +	if (!target)
>> +		target = bug;
>> +
>> +	/* (BUG_func - .) + (target - BUG_func) := target - . */
>> +	*(s32 *)(insn_buff + 1) += target - bug;
> 
> If I squint hard enough, this looks like it is replacing one call with
> another. We have a C macro alternative_call() which does exactly that.
> Why can't you define an asm version ALTERNATIVE_CALL and use it
> instead of using adding a new flag? We have 16 possible ones in total.

It is replacing an _indirect_ call with a _direct_ one, taking the call target
from the pointer used by the indirect call.


Juergen

Download attachment "OpenPGP_0xB0DE9DD628BF132F.asc" of type "application/pgp-keys" (3099 bytes)

Download attachment "OpenPGP_signature.asc" of type "application/pgp-signature" (496 bytes)