linux-kernel - Re: [PATCH v2 09/12] x86/msr: Use the alternatives mechanism for WRMSR

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: <2ad137cb-ed38-42f6-ac0a-a81569051779@suse.com>
Date: Tue, 30 Sep 2025 17:42:13 +0200
From: Jürgen Groß <jgross@...e.com>
To: Peter Zijlstra <peterz@...radead.org>
Cc: linux-kernel@...r.kernel.org, x86@...nel.org, llvm@...ts.linux.dev,
 xin@...or.com, "H. Peter Anvin" <hpa@...or.com>,
 Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>,
 Borislav Petkov <bp@...en8.de>, Dave Hansen <dave.hansen@...ux.intel.com>,
 Nathan Chancellor <nathan@...nel.org>,
 Nick Desaulniers <nick.desaulniers+lkml@...il.com>,
 Bill Wendling <morbo@...gle.com>, Justin Stitt <justinstitt@...gle.com>
Subject: Re: [PATCH v2 09/12] x86/msr: Use the alternatives mechanism for
 WRMSR

On 30.09.25 14:51, Peter Zijlstra wrote:
> On Tue, Sep 30, 2025 at 10:50:44AM +0200, Peter Zijlstra wrote:
>> On Tue, Sep 30, 2025 at 10:46:23AM +0200, Jürgen Groß wrote:
> 
>>>> 	asm_inline volatile goto(
>>>> 		"1:\n"
>>>> 		ALTERNATIVE(PREPARE_RCX_RDX_FOR_WRMSR
>>>> 			    "2:\n"
>>>> 			    ALTERNATIVE("ds wrmsr",
>>>> 			                ASM_WRMSRNS, X86_FEATURE_WRMSRNS),
>>>> 			    ASM_WRMSRNS_IMM, X86_FEATURE_MSR_IMM);
>>>> 		_ASM_EXTABLE_TYPE(1b, %l[badmsr], %c[type])	/* For WRMSRNS immediate */
>>>> 		_ASM_EXTABLE_TYPE(2b, %l[badmsr], %c[type])	/* For WRMSR(NS) */
>>>>
>>>> 		:
>>>> 		: [val] "a" (val), [msr] "i" (msr), [type] "i" (type)
>>>> 		: "memory", "ecx", "rdx"
>>>> 		: badmsr);
>>>>
> 
>> Oh well, lets forget about this :-)
> 
> So I couldn't. I tried the below, which when building a .i generates the
> following:
> 
> 
> static inline __attribute__((__gnu_inline__)) __attribute__((__unused__)) __attribute__((no_instrument_function)) void clear_page(void *page)
> {
> 
>   kmsan_unpoison_memory(page, ((1UL) << 12));
>   asm __inline volatile(
>   	"# ALT: oldinstr\n"
> 	"__UNIQUE_ID_altinstr_9" "_begin:\n\t"
> 
> 		"# ALT: oldinstr\n"
> 		"__UNIQUE_ID_altinstr_8" "_begin:\n\t"
> 		"call %c[old]" "\n"
> 		"__UNIQUE_ID_altinstr_8" "_pad:\n"
> 		"# ALT: padding\n"
> 		".skip -(((" "__UNIQUE_ID_altinstr_8" "_alt_end - " "__UNIQUE_ID_altinstr_8" "_alt_begin" ")-(" "__UNIQUE_ID_altinstr_8" "_pad - " "__UNIQUE_ID_altinstr_8" "_begin" ")) > 0) * "
> 		"((" "__UNIQUE_ID_altinstr_8" "_alt_end - " "__UNIQUE_ID_altinstr_8" "_alt_begin" ")-(" "__UNIQUE_ID_altinstr_8" "_pad - " "__UNIQUE_ID_altinstr_8" "_begin" ")),0x90\n"
> 		"__UNIQUE_ID_altinstr_8" "_end:\n"
> 		".pushsection .altinstructions,\"a\"\n"
> 		" .long " "__UNIQUE_ID_altinstr_8" "_begin - .\n"
> 		" .long " "__UNIQUE_ID_altinstr_8" "_alt_begin - .\n"
> 		" .4byte " "( 3*32+16)" "\n"
> 		" .byte " "__UNIQUE_ID_altinstr_8" "_end - " "__UNIQUE_ID_altinstr_8" "_begin" "\n"
> 		" .byte " "__UNIQUE_ID_altinstr_8" "_alt_end - " "__UNIQUE_ID_altinstr_8" "_alt_begin" "\n"
> 		".popsection\n"
> 		".pushsection .altinstr_replacement, \"ax\"\n"
> 		"# ALT: replacement\n"
> 		"__UNIQUE_ID_altinstr_8" "_alt_begin:\n\t"
> 		"call %c[new1]" "\n"
> 		"__UNIQUE_ID_altinstr_8" "_alt_end:\n"
> 		".popsection\n"
> 		"\n"
> 
> 	"__UNIQUE_ID_altinstr_9" "_pad:\n"
> 	"# ALT: padding\n"
> 	".skip -(((" "__UNIQUE_ID_altinstr_9" "_alt_end - " "__UNIQUE_ID_altinstr_9" "_alt_begin" ")-(" "__UNIQUE_ID_altinstr_9" "_pad - " "__UNIQUE_ID_altinstr_9" "_begin" ")) > 0) * "
> 	"((" "__UNIQUE_ID_altinstr_9" "_alt_end - " "__UNIQUE_ID_altinstr_9" "_alt_begin" ")-(" "__UNIQUE_ID_altinstr_9" "_pad - " "__UNIQUE_ID_altinstr_9" "_begin" ")),0x90\n"
> 	"__UNIQUE_ID_altinstr_9" "_end:\n"
> 	".pushsection .altinstructions,\"a\"\n"
> 	" .long " "__UNIQUE_ID_altinstr_9" "_begin - .\n"
> 	" .long " "__UNIQUE_ID_altinstr_9" "_alt_begin - .\n"
> 	" .4byte " "( 9*32+ 9)" "\n"
> 	" .byte " "__UNIQUE_ID_altinstr_9" "_end - " "__UNIQUE_ID_altinstr_9" "_begin" "\n"
> 	" .byte " "__UNIQUE_ID_altinstr_9" "_alt_end - " "__UNIQUE_ID_altinstr_9" "_alt_begin" "\n"
> 	".popsection\n"
> 	".pushsection .altinstr_replacement, \"ax\"\n"
> 	"# ALT: replacement\n"
> 	"__UNIQUE_ID_altinstr_9" "_alt_begin:\n\t"
> 	"call %c[new2]" "\n"
> 	"__UNIQUE_ID_altinstr_9" "_alt_end:\n"
> 	".popsection\n"
> 	: "+r" (current_stack_pointer), "=D" (page)
> 	: [old] "i" (clear_page_orig), [new1] "i" (clear_page_rep), [new2] "i" (clear_page_erms) , "D" (page)
> 	: "cc", "memory", "rax", "rcx")
> 
>                                     ;
> }
> 
> Which looks right, but utterly fails to build :(

What does the failure look like?

Could it be that the labels should be local ones?


Juergen

Download attachment "OpenPGP_0xB0DE9DD628BF132F.asc" of type "application/pgp-keys" (3684 bytes)

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