linux-kernel - Re: [PATCH] x86/traps: add CONFIG_BUG to the use of _

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: <20220425192856.GG2762@worktop.programming.kicks-ass.net>
Date:   Mon, 25 Apr 2022 21:28:56 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     Dave Hansen <dave.hansen@...el.com>
Cc:     Stephen Zhang <starzhangzsd@...il.com>, tglx@...utronix.de,
        mingo@...hat.com, bp@...en8.de, dave.hansen@...ux.intel.com,
        x86@...nel.org, hpa@...or.com, laijs@...ux.alibaba.com,
        lihuafei1@...wei.com, fenghua.yu@...el.com,
        chang.seok.bae@...el.com, zhangshida@...inos.cn,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] x86/traps: add CONFIG_BUG to the use of __warn()

On Mon, Apr 25, 2022 at 08:24:56PM +0200, Peter Zijlstra wrote:
> On Mon, Apr 25, 2022 at 09:05:17AM -0700, Dave Hansen wrote:
> > On 4/24/22 22:59, Stephen Zhang wrote:
> > > --- a/arch/x86/kernel/traps.c
> > > +++ b/arch/x86/kernel/traps.c
> > > @@ -246,7 +246,9 @@ DEFINE_IDTENTRY_ERRORCODE(exc_control_protection)
> > >  	pr_err("Missing ENDBR: %pS\n", (void *)instruction_pointer(regs));
> > >  	if (!ibt_fatal) {
> > >  		printk(KERN_DEFAULT CUT_HERE);
> > > +#ifdef CONFIG_BUG
> > >  		__warn(__FILE__, __LINE__, (void *)regs->ip, TAINT_WARN, regs, NULL);
> > > +#endif
> > 
> > This really should be done with an #ifdef'd stub in in bug.h, not an
> > #ifdef at every call site.  I assume there was a good reason for not
> > using the normal WARN*() macros.
> 
> I think the idea was to have the pr_err() unconditional, and if you have
> the fancy crap on also print a full backtrace or something.
> 
> But yeah, a few stubs in bug.h ought to make this go away too.

As pointed out by Boris, __warn is actually uncondtionally defined in
panic.c, the below seems to build.

---
diff --git a/include/asm-generic/bug.h b/include/asm-generic/bug.h
index edb0e2a602a8..ba1f860af38b 100644
--- a/include/asm-generic/bug.h
+++ b/include/asm-generic/bug.h
@@ -21,6 +21,12 @@
 #include <linux/panic.h>
 #include <linux/printk.h>
 
+struct warn_args;
+struct pt_regs;
+
+void __warn(const char *file, int line, void *caller, unsigned taint,
+	    struct pt_regs *regs, struct warn_args *args);
+
 #ifdef CONFIG_BUG
 
 #ifdef CONFIG_GENERIC_BUG
@@ -110,11 +116,6 @@ extern __printf(1, 2) void __warn_printk(const char *fmt, ...);
 #endif
 
 /* used internally by panic.c */
-struct warn_args;
-struct pt_regs;
-
-void __warn(const char *file, int line, void *caller, unsigned taint,
-	    struct pt_regs *regs, struct warn_args *args);
 
 #ifndef WARN_ON
 #define WARN_ON(condition) ({						\