linux-kernel - [PATCH] x86: fix ioport unification on 32-bit [was: Re: hwclock failure in x86.git]

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20080112010443.GA5476@sequoia.sous-sol.org>
Date:	Fri, 11 Jan 2008 17:04:44 -0800
From:	Chris Wright <chrisw@...s-sol.org>
To:	Ingo Molnar <mingo@...e.hu>
Cc:	Kevin Winchester <kjwinchester@...il.com>,
	"H. Peter Anvin" <hpa@...or.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	"mboton@...il.com" <mboton@...il.com>,
	Zach Brown <zach.brown@...cle.com>
Subject: [PATCH] x86: fix ioport unification on 32-bit [was: Re: hwclock
	failure in x86.git]

* Ingo Molnar (mingo@...e.hu) wrote:
> thanks for tracking it down. I pulled that commit for now. But it would 
> be nice to figure out what's going on there.

Zach was right. The unification was broken for 32-bit; it was missing
the actual pushf/popf EFLAGS manipluation (set_iopl_mask()) and would've
broken task switching between processes w/ different iopl in paravirt
guests too.  64-bit sys_iopl just does pt_regs->flags modification and
lets syscall/sysret plus ptregscall sync and do EFLAGS update.

Also, use of volatile looks like leftover cruft.

This patch in on top of Miguel's (can respin to standalone if that's better).
Tested (on both 32 and 64-bit) with simple:

  #include <stdlib.h>
  #include <sys/io.h>
  
  main()
  {
  	if (iopl(3) == 0)
  		asm ("cli\nsti\n"::);
  }

thanks,
-chris
--

From: Chris Wright <chrisw@...s-sol.org>
Subject: [PATCH] x86: fix ioport unification on 32-bit

ioport unification was broken for 32-bit; it was missing
the acutal pushf/popf EFLAGS manipulation (set_iopl_mask()).
Also, use of volatile looks like leftover cruft.

Cc: mboton@...il.com
Cc: Kevin Winchester <kjwinchester@...il.com>
Cc: Zach Brown <zach.brown@...cle.com>
Cc: Ingo Molnar <mingo@...e.hu>
Cc: "H. Peter Anvin" <hpa@...or.com>
Cc: Thomas Gleixner <tglx@...utronix.de>
Signed-off-by: Chris Wright <chrisw@...s-sol.org>
---
 arch/x86/kernel/ioport.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kernel/ioport.c b/arch/x86/kernel/ioport.c
index e723ff3..be72d80 100644
--- a/arch/x86/kernel/ioport.c
+++ b/arch/x86/kernel/ioport.c
@@ -116,9 +116,10 @@ asmlinkage long sys_ioperm(unsigned long from, unsigned long num, int turn_on)
 #ifdef CONFIG_X86_32
 asmlinkage long sys_iopl(unsigned long regsp)
 {
-	volatile struct pt_regs *regs = (struct pt_regs *)&regsp;
+	struct pt_regs *regs = (struct pt_regs *)&regsp;
 	unsigned int level = regs->bx;
 	unsigned int old = (regs->flags >> 12) & 3;
+	struct thread_struct *t = &current->thread;
 
 	if (level > 3)
 		return -EINVAL;
@@ -127,8 +128,9 @@ asmlinkage long sys_iopl(unsigned long regsp)
 		if (!capable(CAP_SYS_RAWIO))
 			return -EPERM;
 	}
+	t->iopl = level << 12;
 	regs->flags = (regs->flags & ~X86_EFLAGS_IOPL) | (level << 12);
-
+	set_iopl_mask(t->iopl);
 	return 0;
 }
 #else
--
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/