linux-kernel - Re: [patch 4/4] x86: apic_is_clustered_box to indicate unsynched TSC's on multiboard vSMP systems

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: <20080322185948.GD23139@localdomain>
Date:	Sat, 22 Mar 2008 11:59:48 -0700
From:	Ravikiran G Thirumalai <kiran@...lex86.org>
To:	Ingo Molnar <mingo@...e.hu>
Cc:	Yinghai Lu <yhlu.kernel@...il.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	linux-kernel@...r.kernel.org,
	Glauber de Oliveira Costa <gcosta@...hat.com>,
	shai@...lex86.org
Subject: Re: [patch 4/4] x86: apic_is_clustered_box to indicate unsynched TSC's on multiboard vSMP systems

On Fri, Mar 21, 2008 at 07:58:21PM +0100, Ingo Molnar wrote:
>
>* Ravikiran G Thirumalai <kiran@...lex86.org> wrote:
>
>> As for the observation about probing the pci space early during the 
>> bootup, we call vsmp_init() much earlier during the bootup, which 
>> calls is_vsmp_box(), does the pci probing and caches the result in the 
>> flag, as you suggest. So the call in the above diff context does not 
>> access the pci config space as is.
>
>ah, i see - indeed - the trick with -1 :-)
>
>my point remains though: if you initialize VSMP in a separate function 
>anyway then please move this PCI config space access from is_vsmp_box() 
>into vsmp_init() and keep a pure flag return is_vsmp_box(). That way 
>there can be no question at all whether there are (or can be) any 
>side-effects of that function.
>

OK. Something like this?

vSMP detection: Access pci config space early in boot to detect if the
system is a vSMPowered box, and cache the result in a flag, so that
is_vsmp_box() retrieves the value of the flag always.

Signed-off-by; Ravikiran Thirumalai <kiran@...lex86.org>

Index: linux.git.trees/arch/x86/kernel/vsmp_64.c
===================================================================
--- linux.git.trees.orig/arch/x86/kernel/vsmp_64.c	2008-03-21 13:15:17.000000000 -0700
+++ linux.git.trees/arch/x86/kernel/vsmp_64.c	2008-03-22 11:58:42.313580356 -0700
@@ -108,25 +108,34 @@ static void __init set_vsmp_pv_ops(void)
 #endif
 
 #ifdef CONFIG_PCI
-static int vsmp = -1;
+static int is_vsmp = -1;
 
-int is_vsmp_box(void)
+static void __init detect_vsmp_box(void)
 {
-	if (vsmp != -1)
-		return vsmp;
+	is_vsmp = 0;
 
-	vsmp = 0;
 	if (!early_pci_allowed())
-		return vsmp;
+		return;
 
-	/* Check if we are running on a ScaleMP vSMP box */
+	/* Check if we are running on a ScaleMP vSMPowered box */
 	if (read_pci_config(0, 0x1f, 0, PCI_VENDOR_ID) ==
 	     (PCI_VENDOR_ID_SCALEMP | (PCI_DEVICE_ID_SCALEMP_VSMP_CTL << 16)))
-		vsmp = 1;
+		is_vsmp = 1;
+}
 
-	return vsmp;
+int is_vsmp_box(void)
+{
+	if (is_vsmp != -1)
+		return is_vsmp;
+	else {
+		printk("ScaleMP vSMPowered system detection called too early!\n");
+		return 0;
+	}
 }
 #else
+static int __init detect_vsmp_box(void)
+{
+}
 int is_vsmp_box(void)
 {
 	return 0;
@@ -135,6 +144,7 @@ int is_vsmp_box(void)
 
 void __init vsmp_init(void)
 {
+	detect_vsmp_box();
 	if (!is_vsmp_box())
 		return;
 
--
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/