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-next>] [day] [month] [year] [list]
Date:	Wed, 02 Jun 2010 16:22:02 -0500
From:	Cliff Wickman <cpw@....com>
To:	linux-kernel@...r.kernel.org
Cc:	mingo@...e.hu, hpa@...or.com, gregkh@...e.de
Subject: [PATCH 8/11] x86, UV: remove BAU check for stay-busy


Remove a faulty assumption that a long running BAU request has encountered
a hardware problem and will never finish.

Numalink congestion can make a request appear to have encountered such a
problem, but it is not safe to cancel the request.  If such a cancel is
done but a reply is later received we can miss a TLB shootdown.

We depend upon the max_bau_concurrent 'throttle' to prevent the stay-busy
case from happening.

Diffed against 2.6.34 -tip

Signed-off-by: Cliff Wickman <cpw@....com>
---
 arch/x86/include/asm/uv/uv_bau.h |    1 -
 arch/x86/kernel/tlb_uv.c         |   23 -----------------------
 2 files changed, 24 deletions(-)

Index: 100531.linux-tip/arch/x86/kernel/tlb_uv.c
===================================================================
--- 100531.linux-tip.orig/arch/x86/kernel/tlb_uv.c
+++ 100531.linux-tip/arch/x86/kernel/tlb_uv.c
@@ -405,12 +405,10 @@ static int uv_wait_completion(struct bau
 	unsigned long mmr;
 	unsigned long mask;
 	cycles_t ttime;
-	cycles_t timeout_time;
 	struct ptc_stats *stat = bcp->statp;
 	struct bau_control *hmaster;
 
 	hmaster = bcp->uvhub_master;
-	timeout_time = get_cycles() + bcp->timeout_interval;
 
 	/* spin on the status MMR, waiting for it to go idle */
 	while ((descriptor_status = (((unsigned long)
@@ -450,26 +448,6 @@ static int uv_wait_completion(struct bau
 			 * descriptor_status is still BUSY
 			 */
 			cpu_relax();
-			relaxes++;
-			if (relaxes >= 10000) {
-				relaxes = 0;
-				if (get_cycles() > timeout_time) {
-					quiesce_local_uvhub(hmaster);
-
-					/* single-thread the register change */
-					spin_lock(&hmaster->masks_lock);
-					mmr = uv_read_local_mmr(mmr_offset);
-					mask = 0UL;
-					mask |= (3UL < right_shift);
-					mask = ~mask;
-					mmr &= mask;
-					uv_write_local_mmr(mmr_offset, mmr);
-					spin_unlock(&hmaster->masks_lock);
-					end_uvhub_quiesce(hmaster);
-					stat->s_busy++;
-					return FLUSH_GIVEUP;
-				}
-			}
 		}
 	}
 	bcp->conseccompletes++;
@@ -1580,7 +1558,6 @@ static void uv_init_per_cpu(int nuvhubs)
 	for_each_present_cpu(cpu) {
 		bcp = &per_cpu(bau_control, cpu);
 		memset(bcp, 0, sizeof(struct bau_control));
-		spin_lock_init(&bcp->masks_lock);
 		pnode = uv_cpu_hub_info(cpu)->pnode;
 		uvhub = uv_cpu_hub_info(cpu)->numa_blade_id;
 		uvhub_mask |= (1 << uvhub);
Index: 100531.linux-tip/arch/x86/include/asm/uv/uv_bau.h
===================================================================
--- 100531.linux-tip.orig/arch/x86/include/asm/uv/uv_bau.h
+++ 100531.linux-tip/arch/x86/include/asm/uv/uv_bau.h
@@ -402,7 +402,6 @@ struct bau_control {
 	unsigned short uvhub_quiesce;
 	short socket_acknowledge_count[DEST_Q_SIZE];
 	cycles_t send_message;
-	spinlock_t masks_lock;
 	spinlock_t uvhub_lock;
 	spinlock_t queue_lock;
 	/* tunables */
--
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/

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ