public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Greg KH <gregkh@suse.de>
To: linux-kernel@vger.kernel.org, stable@kernel.org
Cc: stable-review@kernel.org, torvalds@linux-foundation.org,
	akpm@linux-foundation.org, alan@lxorguk.ukuu.org.uk,
	Joerg Roedel <joerg.roedel@amd.com>,
	lethal@linux-sh.org, just.for.lkml@googlemail.com,
	hancockrwd@gmail.com, jens.axboe@oracle.com,
	bharrosh@panasas.com,
	FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>,
	Ingo Molnar <mingo@elte.hu>
Subject: [patch 025/108] dma-debug: change hash_bucket_find from first-fit to best-fit
Date: Tue, 30 Jun 2009 17:23:14 -0700	[thread overview]
Message-ID: <20090701002420.141480274@mini.kroah.org> (raw)
In-Reply-To: <20090701002838.GA7100@kroah.com>

[-- Attachment #1: dma-debug-change-hash_bucket_find-from-first-fit-to-best-fit.patch --]
[-- Type: text/plain, Size: 2991 bytes --]

2.6.30-stable review patch.  If anyone has any objections, please let us know.

------------------

From: Joerg Roedel <joerg.roedel@amd.com>

commit 7caf6a49bb17d0377210693af5737563b31aa5ee upstream.

Some device drivers map the same physical address multiple times to a
dma address. Without an IOMMU this results in the same dma address being
put into the dma-debug hash multiple times. With a first-fit match in
hash_bucket_find() this function may return the wrong dma_debug_entry.

This can result in false positive warnings. This patch fixes it by
changing the first-fit behavior of hash_bucket_find() into a best-fit
algorithm.

Reported-by: Torsten Kaiser <just.for.lkml@googlemail.com>
Reported-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Cc: lethal@linux-sh.org
Cc: just.for.lkml@googlemail.com
Cc: hancockrwd@gmail.com
Cc: jens.axboe@oracle.com
Cc: bharrosh@panasas.com
Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <20090605104132.GE24836@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

---
 lib/dma-debug.c |   43 +++++++++++++++++++++++++++++++++++++++----
 1 file changed, 39 insertions(+), 4 deletions(-)

--- a/lib/dma-debug.c
+++ b/lib/dma-debug.c
@@ -185,15 +185,50 @@ static void put_hash_bucket(struct hash_
 static struct dma_debug_entry *hash_bucket_find(struct hash_bucket *bucket,
 						struct dma_debug_entry *ref)
 {
-	struct dma_debug_entry *entry;
+	struct dma_debug_entry *entry, *ret = NULL;
+	int matches = 0, match_lvl, last_lvl = 0;
 
 	list_for_each_entry(entry, &bucket->list, list) {
-		if ((entry->dev_addr == ref->dev_addr) &&
-		    (entry->dev == ref->dev))
+		if ((entry->dev_addr != ref->dev_addr) ||
+		    (entry->dev != ref->dev))
+			continue;
+
+		/*
+		 * Some drivers map the same physical address multiple
+		 * times. Without a hardware IOMMU this results in the
+		 * same device addresses being put into the dma-debug
+		 * hash multiple times too. This can result in false
+		 * positives being reported. Therfore we implement a
+		 * best-fit algorithm here which returns the entry from
+		 * the hash which fits best to the reference value
+		 * instead of the first-fit.
+		 */
+		matches += 1;
+		match_lvl = 0;
+		entry->size      == ref->size      ? ++match_lvl : match_lvl;
+		entry->type      == ref->type      ? ++match_lvl : match_lvl;
+		entry->direction == ref->direction ? ++match_lvl : match_lvl;
+
+		if (match_lvl == 3) {
+			/* perfect-fit - return the result */
 			return entry;
+		} else if (match_lvl > last_lvl) {
+			/*
+			 * We found an entry that fits better then the
+			 * previous one
+			 */
+			last_lvl = match_lvl;
+			ret      = entry;
+		}
 	}
 
-	return NULL;
+	/*
+	 * If we have multiple matches but no perfect-fit, just return
+	 * NULL.
+	 */
+	ret = (matches == 1) ? ret : NULL;
+
+	return ret;
 }
 
 /*



  parent reply	other threads:[~2009-07-01  0:53 UTC|newest]

Thread overview: 124+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20090701002249.937782934@mini.kroah.org>
2009-07-01  0:28 ` [patch 000/108] 2.6.30-stable review Greg KH
2009-07-01  0:22   ` [patch 001/108] firmware_map: fix hang with x86/32bit Greg KH
2009-07-01  0:22   ` [patch 002/108] fs: remove incorrect I_NEW warnings Greg KH
2009-07-01  0:22   ` [patch 003/108] PCI: disable ASPM on VIA root-port-under-bridge configurations Greg KH
2009-07-01  0:22   ` [patch 004/108] KVM: VMX: Fix handling of a fault during NMI unblocked due to IRET Greg KH
2009-07-01  0:22   ` [patch 005/108] KVM: Move "exit due to NMI" handling into vmx_complete_interrupts() Greg KH
2009-07-01  0:22   ` [patch 006/108] KVM: Add VT-x machine check support Greg KH
2009-07-01  0:22   ` [patch 007/108] KVM: Disable large pages on misaligned memory slots Greg KH
2009-07-01  0:22   ` [patch 008/108] KVM: Prevent overflow in largepages calculation Greg KH
2009-07-01  0:22   ` [patch 009/108] KVM: x86: check for cr3 validity in ioctl_set_sregs Greg KH
2009-07-01  0:22   ` [patch 010/108] KVM: VMX: Handle vmx instruction vmexits Greg KH
2009-07-01  0:23   ` [patch 011/108] KVM: protect concurrent make_all_cpus_request Greg KH
2009-07-01  0:23   ` [patch 012/108] KVM: Fix dirty bit tracking for slots with large pages Greg KH
2009-07-01  0:23   ` [patch 013/108] IMA: use current_cred() instead of current->cred Greg KH
2009-07-01  0:23   ` [patch 014/108] IMA: Handle dentry_open failures Greg KH
2009-07-01  0:23   ` [patch 015/108] IMA: open all files O_LARGEFILE Greg KH
2009-07-01  0:23   ` [patch 016/108] e1000e: stop unnecessary polling when using msi-x Greg KH
2009-07-01  0:23   ` [patch 017/108] pegasus usb-net: Fix endianness bugs Greg KH
2009-07-01  0:23   ` [patch 018/108] ipv4: fix NULL pointer + success return in route lookup path Greg KH
2009-07-01  0:23   ` [patch 019/108] ipv4 routing: Ensure that route cache entries are usable and reclaimable with caching is off Greg KH
2009-07-01  0:23   ` [patch 020/108] sky2: dont look for VPD size Greg KH
2009-07-01  0:23   ` [patch 021/108] tun: Fix unregister race Greg KH
2009-07-01  0:23   ` [patch 022/108] via-velocity: Fix velocity driver unmapping incorrect size Greg KH
2009-07-01  0:23   ` [patch 023/108] x25: Fix sleep from timer on socket destroy Greg KH
2009-07-01  0:23   ` [patch 024/108] bonding: fix multiple module load problem Greg KH
2009-07-01  0:23   ` Greg KH [this message]
2009-07-01  0:23   ` [patch 026/108] char: moxa, prevent opening unavailable ports Greg KH
2009-07-01  0:23   ` [patch 027/108] serial: refactor ASYNC_ flags Greg KH
2009-07-01  0:23   ` [patch 028/108] rocket: fix test_bit parameters Greg KH
2009-07-01  0:23   ` [patch 029/108] epca: " Greg KH
2009-07-01  0:23   ` [patch 030/108] x86: Detect use of extended APIC ID for AMD CPUs Greg KH
2009-07-01  0:23   ` [patch 031/108] USB: usbtmc: fix switch statment Greg KH
2009-07-01  0:23   ` [patch 032/108] DVB: lgdt3305: fix 64bit division in function lgdt3305_set_if Greg KH
2009-07-01  0:23   ` [patch 033/108] V4L: ivtv/cx18: fix regression: class controls are no longer seen Greg KH
2009-07-01  0:23   ` [patch 034/108] V4L: pvrusb2: Fix hardware scaling when used with cx25840 Greg KH
2009-07-01  0:23   ` [patch 035/108] V4L: pvrusb2: Re-fix hardware scaling on video standard change Greg KH
2009-07-01  0:23   ` [patch 036/108] V4L: i2c modules must be linked before the v4l2 drivers Greg KH
2009-07-01  0:23   ` [patch 037/108] sym53c8xx: ratelimit parity errors Greg KH
2009-07-01  0:23   ` [patch 038/108] ISDN: Fix DMA alloc for hfcpci Greg KH
2009-07-01  0:23   ` [patch 039/108] jfs: fix regression preventing coalescing of extents Greg KH
2009-07-01  0:23   ` [patch 040/108] spi: takes size of a pointer to determine the size of the pointed-to type Greg KH
2009-07-01  0:23   ` [patch 041/108] serial: bfin_5xx: add missing spin_lock init Greg KH
2009-07-01  0:23   ` [patch 042/108] x86: memtest: remove 64-bit division Greg KH
2009-07-01  0:23   ` [patch 043/108] x86: Fix UV BAU activation descriptor init Greg KH
2009-07-01  0:23   ` [patch 044/108] x86, UV: Fix macros for multiple coherency domains Greg KH
2009-07-01  0:23   ` [patch 045/108] x86: enable GART-IOMMU only after setting up protection methods Greg KH
2009-07-01  0:23   ` [patch 046/108] x86: move rdtsc_barrier() into the TSC vread method Greg KH
2009-07-01  0:23   ` [patch 047/108] x86: Fix uv bau sending buffer initialization Greg KH
2009-07-01  0:23   ` [patch 048/108] x86: Add quirk for reboot stalls on a Dell Optiplex 360 Greg KH
2009-07-01  0:23   ` [patch 049/108] x86: handle initrd that extends into unusable memory Greg KH
2009-07-01  0:23   ` [patch 050/108] ALSA: ca0106 - Add missing registrations of vmaster controls Greg KH
2009-07-01  0:23   ` [patch 051/108] ALSA: intel8x0 - Fix PCM position craziness Greg KH
2009-07-01  0:23   ` [patch 052/108] ALSA: hda - Get back Input Source for ALC262 toshiba-s06 model Greg KH
2009-07-01  0:23   ` [patch 053/108] ALSA: hda - Add quirk for Sony VAIO Z21MN Greg KH
2009-07-01  0:23   ` [patch 054/108] ALSA: cmi8330: fix MPU-401 PnP init copy&paste bug Greg KH
2009-07-01  0:23   ` [patch 055/108] x86: hpet: Mark per cpu interrupts IRQF_TIMER to prevent resume failure Greg KH
2009-07-01  0:23   ` [patch 056/108] ARM: 5545/2: add flush_kernel_dcache_page() for ARM Greg KH
2009-07-01  0:23   ` [patch 057/108] IB/mlx4: Add strong ordering to local inval and fast reg work requests Greg KH
2009-07-01  0:23   ` [patch 058/108] epoll: fix nested calls support Greg KH
2009-07-01  0:23   ` [patch 059/108] lockdep: Select frame pointers on x86 Greg KH
2009-07-01  0:23   ` [patch 060/108] ASoC: Remove odd bit clock ratios for WM8903 Greg KH
2009-07-01  0:23   ` [patch 061/108] ramfs: ignore unknown mount options Greg KH
2009-07-01  0:23   ` [patch 062/108] mac80211: fix minstrel single-rate memory corruption Greg KH
2009-07-01  0:23   ` [patch 063/108] cfg80211: fix for duplicate userspace replies Greg KH
2009-07-01  0:23   ` [patch 064/108] cfg80211: cleanup return calls on nl80211_set_reg() Greg KH
2009-07-01  0:23   ` [patch 065/108] cfg80211: return immediately if num reg rules > NL80211_MAX_SUPP_REG_RULES Greg KH
2009-07-01  0:23   ` [patch 066/108] cfg80211: fix in nl80211_set_reg() Greg KH
2009-07-01  0:23   ` [patch 067/108] ath9k: Fix bug when using a card with a busted EEPROM Greg KH
2009-07-01  0:23   ` [patch 068/108] ath9k: Fix bug in calibration initialization Greg KH
2009-07-01  0:23   ` [patch 069/108] ath9k: Fix bug in determining calibration support Greg KH
2009-07-01  0:23   ` [patch 070/108] ath9k: Fix bug in checking HT flag Greg KH
2009-07-01  0:24   ` [patch 071/108] ath9k: Fix bug in scan termination Greg KH
2009-07-01  0:24   ` [patch 072/108] ath9k: Fix memleak on TX DMA failure Greg KH
2009-07-01  0:24   ` [patch 073/108] ath9k: Initialize ANI timers Greg KH
2009-07-01  0:24   ` [patch 074/108] ath9k: Fix PCI FATAL interrupts by restoring RETRY_TIMEOUT disabling Greg KH
2009-07-01  0:24   ` [patch 075/108] crypto: aes-ni - Fix cbc mode IV saving Greg KH
2009-07-01  0:24   ` [patch 076/108] md/raid5: add missing call to schedule() after prepare_to_wait() Greg KH
2009-07-01  0:24   ` [patch 077/108] tracing/urgent: fix unbalanced ftrace_start_up Greg KH
2009-07-01  0:24   ` [patch 078/108] cifs: fix fh_mutex locking in cifs_reopen_file Greg KH
2009-07-01  0:24   ` [patch 079/108] vt_ioctl: fix lock imbalance Greg KH
2009-07-01  0:24   ` [patch 080/108] x86: Fix non-lazy GS handling in sys_vm86() Greg KH
2009-07-01  0:24   ` [patch 081/108] x86: Set cpu_llc_id on AMD CPUs Greg KH
2009-07-01  0:24   ` [patch 082/108] usb-serial: replace shutdown with disconnect, release Greg KH
2009-07-01  0:24   ` [patch 083/108] pcmcia/cm4000: fix lock imbalance Greg KH
2009-07-01  0:24   ` [patch 084/108] n_r3964: " Greg KH
2009-07-01  0:24   ` [patch 085/108] parport_pc: set properly the dma_mask for parport_pc device Greg KH
2009-07-01  0:24   ` [patch 086/108] parport_pc: after superio probing restore original register values Greg KH
2009-07-01  0:24   ` [patch 087/108] mv643xx_eth: fix unicast filter programming in promiscuous mode Greg KH
2009-07-01  0:24   ` [patch 088/108] ath5k: avoid PCI FATAL interrupts by restoring RETRY_TIMEOUT disabling Greg KH
2009-07-01  0:24   ` [patch 089/108] sound: seq_midi_event: fix decoding of (N)RPN events Greg KH
2009-07-01  0:24   ` [patch 090/108] PCI PM: Fix handling of devices without PM support by pci_target_state() Greg KH
2009-07-01  0:24   ` [patch 091/108] PCI PM: Follow PCI_PM_CTRL_NO_SOFT_RESET during transitions from D3 Greg KH
2009-07-01  0:24   ` [patch 092/108] qla2xxx: Correct (again) overflow during dump-processing on large-memory ISP23xx parts Greg KH
2009-07-01  0:24   ` [patch 093/108] mm: fix handling of pagesets for downed cpus Greg KH
2009-07-01  0:24   ` [patch 094/108] dm mpath: validate hw_handler argument count Greg KH
2009-07-01  0:24   ` [patch 095/108] dm mpath: validate table " Greg KH
2009-07-01  0:24   ` [patch 096/108] dm: sysfs skip output when device is being destroyed Greg KH
2009-07-01  0:24   ` [patch 097/108] dm mpath: flush keventd queue in destructor Greg KH
2009-07-01  0:24   ` [patch 098/108] dm exception store: fix exstore lookup to be case insensitive Greg KH
2009-07-01  7:11     ` Milan Broz
2009-07-01 17:26       ` [Stable-review] " Greg KH
2009-07-01  0:24   ` [patch 099/108] dm: use i_size_read Greg KH
2009-07-01  0:24   ` [patch 100/108] vmscan: properly account for the number of page cache pages zone_reclaim() can reclaim Greg KH
2009-07-01  0:24   ` [patch 101/108] vmscan: count the number of times zone_reclaim() scans and fails Greg KH
2009-07-01  0:24   ` [patch 102/108] lib/genalloc.c: remove unmatched write_lock() in gen_pool_destroy Greg KH
2009-07-01  0:24   ` [patch 103/108] CONFIG_FILE_LOCKING should not depend on CONFIG_BLOCK Greg KH
2009-07-01  0:24   ` [patch 104/108] serial: bfin_5xx: fix building as module when early printk is enabled Greg KH
2009-07-01  0:24   ` [patch 105/108] ocfs2: Fix ocfs2_osb_dump() Greg KH
2009-07-01  0:24   ` [patch 106/108] ide-cd: prevent null pointer deref via cdrom_newpc_intr Greg KH
2009-07-01  0:24   ` [patch 107/108] drm/i915: correct suspend/resume ordering Greg KH
2009-07-01  0:24   ` [patch 108/108] KVM: x86: silence preempt warning on kvm_write_guest_time Greg KH
2009-07-01 11:22   ` [patch 000/108] 2.6.30-stable review Julien BLACHE
2009-07-01 15:02     ` Stefan Richter
2009-07-01 15:41       ` Julien BLACHE
2009-07-01 16:22         ` [Stable-review] " Luis R. Rodriguez
2009-07-01 16:27           ` Julien BLACHE
2009-07-01 16:43             ` Stefan Richter
2009-07-01 17:21     ` Greg KH
2009-07-01 18:10   ` Greg KH
2009-07-01 18:12     ` [patch 109/108] bsdacct: fix access to invalid filp in acct_on() Greg KH
2009-07-01 18:13     ` [patch 110/108] dm exception store: really fix type lookup Greg KH
2009-07-01 18:14     ` [patch 111/108] xfs: fix freeing memory in xfs_getbmap() Greg KH
2009-07-01 18:33     ` [patch 000/108] 2.6.30-stable review Greg KH
2009-07-01 18:24   ` [Stable-review] " Luis R. Rodriguez

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20090701002420.141480274@mini.kroah.org \
    --to=gregkh@suse.de \
    --cc=akpm@linux-foundation.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=bharrosh@panasas.com \
    --cc=fujita.tomonori@lab.ntt.co.jp \
    --cc=hancockrwd@gmail.com \
    --cc=jens.axboe@oracle.com \
    --cc=joerg.roedel@amd.com \
    --cc=just.for.lkml@googlemail.com \
    --cc=lethal@linux-sh.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=stable-review@kernel.org \
    --cc=stable@kernel.org \
    --cc=torvalds@linux-foundation.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox