From: Greg KH <gregkh@suse.de>
To: ath9k-devel@lists.ath9k.org
Subject: [ath9k-devel] [patch 47/58] ath9k: implement IO serialization
Date: Wed, 29 Apr 2009 15:07:46 -0700 [thread overview]
Message-ID: <20090429220929.995031553@mini.kroah.org> (raw)
In-Reply-To: <20090429221657.GA11765@kroah.com>
An embedded and charset-unspecified text was scrubbed...
Name: ath9k-implement-io-serialization.patch
Url: http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20090429/0eacbaf6/attachment.txt
WARNING: multiple messages have this Message-ID (diff)
From: Greg KH <gregkh@suse.de>
To: linux-kernel@vger.kernel.org, stable@kernel.org
Cc: Justin Forbes <jmforbes@linuxtx.org>,
Zwane Mwaikambo <zwane@arm.linux.org.uk>,
Theodore Ts'o <tytso@mit.edu>,
Randy Dunlap <rdunlap@xenotime.net>,
Dave Jones <davej@redhat.com>,
Chuck Wolber <chuckw@quantumlinux.com>,
Chris Wedgwood <reviews@ml.cw.f00f.org>,
Michael Krufky <mkrufky@linuxtv.org>,
Chuck Ebbert <cebbert@redhat.com>,
Domenico Andreoli <cavokz@gmail.com>, Willy Tarreau <w@1wt.eu>,
Rodrigo Rubira Branco <rbranco@la.checkpoint.com>,
Jake Edge <jake@lwn.net>, Eugene Teo <eteo@redhat.com>,
torvalds@linux-foundation.org, akpm@linux-foundation.org,
alan@lxorguk.ukuu.org.uk,
"Luis R. Rodriguez" <lrodriguez@atheros.com>,
ath9k-devel@venema.h4ckr.net, linux-wireless@vger.kernel.org
Subject: [patch 47/58] ath9k: implement IO serialization
Date: Wed, 29 Apr 2009 15:07:46 -0700 [thread overview]
Message-ID: <20090429220929.995031553@mini.kroah.org> (raw)
In-Reply-To: <20090429221657.GA11765@kroah.com>
2.6.27-stable review patch. If anyone has any objections, please let us know.
------------------
From: Luis R. Rodriguez <lrodriguez@atheros.com>
This is a port of:
commit SHA1 6158425be398936af1fd04451f78ffad01529cb0
for 2.6.27
All 802.11n PCI devices (Cardbus, PCI, mini-PCI) require
serialization of IO when on non-uniprocessor systems. PCI
express devices not not require this.
This should fix our only last standing open ath9k kernel.org
bugzilla bug report:
http://bugzilla.kernel.org/show_bug.cgi?id=12110
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
---
drivers/net/wireless/ath9k/ath9k.h | 4 ++--
drivers/net/wireless/ath9k/core.c | 1 +
drivers/net/wireless/ath9k/core.h | 33 +++++++++++++++++++++++++++++++++
drivers/net/wireless/ath9k/hw.c | 19 +++++++++++++++++++
4 files changed, 55 insertions(+), 2 deletions(-)
--- a/drivers/net/wireless/ath9k/ath9k.h
+++ b/drivers/net/wireless/ath9k/ath9k.h
@@ -591,8 +591,8 @@ struct ath9k_country_entry {
u8 iso[3];
};
-#define REG_WRITE(_ah, _reg, _val) iowrite32(_val, _ah->ah_sh + _reg)
-#define REG_READ(_ah, _reg) ioread32(_ah->ah_sh + _reg)
+#define REG_WRITE(_ah, _reg, _val) ath9k_iowrite32((_ah), (_reg), (_val))
+#define REG_READ(_ah, _reg) ath9k_ioread32((_ah), (_reg))
#define SM(_v, _f) (((_v) << _f##_S) & _f)
#define MS(_v, _f) (((_v) & _f) >> _f##_S)
--- a/drivers/net/wireless/ath9k/core.c
+++ b/drivers/net/wireless/ath9k/core.c
@@ -1120,6 +1120,7 @@ int ath_init(u16 devid, struct ath_softc
sc->sc_cachelsz = csz << 2; /* convert to bytes */
spin_lock_init(&sc->sc_resetlock);
+ spin_lock_init(&sc->sc_serial_rw);
ah = ath9k_hw_attach(devid, sc, sc->mem, &status);
if (ah == NULL) {
--- a/drivers/net/wireless/ath9k/core.h
+++ b/drivers/net/wireless/ath9k/core.h
@@ -1022,6 +1022,7 @@ struct ath_softc {
spinlock_t sc_rxbuflock;
spinlock_t sc_txbuflock;
spinlock_t sc_resetlock;
+ spinlock_t sc_serial_rw;
spinlock_t node_lock;
};
@@ -1069,4 +1070,36 @@ void ath_skb_unmap_single(struct ath_sof
void ath_mcast_merge(struct ath_softc *sc, u32 mfilt[2]);
enum ath9k_ht_macmode ath_cwm_macmode(struct ath_softc *sc);
+/*
+ * Read and write, they both share the same lock. We do this to serialize
+ * reads and writes on Atheros 802.11n PCI devices only. This is required
+ * as the FIFO on these devices can only accept sanely 2 requests. After
+ * that the device goes bananas. Serializing the reads/writes prevents this
+ * from happening.
+ */
+
+static inline void ath9k_iowrite32(struct ath_hal *ah, u32 reg_offset, u32 val)
+{
+ if (ah->ah_config.serialize_regmode == SER_REG_MODE_ON) {
+ unsigned long flags;
+ spin_lock_irqsave(&ah->ah_sc->sc_serial_rw, flags);
+ iowrite32(val, ah->ah_sc->mem + reg_offset);
+ spin_unlock_irqrestore(&ah->ah_sc->sc_serial_rw, flags);
+ } else
+ iowrite32(val, ah->ah_sc->mem + reg_offset);
+}
+
+static inline unsigned int ath9k_ioread32(struct ath_hal *ah, u32 reg_offset)
+{
+ u32 val;
+ if (ah->ah_config.serialize_regmode == SER_REG_MODE_ON) {
+ unsigned long flags;
+ spin_lock_irqsave(&ah->ah_sc->sc_serial_rw, flags);
+ val = ioread32(ah->ah_sc->mem + reg_offset);
+ spin_unlock_irqrestore(&ah->ah_sc->sc_serial_rw, flags);
+ } else
+ val = ioread32(ah->ah_sc->mem + reg_offset);
+ return val;
+}
+
#endif /* CORE_H */
--- a/drivers/net/wireless/ath9k/hw.c
+++ b/drivers/net/wireless/ath9k/hw.c
@@ -369,6 +369,25 @@ static void ath9k_hw_set_defaults(struct
}
ah->ah_config.intr_mitigation = 0;
+
+ /*
+ * We need this for PCI devices only (Cardbus, PCI, miniPCI)
+ * _and_ if on non-uniprocessor systems (Multiprocessor/HT).
+ * This means we use it for all AR5416 devices, and the few
+ * minor PCI AR9280 devices out there.
+ *
+ * Serialization is required because these devices do not handle
+ * well the case of two concurrent reads/writes due to the latency
+ * involved. During one read/write another read/write can be issued
+ * on another CPU while the previous read/write may still be working
+ * on our hardware, if we hit this case the hardware poops in a loop.
+ * We prevent this by serializing reads and writes.
+ *
+ * This issue is not present on PCI-Express devices or pre-AR5416
+ * devices (legacy, 802.11abg).
+ */
+ if (num_possible_cpus() > 1)
+ ah->ah_config.serialize_regmode = SER_REG_MODE_AUTO;
}
static inline void ath9k_hw_override_ini(struct ath_hal *ah,
WARNING: multiple messages have this Message-ID (diff)
From: Greg KH <gregkh@suse.de>
To: linux-kernel@vger.kernel.org, stable@kernel.org
Cc: Justin Forbes <jmforbes@linuxtx.org>,
Zwane Mwaikambo <zwane@arm.linux.org.uk>,
"Theodore Ts'o" <tytso@mit.edu>,
Randy Dunlap <rdunlap@xenotime.net>,
Dave Jones <davej@redhat.com>,
Chuck Wolber <chuckw@quantumlinux.com>,
Chris Wedgwood <reviews@ml.cw.f00f.org>,
Michael Krufky <mkrufky@linuxtv.org>,
Chuck Ebbert <cebbert@redhat.com>,
Domenico Andreoli <cavokz@gmail.com>, Willy Tarreau <w@1wt.eu>,
Rodrigo Rubira Branco <rbranco@la.checkpoint.com>,
Jake Edge <jake@lwn.net>, Eugene Teo <eteo@redhat.com>,
torvalds@linux-foundation.org, akpm@linux-foundation.org,
alan@lxorguk.ukuu.org.uk,
"Luis R. Rodriguez" <lrodriguez@atheros.com>,
ath9k-devel@venema.h4ckr.net, linux-wireless@vger.kernel.org
Subject: [patch 47/58] ath9k: implement IO serialization
Date: Wed, 29 Apr 2009 15:07:46 -0700 [thread overview]
Message-ID: <20090429220929.995031553@mini.kroah.org> (raw)
In-Reply-To: <20090429221657.GA11765@kroah.com>
[-- Attachment #1: ath9k-implement-io-serialization.patch --]
[-- Type: text/plain, Size: 4491 bytes --]
2.6.27-stable review patch. If anyone has any objections, please let us know.
------------------
From: Luis R. Rodriguez <lrodriguez@atheros.com>
This is a port of:
commit SHA1 6158425be398936af1fd04451f78ffad01529cb0
for 2.6.27
All 802.11n PCI devices (Cardbus, PCI, mini-PCI) require
serialization of IO when on non-uniprocessor systems. PCI
express devices not not require this.
This should fix our only last standing open ath9k kernel.org
bugzilla bug report:
http://bugzilla.kernel.org/show_bug.cgi?id=12110
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
---
drivers/net/wireless/ath9k/ath9k.h | 4 ++--
drivers/net/wireless/ath9k/core.c | 1 +
drivers/net/wireless/ath9k/core.h | 33 +++++++++++++++++++++++++++++++++
drivers/net/wireless/ath9k/hw.c | 19 +++++++++++++++++++
4 files changed, 55 insertions(+), 2 deletions(-)
--- a/drivers/net/wireless/ath9k/ath9k.h
+++ b/drivers/net/wireless/ath9k/ath9k.h
@@ -591,8 +591,8 @@ struct ath9k_country_entry {
u8 iso[3];
};
-#define REG_WRITE(_ah, _reg, _val) iowrite32(_val, _ah->ah_sh + _reg)
-#define REG_READ(_ah, _reg) ioread32(_ah->ah_sh + _reg)
+#define REG_WRITE(_ah, _reg, _val) ath9k_iowrite32((_ah), (_reg), (_val))
+#define REG_READ(_ah, _reg) ath9k_ioread32((_ah), (_reg))
#define SM(_v, _f) (((_v) << _f##_S) & _f)
#define MS(_v, _f) (((_v) & _f) >> _f##_S)
--- a/drivers/net/wireless/ath9k/core.c
+++ b/drivers/net/wireless/ath9k/core.c
@@ -1120,6 +1120,7 @@ int ath_init(u16 devid, struct ath_softc
sc->sc_cachelsz = csz << 2; /* convert to bytes */
spin_lock_init(&sc->sc_resetlock);
+ spin_lock_init(&sc->sc_serial_rw);
ah = ath9k_hw_attach(devid, sc, sc->mem, &status);
if (ah == NULL) {
--- a/drivers/net/wireless/ath9k/core.h
+++ b/drivers/net/wireless/ath9k/core.h
@@ -1022,6 +1022,7 @@ struct ath_softc {
spinlock_t sc_rxbuflock;
spinlock_t sc_txbuflock;
spinlock_t sc_resetlock;
+ spinlock_t sc_serial_rw;
spinlock_t node_lock;
};
@@ -1069,4 +1070,36 @@ void ath_skb_unmap_single(struct ath_sof
void ath_mcast_merge(struct ath_softc *sc, u32 mfilt[2]);
enum ath9k_ht_macmode ath_cwm_macmode(struct ath_softc *sc);
+/*
+ * Read and write, they both share the same lock. We do this to serialize
+ * reads and writes on Atheros 802.11n PCI devices only. This is required
+ * as the FIFO on these devices can only accept sanely 2 requests. After
+ * that the device goes bananas. Serializing the reads/writes prevents this
+ * from happening.
+ */
+
+static inline void ath9k_iowrite32(struct ath_hal *ah, u32 reg_offset, u32 val)
+{
+ if (ah->ah_config.serialize_regmode == SER_REG_MODE_ON) {
+ unsigned long flags;
+ spin_lock_irqsave(&ah->ah_sc->sc_serial_rw, flags);
+ iowrite32(val, ah->ah_sc->mem + reg_offset);
+ spin_unlock_irqrestore(&ah->ah_sc->sc_serial_rw, flags);
+ } else
+ iowrite32(val, ah->ah_sc->mem + reg_offset);
+}
+
+static inline unsigned int ath9k_ioread32(struct ath_hal *ah, u32 reg_offset)
+{
+ u32 val;
+ if (ah->ah_config.serialize_regmode == SER_REG_MODE_ON) {
+ unsigned long flags;
+ spin_lock_irqsave(&ah->ah_sc->sc_serial_rw, flags);
+ val = ioread32(ah->ah_sc->mem + reg_offset);
+ spin_unlock_irqrestore(&ah->ah_sc->sc_serial_rw, flags);
+ } else
+ val = ioread32(ah->ah_sc->mem + reg_offset);
+ return val;
+}
+
#endif /* CORE_H */
--- a/drivers/net/wireless/ath9k/hw.c
+++ b/drivers/net/wireless/ath9k/hw.c
@@ -369,6 +369,25 @@ static void ath9k_hw_set_defaults(struct
}
ah->ah_config.intr_mitigation = 0;
+
+ /*
+ * We need this for PCI devices only (Cardbus, PCI, miniPCI)
+ * _and_ if on non-uniprocessor systems (Multiprocessor/HT).
+ * This means we use it for all AR5416 devices, and the few
+ * minor PCI AR9280 devices out there.
+ *
+ * Serialization is required because these devices do not handle
+ * well the case of two concurrent reads/writes due to the latency
+ * involved. During one read/write another read/write can be issued
+ * on another CPU while the previous read/write may still be working
+ * on our hardware, if we hit this case the hardware poops in a loop.
+ * We prevent this by serializing reads and writes.
+ *
+ * This issue is not present on PCI-Express devices or pre-AR5416
+ * devices (legacy, 802.11abg).
+ */
+ if (num_possible_cpus() > 1)
+ ah->ah_config.serialize_regmode = SER_REG_MODE_AUTO;
}
static inline void ath9k_hw_override_ini(struct ath_hal *ah,
next prev parent reply other threads:[~2009-04-29 22:07 UTC|newest]
Thread overview: 65+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20090429220659.339950874@mini.kroah.org>
[not found] ` <20090429221657.GA11765-U8xfFu+wG4EAvxtiuMwx3w@public.gmane.org>
2009-04-29 22:07 ` [patch 29/58] sched: do not count frozen tasks toward load Greg KH
2009-04-29 22:07 ` Greg KH
2009-04-29 22:16 ` [patch 00/58] 2.6.27-stable review Greg KH
2009-04-29 22:07 ` [patch 01/58] USB: EHCI: add software retry for transaction errors Greg KH
2009-04-29 22:07 ` [patch 02/58] USB: fix USB_STORAGE_CYPRESS_ATACB Greg KH
2009-04-29 22:07 ` [patch 03/58] USB: usb-storage: increase max_sectors for tape drives Greg KH
2009-04-29 22:07 ` [patch 04/58] USB: gadget: fix rndis regression Greg KH
2009-04-29 22:07 ` [patch 05/58] cifs: fix buffer format byte on NT Rename/hardlink Greg KH
2009-04-29 22:07 ` [patch 06/58] b43: fix b43_plcp_get_bitrate_idx_ofdm return type Greg KH
2009-04-29 22:07 ` [patch 07/58] CIFS: Fix memory overwrite when saving nativeFileSystem field during mount Greg KH
2009-04-29 22:07 ` [patch 08/58] Add a missing unlock_kernel() in raw_open() Greg KH
2009-04-29 22:07 ` [patch 09/58] x86, PAT, PCI: Change vma prot in pci_mmap to reflect inherited prot Greg KH
2009-04-29 22:07 ` [patch 10/58] x86: mtrr: dont modify RdDram/WrDram bits of fixed MTRRs Greg KH
2009-04-29 22:07 ` [patch 11/58] bas_gigaset: correctly allocate USB interrupt transfer buffer Greg KH
2009-04-29 22:07 ` [patch 12/58] bonding: Fix updating of speed/duplex changes Greg KH
2009-04-29 22:07 ` [patch 13/58] bridge: bad error handling when adding invalid ether address Greg KH
2009-04-29 22:07 ` [patch 14/58] ipv6: dont use tw net when accounting for recycled tw Greg KH
2009-04-29 22:07 ` [patch 15/58] ipv6: Plug sk_buff leak in ipv6_rcv (net/ipv6/ip6_input.c) Greg KH
2009-04-29 22:07 ` [patch 16/58] netfilter: nf_conntrack_tcp: fix unaligned memory access in tcp_sack Greg KH
2009-04-29 22:07 ` [patch 17/58] net: fix sctp breakage Greg KH
2009-04-29 22:07 ` [patch 18/58] security/smack: fix oops when setting a size 0 SMACK64 xattr Greg KH
2009-04-29 22:07 ` [patch 19/58] x86, setup: mark %esi as clobbered in E820 BIOS call Greg KH
2009-04-29 22:07 ` [patch 20/58] mm: do_xip_mapping_read: fix length calculation Greg KH
2009-04-29 22:07 ` [patch 21/58] vfs: skip I_CLEAR state inodes Greg KH
2009-04-29 22:07 ` [patch 22/58] af_rose/x25: Sanity check the maximum user frame size Greg KH
2009-04-29 22:07 ` [patch 23/58] net/netrom: Fix socket locking Greg KH
2009-04-29 22:07 ` [patch 24/58] netfilter: {ip, ip6, arp}_tables: fix incorrect loop detection Greg KH
2009-04-29 22:07 ` [patch 25/58] splice: fix deadlock in splicing to file Greg KH
2009-04-29 22:07 ` [patch 26/58] ALSA: hda - add missing comma in ad1884_slave_vols Greg KH
2009-04-29 22:07 ` [patch 27/58] SCSI: libiscsi: fix iscsi pool error path Greg KH
2009-04-29 22:07 ` [patch 28/58] SCSI: libiscsi: fix iscsi pool error path again Greg KH
2009-04-29 22:07 ` [patch 29/58] sched: do not count frozen tasks toward load Greg KH
2009-04-29 22:07 ` [patch 30/58] add some long-missing capabilities to fs_mask Greg KH
2009-04-29 22:07 ` [patch 31/58] powerpc: Fix data-corrupting bug in __futex_atomic_op Greg KH
2009-04-29 22:07 ` [patch 32/58] hpt366: fix HPT370 DMA timeouts Greg KH
2009-04-29 22:07 ` [patch 33/58] pata_hpt37x: " Greg KH
2009-04-29 22:07 ` [patch 34/58] usb gadget: fix ethernet link reports to ethtool Greg KH
2009-04-29 22:07 ` [patch 35/58] USB: ftdi_sio: add vendor/project id for JETI specbos 1201 spectrometer Greg KH
2009-04-29 22:07 ` [patch 36/58] USB: fix oops in cdc-wdm in case of malformed descriptors Greg KH
2009-04-29 22:07 ` [patch 37/58] USB: usb-storage: augment unusual_devs entry for Simple Tech/Datafab Greg KH
2009-04-29 22:07 ` [patch 38/58] agp: zero pages before sending to userspace Greg KH
2009-04-29 22:07 ` [patch 39/58] hugetlbfs: return negative error code for bad mount option Greg KH
2009-04-29 22:07 ` [patch 40/58] kprobes: Fix locking imbalance in kretprobes Greg KH
2009-04-29 22:07 ` [patch 41/58] block: revert part of 18ce3751ccd488c78d3827e9f6bf54e6322676fb Greg KH
2009-04-29 22:07 ` [patch 42/58] r8169: Dont update statistics counters when interface is down Greg KH
2009-04-29 22:07 ` [patch 43/58] r8169: use hardware auto-padding Greg KH
2009-04-29 22:07 ` [patch 44/58] r8169: reset IntrStatus after chip reset Greg KH
2009-04-29 22:07 ` [patch 45/58] mm: check for no mmaps in exit_mmap() Greg KH
2009-04-29 22:07 ` [patch 46/58] powerpc: Sanitize stack pointer in signal handling code Greg KH
2009-04-29 22:07 ` Greg KH [this message]
2009-04-29 22:07 ` [patch 47/58] ath9k: implement IO serialization Greg KH
2009-04-29 22:07 ` Greg KH
2009-04-29 22:07 ` [ath9k-devel] [patch 48/58] ath9k: AR9280 PCI devices must serialize IO as well Greg KH
2009-04-29 22:07 ` Greg KH
2009-04-29 22:07 ` Greg KH
2009-04-29 22:07 ` [patch 49/58] b44: Use kernel DMA addresses for the kernel DMA API Greg KH
2009-04-29 22:07 ` [patch 50/58] crypto: ixp4xx - Fix handling of chained sg buffers Greg KH
2009-04-29 22:07 ` [patch 51/58] exit_notify: kill the wrong capable(CAP_KILL) check (CVE-2009-1337) Greg KH
2009-04-29 22:07 ` [patch 52/58] fix ptrace slowness Greg KH
2009-04-29 22:07 ` [patch 53/58] fs core fixes Greg KH
2009-04-29 22:07 ` [patch 54/58] PCI: fix incorrect mask of PM No_Soft_Reset bit Greg KH
2009-04-29 22:07 ` [patch 55/58] thinkpad-acpi: fix LED blinking through timer trigger Greg KH
2009-04-29 22:07 ` [patch 56/58] ACPI: EC: Add some basic check for ECDT data Greg KH
2009-04-29 22:07 ` [patch 57/58] ACPI: EC: fix compilation warning Greg KH
2009-04-29 22:07 ` [patch 58/58] unreached code in selinux_ip_postroute_iptables_compat() (CVE-2009-1184) Greg KH
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=20090429220929.995031553@mini.kroah.org \
--to=gregkh@suse.de \
--cc=ath9k-devel@lists.ath9k.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.