headers
Richard Zhao | 1 Jan 2012 04:27
Favicon

[PATCH] dma/imx-sdma: add wmb before enable/run a channel

dma_alloc_coherent memory may be bufferable. We need to add
nececcary memory barrier.

Signed-off-by: Richard Zhao <richard.zhao <at> linaro.org>
---
 drivers/dma/imx-sdma.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
index 2ebb2bc..8cd9492 100644
--- a/drivers/dma/imx-sdma.c
+++ b/drivers/dma/imx-sdma.c
 <at>  <at>  -403,6 +403,7  <at>  <at>  static int sdma_run_channel(struct sdma_channel *sdmac)

 	init_completion(&sdmac->done);

+	wmb();
 	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);

 	ret = wait_for_completion_timeout(&sdmac->done, HZ);
 <at>  <at>  -809,6 +810,7  <at>  <at>  out:

 static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
 {
+	wmb();
 	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
 }

--

-- 
1.7.4.1
(Continue reading)

Permalink | Reply | 2 comments |
headers
Richard Zhao | 1 Jan 2012 04:30
Favicon

[PATCH] dma/imx-sdma: save irq flags when use spin_lock in sdma_tx_submit

From: Haitao Zhang <minipanda <at> linuxrobot.org>

mx53_loco: fix deadlock report from sdma_tx_submit() during boot

BugLink: http://bugs.launchpad.net/bugs/878701

Adjust to use spin_lock_irqsave()/spin_unlock_irqresotre(), so to
make it safe when called from interrupt context.

Signed-off-by: Haitao Zhang <haitao.zhang <at> linaro.org>
Signed-off-by: Eric Miao <eric.miao <at> linaro.org>
Signed-off-by: Richard Zhao <richard.zhao <at> linaro.org>
---
 drivers/dma/imx-sdma.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
index f993955..2b7a88b 100644
--- a/drivers/dma/imx-sdma.c
+++ b/drivers/dma/imx-sdma.c
 <at>  <at>  -832,17 +832,18  <at>  <at>  static struct sdma_channel *to_sdma_chan(struct dma_chan *chan)

 static dma_cookie_t sdma_tx_submit(struct dma_async_tx_descriptor *tx)
 {
+	unsigned long flags;
 	struct sdma_channel *sdmac = to_sdma_chan(tx->chan);
 	struct sdma_engine *sdma = sdmac->sdma;
 	dma_cookie_t cookie;

-	spin_lock_irq(&sdmac->lock);
(Continue reading)

Permalink | Reply | 2 comments |
headers
Steinar H. Gunderson | 1 Jan 2012 03:47
Picon

3.2.0-rc5 oops in ext4

Hi,

During relatively heavy write load, my 3.2.0-rc5 kernel oopsed (after sixteen
days of uptime, so this is not easily reproducible):

Jan  1 03:21:41 pannekake kernel: [1439120.482015] PGD 0 
Jan  1 03:21:41 pannekake kernel: [1439120.488213] CPU 4 
Jan  1 03:21:41 pannekake kernel: [1439120.490228] Modules linked in: ipt_REJECT iptable_filter
ip_tables sha256_generic cryptd aes_x86_64 aes_generic af_packet microcode bridge stp llc ext4 jbd2
crc16 ext2 fuse dm_snapshot dm_crypt kvm_intel kvm coretemp w83627ehf hwmon_vid ip_gre gre sg
ide_generic ide_gd_mod ide_cd_mod cdrom i2c_i801 psmouse i2c_core serio_raw pcspkr evdev ext3 jbd
mbcache dm_mod raid456 async_pq async_xor xor async_memcpy async_raid6_recov raid6_pq async_tx
raid1 md_mod usbhid ide_pci_generic ide_core uhci_hcd ata_piix e1000e ehci_hcd sd_mod unix [last
unloaded: scsi_wait_scan]
Jan  1 03:21:41 pannekake kernel: [1439120.543854] 
Jan  1 03:21:41 pannekake kernel: [1439120.545732] Pid: 26600, comm: apache2 Not tainted 3.2.0-rc5 #2
Supermicro X8DTL/X8DTL
Jan  1 03:21:41 pannekake kernel: [1439120.554226] RIP: 0010:[<ffffffff81355161>] 
[<ffffffff81355161>] _raw_spin_lock+0x9/0x20
Jan  1 03:21:41 pannekake kernel: [1439120.562935] RSP: 0018:ffff88002a6259e8  EFLAGS: 00010202
Jan  1 03:21:41 pannekake kernel: [1439120.568660] RAX: 0000000000000100 RBX: ffffea0000bdb240 RCX: 0000000000000000
Jan  1 03:21:41 pannekake kernel: [1439120.576458] RDX: ffff880033231f70 RSI: ffff88002f6c9000 RDI: 0000000000000098
Jan  1 03:21:41 pannekake kernel: [1439120.584190] RBP: ffff88002a6259e8 R08: 0000000000000001 R09: 0000000000000001
Jan  1 03:21:41 pannekake kernel: [1439120.591952] R10: ffff880623d0fe28 R11: 0000000000000000 R12: 0000000000000000
Jan  1 03:21:41 pannekake kernel: [1439120.599697] R13: ffff880033231f70 R14: 0000000000000001 R15: 000000000003300c
Jan  1 03:21:41 pannekake kernel: [1439120.607482] FS:  00007f43323c0740(0000)
GS:ffff880627280000(0000) knlGS:0000000000000000
Jan  1 03:21:41 pannekake kernel: [1439120.616190] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
Jan  1 03:21:41 pannekake kernel: [1439120.622311] CR2: 0000000000000098 CR3: 00000001289cd000 CR4: 00000000000006e0
Jan  1 03:21:41 pannekake kernel: [1439120.630095] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
(Continue reading)

Permalink | Reply | 3 comments |
headers
Larry Finger | 1 Jan 2012 04:54
Favicon

[PATCH] rtlwifi: rtl8192ce: rtl8192cu: Delay firmware file read until needed to fix at drivers/base/firmware_class.c:537

A recent LKML thread (http://lkml.indiana.edu/hypermail/linux/kernel/1112.3/00965.html)
discusses the the WARNING listed below, which occurs after a suspend/resume cycle. For
rtl8192cu, the problem was fixed by delaying the firmware loading until the interface is
brought up. By this time, the userspace support is available.

Dec 30 18:58:25 larrylap kernel: [186060.505125] ------------[ cut here ]------------
Dec 30 18:58:25 larrylap kernel: [186060.505133] WARNING: at drivers/base/firmware_class.c:537 _request_firmware+0x3f6/0x420()
Dec 30 18:58:25 larrylap kernel: [186060.505135] Hardware name: HP Pavilion dv2700 Notebook PC
Dec 30 18:58:25 larrylap kernel: [186060.505137] Modules linked in: rtl8192cu rtl8192c_common
rtl8192se rtlwifi mac80211 cfg80211 rfkill xt_state iptable_filter ipt_MASQUERADE iptable_nat
nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables fuse aes_x86_64
aes_generic nfs lockd auth_rpcgss nfs_acl sunrpc vboxpci(O) vboxnetadp(O) af_packet vboxnetflt(O)
vboxdrv(O) cpufreq_conservative cpufreq_userspace cpufreq_powersave powernow_k8 mperf
ide_cd_mod cdrom ext3 jbd snd_hda_codec_conexant arc4 ide_pci_generic snd_hda_intel snd_hda_codec
snd_pcm amd74xx snd_timer snd soundcore snd_page_alloc k8temp i2c_nforce2 ide_core battery ac joydev
button forcedeth serio_raw i2c_core hwmon sg video ipv6 autofs4 ext4 mbcache jbd2 crc16 sd_mod ohci_hcd
ahci libahci libata scsi_mod ehci_hcd
  usbcore usb_common fan processor thermal [last unloaded: rtlwifi]
Dec 30 18:58:25 larrylap kernel: [186060.505206] Pid: 20502, comm: s2ram Tainted: G        W  O 3.2.0-rc7-wl+ #161
Dec 30 18:58:25 larrylap kernel: [186060.505209] Call Trace:
Dec 30 18:58:25 larrylap kernel: [186060.505214]  [<ffffffff81048aea>] warn_slowpath_common+0x7a/0xb0
Dec 30 18:58:25 larrylap kernel: [186060.505218]  [<ffffffff81048b35>] warn_slowpath_null+0x15/0x20
Dec 30 18:58:25 larrylap kernel: [186060.505223]  [<ffffffff81288396>] _request_firmware+0x3f6/0x420
Dec 30 18:58:25 larrylap kernel: [186060.505226]  [<ffffffff812883d1>] request_firmware+0x11/0x20
Dec 30 18:58:25 larrylap kernel: [186060.505234]  [<ffffffffa0400cfe>]
rtl92cu_init_sw_vars+0x6e/0x210 [rtl8192cu]
Dec 30 18:58:25 larrylap kernel: [186060.505243]  [<ffffffffa01189a6>] rtl_usb_probe+0x1cd/0x7b6 [rtlwifi]
Dec 30 18:58:25 larrylap kernel: [186060.505263]  [<ffffffffa002a595>]
usb_probe_interface+0xb5/0x160 [usbcore]
Dec 30 18:58:25 larrylap kernel: [186060.505268]  [<ffffffff8127f2a3>] driver_probe_device+0x73/0x190
(Continue reading)

Permalink | Reply | 1 comment |
headers
Ted Ts'o | 1 Jan 2012 05:23
Picon
Picon
Favicon
Gravatar

Re: 3.2.0-rc5 oops in ext4

On Sun, Jan 01, 2012 at 03:47:58AM +0100, Steinar H. Gunderson wrote:
> Hi,
> 
> During relatively heavy write load, my 3.2.0-rc5 kernel oopsed (after sixteen
> days of uptime, so this is not easily reproducible):

This wasn't the whole kernel oops message.  Can you get any lines
earlier than this?  (Preferably beginning wit
 "-----[ cut here ]-----"?)

     	    			       - Ted
Permalink | Reply | 2 comments |
headers
Hugh Dickins | 1 Jan 2012 07:15
Picon
Favicon

Re: 3.2.0-rc5 oops in ext4

On Sat, 31 Dec 2011, Ted Ts'o wrote:
> On Sun, Jan 01, 2012 at 03:47:58AM +0100, Steinar H. Gunderson wrote:
> > Hi,
> > 
> > During relatively heavy write load, my 3.2.0-rc5 kernel oopsed (after sixteen
> > days of uptime, so this is not easily reproducible):
> 
> This wasn't the whole kernel oops message.  Can you get any lines
> earlier than this?  (Preferably beginning wit
>  "-----[ cut here ]-----"?)

Nah, not necessary:

RIP: 0010:[<ffffffff81355161>]  [<ffffffff81355161>] _raw_spin_lock+0x9/0x20
[<ffffffff8110d19e>] create_empty_buffers+0x43/0xae
[<ffffffffa030c003>] ext4_discard_partial_page_buffers_no_lock+0x9d/0x3ee [ext4]
[<ffffffffa030c60f>] ext4_da_write_end+0x231/0x2df [ext4]

is one of the symptoms I was seeing - fixed in -rc6 by Yongqiang's
093e6e3666f47d29763a235b404c84ee47ba8bb0
ext4: correctly handle pages w/o buffers in ext4_discard_partial_buffers()

Hugh
Permalink | Reply | 0 comments |
headers
Vasco Dias | 1 Jan 2012 07:20
Picon

Problem with efibootmgr on asus 1215b

Hi,

the problem is that after booting the system from uefi (with the uefi
shell) and after loading
the module efivars i can't create a new boot entry because the system
just locks completely.

Laptop is using the latest stable kernel and latest bios version
available. The problem occurred with
previous bios version and previous with kernels too.

After the system locking it becomes completely unresponsive and the
only way too shutdown is to cut the power,
including removing the battery.

I'm reporting here on the list after asking about it on the arch linux
forums ( https://bbs.archlinux.org/viewtopic.php?pid=1024451 ).

I'm a noob with mailing lists and all these kernel problems, so please
be kind :P.

Thanks in advance.

[>>] efibootmgr -v
Timeout: 0 seconds
BootOrder: 0001
Boot0001  Hard Drive    BIOS(2,0,00)
Attachment (kernel.log): text/x-log, 15 KiB
Attachment (lastest_dmesg): application/octet-stream, 64 KiB
(Continue reading)

Permalink | Reply | 9 comments |
headers
Vasco Dias | 1 Jan 2012 07:27
Picon

Re: Problem with efibootmgr on asus 1215b

Sorry for the attachments.

relevant kernel.log:
Jan  1 04:30:12 localhost kernel: [ 3103.586496] INFO:
rcu_preempt_state detected stalls on CPUs/tasks: { 1} (detected by 0,
t=450258 jiffies)
Jan  1 04:30:12 localhost kernel: [ 3103.586519] sending NMI to all CPUs:
Jan  1 04:30:12 localhost kernel: [ 2420.190015] NMI backtrace for cpu 1
Jan  1 04:30:12 localhost kernel: [ 2420.190015] CPU 1
Jan  1 04:30:12 localhost kernel: [ 2420.190015] Modules linked in:
efivars ipv6 fuse uvcvideo videodev media joydev v4l2_compat_ioctl32
bcma arc4 snd_hda_codec_realtek snd_hda_codec_hdmi i2c_piix4
brcmsmac(C) snd_hda_intel mac80211 snd_hda_codec snd_hwdep snd_pcm
snd_timer brcmutil(C) cfg80211 serio_raw snd eeepc_wmi asus_wmi
sparse_keymap ohci_hcd xhci_hcd ehci_hcd pcspkr psmouse evdev usbcore
soundcore snd_page_alloc sp5100_tco k10temp rfkill pci_hotplug thermal
wmi video ac battery button cpufreq_ondemand powernow_k8 freq_table
processor mperf ext4 mbcache jbd2 crc16 sd_mod ahci libahci libata
scsi_mod radeon ttm drm_kms_helper drm i2c_algo_bit i2c_core
Jan  1 04:30:12 localhost kernel: [ 2420.190015]
Jan  1 04:30:12 localhost kernel: [ 2420.190015] Pid: 1426, comm:
efibootmgr Tainted: G      D WC  3.1.6-1-ARCH #1 ASUSTeK Computer INC.
1215B/1215B
Jan  1 04:30:12 localhost kernel: [ 2420.190015] RIP:
0010:[<ffffffff81409b69>]  [<ffffffff81409b69>]
_raw_spin_lock+0x29/0x30
Jan  1 04:30:12 localhost kernel: [ 2420.190015] RSP:
0018:ffff88012c51d558  EFLAGS: 00000097
Jan  1 04:30:12 localhost kernel: [ 2420.190015] RAX: 0000000000002625
RBX: 0000000000000000 RCX: 0000000000000009
(Continue reading)

Permalink | Reply | 0 comments |
headers
Uwe Kleine-König | 1 Jan 2012 08:14
Picon
Favicon
Gravatar

Re: [PATCH] dma/imx-sdma: add wmb before enable/run a channel

Hello Richard,

On Sun, Jan 01, 2012 at 11:27:40AM +0800, Richard Zhao wrote:
> dma_alloc_coherent memory may be bufferable. We need to add
> nececcary memory barrier.
necessary

> 
> Signed-off-by: Richard Zhao <richard.zhao <at> linaro.org>
> ---
>  drivers/dma/imx-sdma.c |    2 ++
>  1 files changed, 2 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
> index 2ebb2bc..8cd9492 100644
> --- a/drivers/dma/imx-sdma.c
> +++ b/drivers/dma/imx-sdma.c
>  <at>  <at>  -403,6 +403,7  <at>  <at>  static int sdma_run_channel(struct sdma_channel *sdmac)
>  
>  	init_completion(&sdmac->done);
>  
> +	wmb();
>  	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
>  
>  	ret = wait_for_completion_timeout(&sdmac->done, HZ);
>  <at>  <at>  -809,6 +810,7  <at>  <at>  out:
>  
>  static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
>  {
> +	wmb();
(Continue reading)

Permalink | Reply | 1 comment |
headers
Roland Dreier | 1 Jan 2012 08:16
Gravatar

Re: [PATCH] btree: Export btree_get_prev() so modules can use btree_for_each

On Sat, Dec 31, 2011 at 7:22 PM, Steve Hodgson <steve <at> purestorage.com> wrote:
> Signed-off-by: Steve Hodgson
>
> Or you would prefer the patch is resent with my Sign-off in the commit
> message?

A properly formed S-o-b line has the email address too... but in any case I
don't think it should be necessary in this case.  First of all the patch is a
pretty trivial one-liner... but in any case as I understand things I should be
able to submit patches from our company's internal tree and both preserve
authorship (ie "From: Steve") but also be the first person sending it out.

Looking at "Developer's Certificate of Origin 1.1" in SubmittingPatches, I
guess I am certifying (b), ie

        (b) The contribution is based upon previous work that, to the best
            of my knowledge, is covered under an appropriate open source
            license and I have the right under that license to submit that
            work with modifications, whether created in whole or in part
            by me, under the same open source license (unless I am
            permitted to submit under a different license), as indicated
            in the file; or

(Since I actually did write the changelog for the patch)

 - R.
Permalink | Reply | 0 comments |

Gmane