| Age | Commit message (Collapse) | Author |
|---|
|
The bulk of ATA host state machine is implemented by
ata_sff_hsm_move(). The function is called from either the interrupt
handler or, if polling, a work item. Unlike from the interrupt path,
the polling path calls the function without holding the host lock and
ata_sff_hsm_move() selectively grabs the lock.
This is completely broken. If an IRQ triggers while polling is in
progress, the two can easily race and end up accessing the hardware
and updating state machine state at the same time. This can put the
state machine in an illegal state and lead to a crash like the
following.
kernel BUG at drivers/ata/libata-sff.c:1302!
invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN
Modules linked in:
CPU: 1 PID: 10679 Comm: syz-executor Not tainted 4.5.0-rc1+ #300
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
task: ffff88002bd00000 ti: ffff88002e048000 task.ti: ffff88002e048000
RIP: 0010:[<ffffffff83a83409>] [<ffffffff83a83409>] ata_sff_hsm_move+0x619/0x1c60
...
Call Trace:
<IRQ>
[<ffffffff83a84c31>] __ata_sff_port_intr+0x1e1/0x3a0 drivers/ata/libata-sff.c:1584
[<ffffffff83a85611>] ata_bmdma_port_intr+0x71/0x400 drivers/ata/libata-sff.c:2877
[< inline >] __ata_sff_interrupt drivers/ata/libata-sff.c:1629
[<ffffffff83a85bf3>] ata_bmdma_interrupt+0x253/0x580 drivers/ata/libata-sff.c:2902
[<ffffffff81479f98>] handle_irq_event_percpu+0x108/0x7e0 kernel/irq/handle.c:157
[<ffffffff8147a717>] handle_irq_event+0xa7/0x140 kernel/irq/handle.c:205
[<ffffffff81484573>] handle_edge_irq+0x1e3/0x8d0 kernel/irq/chip.c:623
[< inline >] generic_handle_irq_desc include/linux/irqdesc.h:146
[<ffffffff811a92bc>] handle_irq+0x10c/0x2a0 arch/x86/kernel/irq_64.c:78
[<ffffffff811a7e4d>] do_IRQ+0x7d/0x1a0 arch/x86/kernel/irq.c:240
[<ffffffff86653d4c>] common_interrupt+0x8c/0x8c arch/x86/entry/entry_64.S:520
<EOI>
[< inline >] rcu_lock_acquire include/linux/rcupdate.h:490
[< inline >] rcu_read_lock include/linux/rcupdate.h:874
[<ffffffff8164b4a1>] filemap_map_pages+0x131/0xba0 mm/filemap.c:2145
[< inline >] do_fault_around mm/memory.c:2943
[< inline >] do_read_fault mm/memory.c:2962
[< inline >] do_fault mm/memory.c:3133
[< inline >] handle_pte_fault mm/memory.c:3308
[< inline >] __handle_mm_fault mm/memory.c:3418
[<ffffffff816efb16>] handle_mm_fault+0x2516/0x49a0 mm/memory.c:3447
[<ffffffff8127dc16>] __do_page_fault+0x376/0x960 arch/x86/mm/fault.c:1238
[<ffffffff8127e358>] trace_do_page_fault+0xe8/0x420 arch/x86/mm/fault.c:1331
[<ffffffff8126f514>] do_async_page_fault+0x14/0xd0 arch/x86/kernel/kvm.c:264
[<ffffffff86655578>] async_page_fault+0x28/0x30 arch/x86/entry/entry_64.S:986
Fix it by ensuring that the polling path is holding the host lock
before entering ata_sff_hsm_move() so that all hardware accesses and
state updates are performed under the host lock.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-and-tested-by: Dmitry Vyukov <dvyukov@google.com>
Link: http://lkml.kernel.org/g/CACT4Y+b_JsOxJu2EZyEf+mOXORc_zid5V1-pLZSroJVxyWdSpw@mail.gmail.com
Cc: stable@vger.kernel.org
|
|
ata_sff_hsm_move() triggers BUG if it sees a host state machine state
that it dind't expect. The risk for data corruption when the
condition occurs is low as it's highly unlikely that it would lead to
spurious completion of commands. The BUG occasionally triggered for
subtle race conditions in the driver. Let's downgrade it to WARN so
that it doesn't kill the machine unnecessarily.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Dmitry Vyukov <dvyukov@google.com>
|
|
Replace occurences of the pci api by appropriate call to the dma api.
A simplified version of the semantic patch that finds this problem is as
follows: (http://coccinelle.lip6.fr)
@deprecated@
idexpression id;
position p;
@@
(
pci_dma_supported@p ( id, ...)
|
pci_alloc_consistent@p ( id, ...)
)
@bad1@
idexpression id;
position deprecated.p;
@@
...when != &id->dev
when != pci_get_drvdata ( id )
when != pci_enable_device ( id )
(
pci_dma_supported@p ( id, ...)
|
pci_alloc_consistent@p ( id, ...)
)
@depends on !bad1@
idexpression id;
expression direction;
position deprecated.p;
@@
(
- pci_dma_supported@p ( id,
+ dma_supported ( &id->dev,
...
+ , GFP_ATOMIC
)
|
- pci_alloc_consistent@p ( id,
+ dma_alloc_coherent ( &id->dev,
...
+ , GFP_ATOMIC
)
)
Signed-off-by: Quentin Lambert <lambert.quentin@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
|
|
It is possible for ata_sff_flush_pio_task() to set ap->hsm_task_state to
HSM_ST_IDLE in between the time __ata_sff_port_intr() checks for HSM_ST_IDLE
and before it calls ata_sff_hsm_move() causing ata_sff_hsm_move() to BUG().
This problem is hard to reproduce making this patch hard to verify, but this
fix will prevent the race.
I have not been able to reproduce the problem, but here is a crash dump from
a 2.6.32 kernel.
On examining the ata port's state, its hsm_task_state field has a value of HSM_ST_IDLE:
crash> struct ata_port.hsm_task_state ffff881c1121c000
hsm_task_state = 0
Normally, this should not be possible as ata_sff_hsm_move() was called from ata_sff_host_intr(),
which checks hsm_task_state and won't call ata_sff_hsm_move() if it has a HSM_ST_IDLE value.
PID: 11053 TASK: ffff8816e846cae0 CPU: 0 COMMAND: "sshd"
#0 [ffff88008ba03960] machine_kexec at ffffffff81038f3b
#1 [ffff88008ba039c0] crash_kexec at ffffffff810c5d92
#2 [ffff88008ba03a90] oops_end at ffffffff8152b510
#3 [ffff88008ba03ac0] die at ffffffff81010e0b
#4 [ffff88008ba03af0] do_trap at ffffffff8152ad74
#5 [ffff88008ba03b50] do_invalid_op at ffffffff8100cf95
#6 [ffff88008ba03bf0] invalid_op at ffffffff8100bf9b
[exception RIP: ata_sff_hsm_move+317]
RIP: ffffffff813a77ad RSP: ffff88008ba03ca0 RFLAGS: 00010097
RAX: 0000000000000000 RBX: ffff881c1121dc60 RCX: 0000000000000000
RDX: ffff881c1121dd10 RSI: ffff881c1121dc60 RDI: ffff881c1121c000
RBP: ffff88008ba03d00 R8: 0000000000000000 R9: 000000000000002e
R10: 000000000001003f R11: 000000000000009b R12: ffff881c1121c000
R13: 0000000000000000 R14: 0000000000000050 R15: ffff881c1121dd78
ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0018
#7 [ffff88008ba03d08] ata_sff_host_intr at ffffffff813a7fbd
#8 [ffff88008ba03d38] ata_sff_interrupt at ffffffff813a821e
#9 [ffff88008ba03d78] handle_IRQ_event at ffffffff810e6ec0
--- <IRQ stack> ---
[exception RIP: pipe_poll+48]
RIP: ffffffff81192780 RSP: ffff880f26d459b8 RFLAGS: 00000246
RAX: 0000000000000000 RBX: ffff880f26d459c8 RCX: 0000000000000000
RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff881a0539fa80
RBP: ffffffff8100bb8e R8: ffff8803b23324a0 R9: 0000000000000000
R10: ffff880f26d45dd0 R11: 0000000000000008 R12: ffffffff8109b646
R13: ffff880f26d45948 R14: 0000000000000246 R15: 0000000000000246
ORIG_RAX: ffffffffffffff10 CS: 0010 SS: 0018
RIP: 00007f26017435c3 RSP: 00007fffe020c420 RFLAGS: 00000206
RAX: 0000000000000017 RBX: ffffffff8100b072 RCX: 00007fffe020c45c
RDX: 00007f2604a3f120 RSI: 00007f2604a3f140 RDI: 000000000000000d
RBP: 0000000000000000 R8: 00007fffe020e570 R9: 0101010101010101
R10: 0000000000000000 R11: 0000000000000246 R12: 00007fffe020e5f0
R13: 00007fffe020e5f4 R14: 00007f26045f373c R15: 00007fffe020e5e0
ORIG_RAX: 0000000000000017 CS: 0033 SS: 002b
Somewhere between the ata_sff_hsm_move() check and the ata_sff_host_intr() check, the value changed.
On examining the other cpus to see what else was running, another cpu was running the error handler
routines:
PID: 326 TASK: ffff881c11014aa0 CPU: 1 COMMAND: "scsi_eh_1"
#0 [ffff88008ba27e90] crash_nmi_callback at ffffffff8102fee6
#1 [ffff88008ba27ea0] notifier_call_chain at ffffffff8152d515
#2 [ffff88008ba27ee0] atomic_notifier_call_chain at ffffffff8152d57a
#3 [ffff88008ba27ef0] notify_die at ffffffff810a154e
#4 [ffff88008ba27f20] do_nmi at ffffffff8152b1db
#5 [ffff88008ba27f50] nmi at ffffffff8152aaa0
[exception RIP: _spin_lock_irqsave+47]
RIP: ffffffff8152a1ff RSP: ffff881c11a73aa0 RFLAGS: 00000006
RAX: 0000000000000001 RBX: ffff881c1121deb8 RCX: 0000000000000000
RDX: 0000000000000246 RSI: 0000000000000020 RDI: ffff881c122612d8
RBP: ffff881c11a73aa0 R8: ffff881c17083800 R9: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: ffff881c1121c000
R13: 000000000000001f R14: ffff881c1121dd50 R15: ffff881c1121dc60
ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0000
--- <NMI exception stack> ---
#6 [ffff881c11a73aa0] _spin_lock_irqsave at ffffffff8152a1ff
#7 [ffff881c11a73aa8] ata_exec_internal_sg at ffffffff81396fb5
#8 [ffff881c11a73b58] ata_exec_internal at ffffffff81397109
#9 [ffff881c11a73bd8] atapi_eh_request_sense at ffffffff813a34eb
Before it tried to acquire a spinlock, ata_exec_internal_sg() called ata_sff_flush_pio_task().
This function will set ap->hsm_task_state to HSM_ST_IDLE, and has no locking around setting this
value. ata_sff_flush_pio_task() can then race with the interrupt handler and potentially set
HSM_ST_IDLE at a fatal moment, which will trigger a kernel BUG.
v2: Fixup comment in ata_sff_flush_pio_task()
tj: Further updated comment. Use ap->lock instead of shost lock and
use the [un]lock_irq variant instead of the irqsave/restore one.
Signed-off-by: David Milburn <dmilburn@redhat.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: stable@vger.kernel.org
|
|
Currently, ata_sff_softreset is skipped for controllers with no ctl port.
But that also skips ata_sff_dev_classify required for device detection.
This means that libata is currently broken on controllers with no ctl port.
No device connected:
[ 1.872480] pata_isapnp 01:01.02: activated
[ 1.889823] scsi2 : pata_isapnp
[ 1.890109] ata3: PATA max PIO0 cmd 0x1e8 ctl 0x0 irq 11
[ 6.888110] ata3.01: qc timeout (cmd 0xec)
[ 6.888179] ata3.01: failed to IDENTIFY (I/O error, err_mask=0x5)
[ 16.888085] ata3.01: qc timeout (cmd 0xec)
[ 16.888147] ata3.01: failed to IDENTIFY (I/O error, err_mask=0x5)
[ 46.888086] ata3.01: qc timeout (cmd 0xec)
[ 46.888148] ata3.01: failed to IDENTIFY (I/O error, err_mask=0x5)
[ 51.888100] ata3.00: qc timeout (cmd 0xec)
[ 51.888160] ata3.00: failed to IDENTIFY (I/O error, err_mask=0x5)
[ 61.888079] ata3.00: qc timeout (cmd 0xec)
[ 61.888141] ata3.00: failed to IDENTIFY (I/O error, err_mask=0x5)
[ 91.888089] ata3.00: qc timeout (cmd 0xec)
[ 91.888152] ata3.00: failed to IDENTIFY (I/O error, err_mask=0x5)
ATAPI device connected:
[ 1.882061] pata_isapnp 01:01.02: activated
[ 1.893430] scsi2 : pata_isapnp
[ 1.893719] ata3: PATA max PIO0 cmd 0x1e8 ctl 0x0 irq 11
[ 6.892107] ata3.01: qc timeout (cmd 0xec)
[ 6.892171] ata3.01: failed to IDENTIFY (I/O error, err_mask=0x5)
[ 16.892079] ata3.01: qc timeout (cmd 0xec)
[ 16.892138] ata3.01: failed to IDENTIFY (I/O error, err_mask=0x5)
[ 46.892079] ata3.01: qc timeout (cmd 0xec)
[ 46.892138] ata3.01: failed to IDENTIFY (I/O error, err_mask=0x5)
[ 46.908586] ata3.00: ATAPI: ACER CD-767E/O, V1.5X, max PIO2, CDB intr
[ 46.924570] ata3.00: configured for PIO0 (device error ignored)
[ 46.926295] scsi 2:0:0:0: CD-ROM ACER CD-767E/O 1.5X PQ: 0 ANSI: 5
[ 46.984519] sr0: scsi3-mmc drive: 6x/6x xa/form2 tray
[ 46.984592] cdrom: Uniform CD-ROM driver Revision: 3.20
So don't skip ata_sff_softreset, just skip the reset part of ata_bus_softreset
if the ctl port is not available.
This makes IDE port on ES968 behave correctly:
No device connected:
[ 4.670888] pata_isapnp 01:01.02: activated
[ 4.673207] scsi host2: pata_isapnp
[ 4.673675] ata3: PATA max PIO0 cmd 0x1e8 ctl 0x0 irq 11
[ 7.081840] Adding 2541652k swap on /dev/sda2. Priority:-1 extents:1 across:2541652k
ATAPI device connected:
[ 4.704362] pata_isapnp 01:01.02: activated
[ 4.706620] scsi host2: pata_isapnp
[ 4.706877] ata3: PATA max PIO0 cmd 0x1e8 ctl 0x0 irq 11
[ 4.872782] ata3.00: ATAPI: ACER CD-767E/O, V1.5X, max PIO2, CDB intr
[ 4.888673] ata3.00: configured for PIO0 (device error ignored)
[ 4.893984] scsi 2:0:0:0: CD-ROM ACER CD-767E/O 1.5X PQ: 0 ANSI: 5
[ 7.015578] Adding 2541652k swap on /dev/sda2. Priority:-1 extents:1 across:2541652k
Signed-off-by: Ondrej Zary <linux@rainbow-software.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: stable@vger.kernel.org
|
|
Ever since v2.6.19 the code contains a check for CONFIG_NO_ATA_LEGACY.
But that macro has never been defined. Apparently no one ran into
problems on platforms that do not support compatibility mode. So remove
this code that has been dead for over seven years.
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: Tejun Heo <tj@kernel.org>
|
|
Jeff moved on to a greener pasture.
s/Maintained by: Jeff Garzik/Maintained by: Tejun Heo/g
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jeff Garzik <jgarzik@pobox.com>
|
|
Acked-by: Jeff Garzik <jgarzik@redhat.com>
Signed-off-by: Cong Wang <amwang@redhat.com>
|
|
atapi_pio_bytes() uses bare numbers for the ATAPI interrupt reason bits despite
these are #define'd in <linux/ata.h>.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
fix these errors:
drivers/ata/libata-sff.c:2538:3: error: implicit declaration of function
'ata_pci_bmdma_prepare_host'
drivers/ata/libata-sff.c:2549:40: error: 'ata_bmdma_interrupt'
undeclared (first use in this function)
Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
A pending cleanup will mean that module.h won't be implicitly
everywhere anymore. Make sure the modular drivers in the ide dir
are actually calling out for <module.h> explicitly in advance.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
|
|
Always use ata_sff_data_xfer32() in ata_sff_data_xfer_noirq()
so the latter can be also used for host controllers supporting
32-bit PIO operations.
It is a completely safe thing to do because if 32-bit PIO is
not supported or enabled ata_sff_data_xfer32() will fallback
to a standard method.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Turn both helpers (which are used only during LLDs initialization
time and thus are not performance sensitive) into wrappers around
the new ata_pci_init_one() function, this cuts 20 LOC and saves
~1.1k of the output code size (x86-64):
text data bss dec hex filename
21392 0 19 21411 53a3 drivers/ata/libata-sff.o.before
20256 0 19 20275 4f33 drivers/ata/libata-sff.o.after
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
ata_sff_data_xfer[32]() use pad area if the transfer size isn't
multiple of transfer size; however, this area wasn't cleared and
garbage data in pad area could be transferred to the device. Make
sure the pad area is cleared.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Lei Ming <tom.leiming@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Cleanup sff_pio_task_link when a command is cancel while the
pio_task thread has been scheduled.
Signed-off-by: Gwendal Grignou <gwendal@google.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
|
|
Saves text by removing nearly duplicated text format strings by
creating ata_<foo>_printk functions and printf extension %pV.
ata defconfig size shrinks ~5% (~8KB), allyesconfig ~2.5% (~13KB)
Format string duplication comes from:
#define ata_link_printk(link, lv, fmt, args...) do { \
if (sata_pmp_attached((link)->ap) || (link)->ap->slave_link) \
printk("%sata%u.%02u: "fmt, lv, (link)->ap->print_id, \
(link)->pmp , ##args); \
else \
printk("%sata%u: "fmt, lv, (link)->ap->print_id , ##args); \
} while(0)
Coalesce long formats.
$ size drivers/ata/built-in.*
text data bss dec hex filename
544969 73893 116584 735446 b38d6 drivers/ata/built-in.allyesconfig.ata.o
558429 73893 117864 750186 b726a drivers/ata/built-in.allyesconfig.dev_level.o
141328 14689 4220 160237 271ed drivers/ata/built-in.defconfig.ata.o
149567 14689 4220 168476 2921c drivers/ata/built-in.defconfig.dev_level.o
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
|
|
Saves a bit of text as the call takes fewer args.
Coalesce a few formats.
Convert a few bare printks to pr_cont.
$ size drivers/ata/built-in.o*
text data bss dec hex filename
558429 73893 117864 750186 b726a drivers/ata/built-in.o.allyesconfig.new
559574 73893 117888 751355 b76fb drivers/ata/built-in.o.allyesconfig.old
149567 14689 4220 168476 2921c drivers/ata/built-in.o.defconfig.new
149851 14689 4220 168760 29338 drivers/ata/built-in.o.defconfig.old
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
|
|
This is a cosmetic change to prevent libata-sff adding irq
descriptions to dummy ports, since the information, while largely
unused, is erroneous.
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
|
|
Fixes generated by 'codespell' and manually reviewed.
Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
|
|
This patch adds ata_sff_queue_work() & ata_sff_queue_delayed_work() routine in
libata-sff.c file. This routine can be used by ata drivers to use ata_sff_wq.
Signed-off-by: Viresh Kumar <viresh.kumar@st.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
|
|
|
|
* 'for-2.6.38' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq: (33 commits)
usb: don't use flush_scheduled_work()
speedtch: don't abuse struct delayed_work
media/video: don't use flush_scheduled_work()
media/video: explicitly flush request_module work
ioc4: use static work_struct for ioc4_load_modules()
init: don't call flush_scheduled_work() from do_initcalls()
s390: don't use flush_scheduled_work()
rtc: don't use flush_scheduled_work()
mmc: update workqueue usages
mfd: update workqueue usages
dvb: don't use flush_scheduled_work()
leds-wm8350: don't use flush_scheduled_work()
mISDN: don't use flush_scheduled_work()
macintosh/ams: don't use flush_scheduled_work()
vmwgfx: don't use flush_scheduled_work()
tpm: don't use flush_scheduled_work()
sonypi: don't use flush_scheduled_work()
hvsi: don't use flush_scheduled_work()
xen: don't use flush_scheduled_work()
gdrom: don't use flush_scheduled_work()
...
Fixed up trivial conflict in drivers/media/video/bt8xx/bttv-input.c
as per Tejun.
|
|
While separating out BMDMA irq handler from SFF, commit c3b28894
(libata-sff: separate out BMDMA irq handler) incorrectly made
__ata_sff_port_intr() consider an IRQ to be an idle one if the host
state was transitioned to HSM_ST_ERR by ata_bmdma_port_intr().
This makes BMDMA drivers ignore IRQs reporting host bus error which
leads to timeouts instead of triggering EH immediately. Fix it by
making __ata_sff_port_intr() consider the IRQ to be an idle one iff
the state is HSM_ST_IDLE. This is equivalent to adding HSM_ST_ERR to
the "break"ing case but less error-prone.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Antonio Toma <antonio.toma@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
cancel_delayed_work_sync()
cancel_rearming_delayed_work[queue]() has been superceded by
cancel_delayed_work_sync() quite some time ago. Convert all the
in-kernel users. The conversions are completely equivalent and
trivial.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: "David S. Miller" <davem@davemloft.net>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Acked-by: Evgeniy Polyakov <zbr@ioremap.net>
Cc: Jeff Garzik <jgarzik@pobox.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Mauro Carvalho Chehab <mchehab@infradead.org>
Cc: netdev@vger.kernel.org
Cc: Anton Vorontsov <cbou@mail.ru>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Neil Brown <neilb@suse.de>
Cc: Alex Elder <aelder@sgi.com>
Cc: xfs-masters@oss.sgi.com
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: netfilter-devel@vger.kernel.org
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: linux-nfs@vger.kernel.org
|
|
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
workqueue: remove in_workqueue_context()
workqueue: Clarify that schedule_on_each_cpu is synchronous
memory_hotplug: drop spurious calls to flush_scheduled_work()
shpchp: update workqueue usage
pciehp: update workqueue usage
isdn/eicon: don't call flush_scheduled_work() from diva_os_remove_soft_isr()
workqueue: add and use WQ_MEM_RECLAIM flag
workqueue: fix HIGHPRI handling in keep_working()
workqueue: add queue_work and activate_work trace points
workqueue: prepare for more tracepoints
workqueue: implement flush[_delayed]_work_sync()
workqueue: factor out start_flush_work()
workqueue: cleanup flush/cancel functions
workqueue: implement alloc_ordered_workqueue()
Fix up trivial conflict in fs/gfs2/main.c as per Tejun
|
|
Add optional @ap argument to ata_wait_register() and replace msleep()
calls with ata_msleep() which take optional @ap in addition to the
duration. These will be used to implement EH exclusion.
This patch doesn't cause any behavior difference.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
This build error showed up in linux-next tag next-20100820 for ia64:
WARNING: vmlinux.o(.init.text+0x4a952): Section mismatch in reference from the function ata_init() to the function .exit.text:ata_sff_exit()
The function __init ata_init() references
a function __exit ata_sff_exit().
This is often seen when error handling in the init function
uses functionality in the exit path.
The fix is often to remove the __exit annotation of
ata_sff_exit() so it may be used outside an exit section.
Sure enough, dropping the __exit fixes the problem.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Add WQ_MEM_RECLAIM flag which currently maps to WQ_RESCUER, mark
WQ_RESCUER as internal and replace all external WQ_RESCUER usages to
WQ_MEM_RECLAIM.
This makes the API users express the intent of the workqueue instead
of indicating the internal mechanism used to guarantee forward
progress. This is also to make it cleaner to add more semantics to
WQ_MEM_RECLAIM. For example, if deemed necessary, memory reclaim
workqueues can be made highpri.
This patch doesn't introduce any functional change.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jeff Garzik <jgarzik@pobox.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
|
|
Keep track of the link on the which the current request is in progress.
It allows support of links behind port multiplier.
Not all libata-sff is PMP compliant. Code for native BMDMA controller
does not take in accound PMP.
Tested on Marvell 7042 and Sil7526.
Signed-off-by: Gwendal Grignou <gwendal@google.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Commit 978c0666 (libata: Remove excess delay in the tf_load path)
removed ata_wait_idle() from ata_sff_tf_load() and via_tf_load().
This caused obscure detection problems in sata_sil.
https://bugzilla.kernel.org/show_bug.cgi?id=16606
The commit was pure performance optimization. Revert it for now.
Reported-by: Dieter Plaetinck <dieter@plaetinck.be>
Reported-by: Jan Beulich <JBeulich@novell.com>
Bisected-by: gianluca <gianluca@sottospazio.it>
Cc: stable@kernel.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Remove harmful BUG_ON() from ata_bmdma_qc_issue(),
as it casts too wide of a net and breaks sata_mv.
It also crashes the kernel while doing the BUG_ON().
There's already a WARN_ON_ONCE() further down to catch
the case of POLLING for a BMDMA operation.
Signed-off-by: Mark Lord <mlord@pobox.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Cc: stable@kernel.org
|
|
libata has two concurrency related limitations.
a. ata_wq which is used for polling PIO has single thread per CPU. If
there are multiple devices doing polling PIO on the same CPU, they
can't be executed simultaneously.
b. ata_aux_wq which is used for SCSI probing has single thread. In
cases where SCSI probing is stalled for extended period of time
which is possible for ATAPI devices, this will stall all probing.
#a is solved by increasing maximum concurrency of ata_wq. Please note
that polling PIO might be used under allocation path and thus needs to
be served by a separate wq with a rescuer.
#b is solved by using the default wq instead and achieving exclusion
via per-port mutex.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Jeff Garzik <jgarzik@pobox.com>
|
|
Make BMDMA optional depending on new config variable CONFIG_ATA_BMDMA.
In Kconfig, drivers are grouped into five groups - non-SFF native, SFF
w/ custom DMA interface, SFF w/ BMDMA, PIO-only SFF, and generic
fallback / legacy ones. Kconfig and Makefile are reorganized
according to the groups and ordered alphabetically inside each group.
ata_ioports.bmdma_addr and ata_port.bmdma_prd[_dma] are put into
CONFIG_ATA_BMDMA, as are all bmdma related ops, variables and
functions.
This increase the binary size slightly when BMDMA is enabled but on
both native-only and PIO-only configurations the size is slightly
reduced. Either way, the size difference is insignificant. This
change is more meaningful to signify the separation between SFF and
BMDMA and as a tool to verify the separation.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Separate out ata_pci_bmdma_prepare_host() and ata_pci_bmdma_init_one()
from their SFF counterparts. SFF ones no longer try to initialize
BMDMA or set PCI master.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Separate out BMDMA irq handler from SFF irq handler. The misnamed
host_intr() functions are renamed to ata_sff_port_intr() and
ata_bmdma_port_intr(). Common parts are factored into
__ata_sff_port_intr() and __ata_sff_interrupt() and used by sff and
bmdma interrupt routines.
All BMDMA drivers now use ata_bmdma_interrupt() or
ata_bmdma_port_intr() while all non-BMDMA SFF ones use
ata_sff_interrupt() or ata_sff_port_intr().
For now, ata_pci_sff_init_one() uses ata_bmdma_interrupt() as it's
used by both SFF and BMDMA drivers.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
ata_sff_irq_clear() is BMDMA specific. Rename it to
ata_bmdma_irq_clear(), move it to ata_bmdma_port_ops and make
->sff_irq_clear() optional.
Note: ata_bmdma_irq_clear() is actually only needed by ata_piix and
possibly by sata_sil. This should be moved to respective low
level drivers later.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Separate out ata_bmdma_qc_issue() from ata_sff_qc_issue() such that
ata_sff_qc_issue() only deals with non-BMDMA SFF protocols (PIO and
nodata) while ata_bmdma_qc_issue() deals with the BMDMA protocols and
uses ata_sff_qc_issue() for non-DMA commands. All the users are
updated accordingly.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
struct ata_prd and ap->prd are BMDMA specific. Add bmdma_ prefix to
them and move them inside CONFIG_ATA_SFF.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Both qc_prep functions deal only with BMDMA PRD setup and PIO only SFF
drivers don't need them. Rename to ata_bmdma_[dumb_]qc_prep() and
relocate.
All usages are renamed except for pdc_adma and sata_qstor. Those two
drivers are not BMDMA drivers and don't need to call BMDMA qc_prep
functions. Calls to ata_sff_qc_prep() in the two drivers are removed.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Some of error handling logic in ata_sff_error_handler() and all of
ata_sff_post_internal_cmd() are for BMDMA. Create
ata_bmdma_error_handler() and ata_bmdma_post_internal_cmd() and move
BMDMA part into those.
While at it, change DMA protocol check to ata_is_dma(), fix
post_internal_cmd to call ap->ops->bmdma_stop instead of directly
calling ata_bmdma_stop() and open code hardreset selection so that
ata_std_error_handler() doesn't have to know about sff hardreset.
As these two functions are BMDMA specific, there's no reason to check
for bmdma_addr before calling bmdma methods if the protocol of the
failed command is DMA. sata_mv and pata_mpc52xx now don't need to set
.post_internal_cmd to ATA_OP_NULL and pata_icside and sata_qstor don't
need to set it to their bmdma_stop routines.
ata_sff_post_internal_cmd() becomes noop and is removed.
This fixes p3 described in clean-up-BMDMA-initialization patch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
port_task is tightly bound to the standard SFF PIO HSM implementation.
Using it for any other purpose would be error-prone and there's no
such user and if some drivers need such feature, it would be much
better off using its own. Move it inside CONFIG_ATA_SFF and rename it
to sff_pio_task.
The only function which is exposed to the core layer is
ata_sff_flush_pio_task() which is renamed from ata_port_flush_task()
and now also takes care of resetting hsm_task_state to HSM_ST_IDLE,
which is possible as it's now specific to PIO HSM.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
ap->[last_]ctl are specific to SFF controllers. Put them inside
CONFIG_ATA_SFF and move initialization into ata_sff_port_init().
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
->drain_fifo() is SFF specific. Rename and relocate it.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
In preparation of proper SFF/BMDMA separation, introduce
ata_sff_init/exit() and ata_sff_port_init(). These functions
currently don't do anything.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
When BMDMA initialization failed or BMDMA was not available for
whatever reason, bmdma_addr was left at zero and used as an indication
that BMDMA shouldn't be used. This leads to the following problems.
p1. For BMDMA drivers which don't use traditional BMDMA register,
ata_bmdma_mode_filter() incorrectly inhibits DMA modes. Those
drivers either have to inherit from ata_sff_port_ops or clear
->mode_filter explicitly.
p2. non-BMDMA drivers call into BMDMA PRD table allocation. It
doesn't actually allocate PRD table if bmdma_addr is not
initialized but is still confusing.
p3. For BMDMA drivers which don't use traditional BMDMA register, some
methods might not be invoked as expected (e.g. bmdma_stop from
ata_sff_post_internal_cmd()).
p4. SFF drivers w/ custom DMA interface implement noop BMDMA ops
worrying libata core might call into one of them.
These problems are caused by the muddy line between SFF and BMDMA and
the assumption that all BMDMA controllers initialize bmdma_addr.
This patch fixes p1 and p2 by removing the bmdma_addr assumption and
moving prd allocation to BMDMA port start. Later patches will fix the
remaining issues.
This patch improves BMDMA initialization such that
* When BMDMA register initialization fails, falls back to PIO instead
of failing. ata_pci_bmdma_init() never fails now.
* When ata_pci_bmdma_init() falls back to PIO, it clears
ap->mwdma_mask and udma_mask instead of depending on
ata_bmdma_mode_filter(). This makes ata_bmdma_mode_filter()
unnecessary thus resolving p1.
* ata_port_start() which actually is BMDMA specific is moved to
ata_bmdma_port_start(). ata_port_start() and ata_sff_port_start()
are killed.
* ata_sff_port_start32() is moved and renamed to
ata_bmdma_port_start32().
Drivers which no longer call into PRD table allocation are...
pdc_adma, sata_inic162x, sata_qstor, sata_sx4, pata_cmd640 and all
drivers which inherit from ata_sff_port_ops.
pata_icside sets ->port_start to ATA_OP_NULL as it doesn't need PRD
but is a BMDMA controller and doesn't have custom port_start like
other such controllers.
Note that with the previous patch which makes all and only BMDMA
drivers inherit from ata_bmdma_port_ops, this change doesn't break
drivers which need PRD table.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Reorder functions such that SFF and BMDMA functions are grouped.
While at it, s/BMDMA/SFF in a few comments where it actually meant
SFF.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
ATA_FLAG_DISABLED is only used by drivers which don't use
->error_handler framework and is largely broken. Its only meaningful
function is to make irq handlers skip processing if the flag is set,
which is largely useless and even harmful as it makes those ports more
likely to cause IRQ storms.
Kill ATA_FLAG_DISABLED and makes the callers disable attached devices
instead. ata_port_probe() and ata_port_disable() which manipulate the
flag are also killed.
This simplifies condition check in IRQ handlers. While updating IRQ
handlers, remove ap NULL check as libata guarantees consecutive port
allocation (unoccupied ports are initialized with dummies) and
long-obsolete ATA_QCFLAG_ACTIVE check (checked by ata_qc_from_tag()).
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
ata_irq_on() was renamed to ata_sff_irq_on() and exported a while ago
but prototype for the original function lingered in
drivers/ata/libata.h. Kill it. Also, ata_dev_select() is only used
inside drivers/ata/libata-sff.c. Make it static.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
* Clearing IRQ from ata_sff_error_handler() is necessary only when the
port is gonna be thawed before performing EH actions and some
controllers don't like being accessed after certain failure modes
until they're reset. Clear IRQ iff the port is being thawed.
* When the controller succesfully indicated bus error, the point of
thawing doesn't matter. Move thawing inside bmdma part of EH. This
is a bit ugly but will ease code reorganization later.
* Remove the unneeded ata_sff_sync().
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
We don't need to stall and wait after loading the task file and before
issuing a command, so don't do it. This shows up on profiles and is not
needed.
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
