Skip to content

K
Kernel

openeuler / Kernel
1 年多 前同步成功

通知 8
Star 0
Fork 0
K
Kernel
切换分支/标签
  1. 20 5月, 2014 1 次提交
  3. 18 5月, 2014 1 次提交
    • M
      target: fix memory leak on XCOPY · 1e1110c4
      由 Mikulas Patocka 提交于 
      On each processed XCOPY command, two "kmalloc-512" memory objects are
leaked. These represent two allocations of struct xcopy_pt_cmd in
target_core_xcopy.c.

The reason for the memory leak is that the cmd_kref field is not
initialized (thus, it is zero because the allocations were done with
kzalloc). When we decrement zero kref in target_put_sess_cmd, the result
is not zero, thus target_release_cmd_kref is not called.

This patch fixes the bug by moving kref initialization from
target_get_sess_cmd to transport_init_se_cmd (this function is called from
target_core_xcopy.c, so it will correctly initialize cmd_kref). It can be
easily verified that all code that calls target_get_sess_cmd also calls
transport_init_se_cmd earlier, thus moving kref_init shouldn't introduce
any new problems.
Signed-off-by: NMikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org	# 3.12+
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>
      1e1110c4
  5. 16 5月, 2014 8 次提交
  7. 26 4月, 2014 11 次提交
  9. 25 4月, 2014 17 次提交
    • T
      serial_core: fix uart PORT_UNKNOWN handling · 7deb39ed
      由 Thomas Pfaff 提交于 
      While porting a RS485 driver from 2.6.29 to 3.14, i noticed that the serial tty
driver could break it by using uart ports that it does not own :

1. uart_change_pm ist called during uart_open and calls the uart pm function
   without checking for PORT_UNKNOWN.
   The fix is to move uart_change_pm from uart_open to uart_port_startup.
2. The return code from the uart request_port call in uart_set_info is not
   handled properly, leading to the situation that the serial driver also
   thinks it owns the uart ports.
   This can triggered by doing following actions :

   setserial /dev/ttyS0 uart none    # release the uart ports
   modprobe lirc-serial              # or any other device that uses the uart
   setserial /dev/ttyS0 uart 16550   # gives no error and the uart tty driver
                                     # can use the ports as well
Signed-off-by: NThomas Pfaff <tpfaff@pcs.com>
Reviewed-by: NAlan Cox <alan@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      7deb39ed
    • D
      serial: samsung: Change barrier() to cpu_relax() in console output · f94b0572
      由 Doug Anderson 提交于 
      The two functions to write out to the console (one used in normal
console mode and one in polling console mode) were slightly different.
One used a barrier() in its loop and the other a cpu_relax().  The
barrier() really doesn't do anything since we're using rd_regl() to
read the port anyway.  Switch it to cpu_relax() to make things
consistent.

No known bugs / issues are fixed by this change--it just makes things
more consistent.
Signed-off-by: NDoug Anderson <dianders@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      f94b0572
    • D
      serial: samsung: don't check config for every character · ab88c8dc
      由 Doug Anderson 提交于 
      The s3c24xx_serial_console_putchar() is _only_ ever used by
s3c24xx_serial_console_write() and is called in a loop (indirectly
through uart_console_write()).  There's no reason to call
s3c24xx_port_configured() for every iteration through the loop.  Move
it outside the loop.
Signed-off-by: NDoug Anderson <dianders@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      ab88c8dc
    • D
      serial: samsung: Use the passed in "port", fixing kgdb w/ no console · bb7f09ba
      由 Doug Anderson 提交于 
      The two functions in the samsung serial driver used for writing
characters out to the port were inconsistent about whether they used
the passed in "port" or the global "cons_uart".  There was no reason
to use the global and the use of the global in
s3c24xx_serial_put_poll_char() caused a crash in the case where you
used the serial port for kgdboc but not for console.

Fix it so we used the passed in variable.

Note that this doesn't fix all problems with the samsung serial
driver.  Specifically:
* s3c24xx_serial_console_putchar() is still 99% identical to
  s3c24xx_serial_put_poll_char() (the function signature is different,
  but that's about it).  A future patch will make them slightly less
  identical and judging by other serial drivers we may need yet more
  differences eventually.
* The samsung serial driver still doesn't allow you to have more than
  one console port since it still uses the global cons_uart in
  s3c24xx_serial_console_write().
Signed-off-by: NDoug Anderson <dianders@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      bb7f09ba
    • L
      serial: 8250: Fix thread unsafe __dma_tx_complete function · f8fd1b03
      由 Loic Poulain 提交于 
      __dma_tx_complete is not protected against concurrent
call of serial8250_tx_dma. it can lead to circular tail
index corruption or parallel call of serial_tx_dma on the
same data portion.

This patch fixes this issue by holding the port lock.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Reviewed-by: NHeikki Krogerus <heikki.krogerus@linux.intel.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      f8fd1b03
    • L
      8250_core: Fix unwanted TX chars write · b08c9c31
      由 Loic Poulain 提交于 
      On transmit-hold-register empty, serial8250_tx_chars
should be called only if we don't use DMA.
DMA has its own tx cycle.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Reviewed-by: NHeikki Krogerus <heikki.krogerus@linux.intel.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      b08c9c31
    • M
      tty: Fix race condition between __tty_buffer_request_room and flush_to_ldisc · 6a20dbd6
      由 Manfred Schlaegl 提交于 
      The race was introduced while development of linux-3.11 by
e8437d7e and
e9975fde.
Originally it was found and reproduced on linux-3.12.15 and
linux-3.12.15-rt25, by sending 500 byte blocks with 115kbaud to the
target uart in a loop with 100 milliseconds delay.

In short:
 1. The consumer flush_to_ldisc is on to remove the head tty_buffer.
 2. The producer adds a number of bytes, so that a new tty_buffer must
	be allocated and added by __tty_buffer_request_room.
 3. The consumer removes the head tty_buffer element, without handling
	newly committed data.

Detailed example:
 * Initial buffer:
   * Head, Tail -> 0: used=250; commit=250; read=240; next=NULL
 * Consumer: ''flush_to_ldisc''
   * consumed 10 Byte
   * buffer:
     * Head, Tail -> 0: used=250; commit=250; read=250; next=NULL
{{{
		count = head->commit - head->read;	// count = 0
		if (!count) {				// enter
			// INTERRUPTED BY PRODUCER ->
			if (head->next == NULL)
				break;
			buf->head = head->next;
			tty_buffer_free(port, head);
			continue;
		}
}}}
 * Producer: tty_insert_flip_... 10 bytes + tty_flip_buffer_push
   * buffer:
     * Head, Tail -> 0: used=250; commit=250; read=250; next=NULL
   * added 6 bytes: head-element filled to maximum.
     * buffer:
       * Head, Tail -> 0: used=256; commit=250; read=250; next=NULL
   * added 4 bytes: __tty_buffer_request_room is called
     * buffer:
       * Head -> 0: used=256; commit=256; read=250; next=1
       * Tail -> 1: used=4; commit=0; read=250 next=NULL
   * push (tty_flip_buffer_push)
     * buffer:
       * Head -> 0: used=256; commit=256; read=250; next=1
       * Tail -> 1: used=4; commit=4; read=250 next=NULL
 * Consumer
{{{
		count = head->commit - head->read;
		if (!count) {
			// INTERRUPTED BY PRODUCER <-
			if (head->next == NULL)		// -> no break
				break;
			buf->head = head->next;
			tty_buffer_free(port, head);
			// ERROR: tty_buffer head freed -> 6 bytes lost
			continue;
		}
}}}

This patch reintroduces a spin_lock to protect this case. Perhaps later
a lock-less solution could be found.
Signed-off-by: NManfred Schlaegl <manfred.schlaegl@gmx.at>
Cc: stable <stable@vger.kernel.org> # 3.11
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      6a20dbd6
    • S
      phy: fix kernel oops in phy_lookup() · 743bb387
      由 Sergei Shtylyov 提交于 
      The kernel oopses in phy_lookup() due to 'phy->init_data' being NULL if we
register PHYs from a device tree probing driver and then call phy_get() on a
device that has no representation in the device tree (e.g. a PCI device).
Checking the pointer before dereferening it and skipping an interation if
it's NULL prevents this kernel oops.
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NKishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      743bb387
    • J
      phy: restore OMAP_CONTROL_PHY dependencies · 907aa3aa
      由 Jean Delvare 提交于 
      When OMAP_CONTROL_USB was renamed to OMAP_CONTROL_PHY (commit
14da699b), its dependencies were lost in the process. Nothing in the
commit message indicates that this removal was intentional, so I think
it was by accident and the dependencies should be restored.
Signed-off-by: NJean Delvare <jdelvare@suse.de>
Acked-by: NRoger Quadros <rogerq@ti.com>
Cc: Felipe Balbi <balbi@ti.com>
Signed-off-by: NKishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      907aa3aa
    • A
      phy: exynos: fix building as a module · d1481832
      由 Arnd Bergmann 提交于 
      The top-level phy-samsung-usb2 driver may be configured as a
loadable module, which currently causes link errors because
of the dependency on the exynos{5250,4x12,4210}_usb2_phy_config
symbol. Solving this could be achieved by exporting these
symbols, but as the SoC-specific parts of the driver are not
currently built as modules, it seems better to just link
everything into one module and avoid the need for the export.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NKamil Debski <k.debski@samsung.com>
Signed-off-by: NKishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      d1481832
    • J
      USB: serial: fix sysfs-attribute removal deadlock · 10164c2a
      由 Johan Hovold 提交于 
      Fix driver new_id sysfs-attribute removal deadlock by making sure to
not hold any locks that the attribute operations grab when removing the
attribute.

Specifically, usb_serial_deregister holds the table mutex when
deregistering the driver, which includes removing the new_id attribute.
This can lead to a deadlock as writing to new_id increments the
attribute's active count before trying to grab the same mutex in
usb_serial_probe.

The deadlock can easily be triggered by inserting a sleep in
usb_serial_deregister and writing the id of an unbound device to new_id
during module unload.

As the table mutex (in this case) is used to prevent subdriver unload
during probe, it should be sufficient to only hold the lock while
manipulating the usb-serial driver list during deregister. A racing
probe will then either fail to find a matching subdriver or fail to get
the corresponding module reference.

Since v3.15-rc1 this also triggers the following lockdep warning:

======================================================
[ INFO: possible circular locking dependency detected ]
3.15.0-rc2 #123 Tainted: G        W
-------------------------------------------------------
modprobe/190 is trying to acquire lock:
 (s_active#4){++++.+}, at: [<c0167aa0>] kernfs_remove_by_name_ns+0x4c/0x94

but task is already holding lock:
 (table_lock){+.+.+.}, at: [<bf004d84>] usb_serial_deregister+0x3c/0x78 [usbserial]

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (table_lock){+.+.+.}:
       [<c0075f84>] __lock_acquire+0x1694/0x1ce4
       [<c0076de8>] lock_acquire+0xb4/0x154
       [<c03af3cc>] _raw_spin_lock+0x4c/0x5c
       [<c02bbc24>] usb_store_new_id+0x14c/0x1ac
       [<bf007eb4>] new_id_store+0x68/0x70 [usbserial]
       [<c025f568>] drv_attr_store+0x30/0x3c
       [<c01690e0>] sysfs_kf_write+0x5c/0x60
       [<c01682c0>] kernfs_fop_write+0xd4/0x194
       [<c010881c>] vfs_write+0xbc/0x198
       [<c0108e4c>] SyS_write+0x4c/0xa0
       [<c000f880>] ret_fast_syscall+0x0/0x48

-> #0 (s_active#4){++++.+}:
       [<c03a7a28>] print_circular_bug+0x68/0x2f8
       [<c0076218>] __lock_acquire+0x1928/0x1ce4
       [<c0076de8>] lock_acquire+0xb4/0x154
       [<c0166b70>] __kernfs_remove+0x254/0x310
       [<c0167aa0>] kernfs_remove_by_name_ns+0x4c/0x94
       [<c0169fb8>] remove_files.isra.1+0x48/0x84
       [<c016a2fc>] sysfs_remove_group+0x58/0xac
       [<c016a414>] sysfs_remove_groups+0x34/0x44
       [<c02623b8>] driver_remove_groups+0x1c/0x20
       [<c0260e9c>] bus_remove_driver+0x3c/0xe4
       [<c026235c>] driver_unregister+0x38/0x58
       [<bf007fb4>] usb_serial_bus_deregister+0x84/0x88 [usbserial]
       [<bf004db4>] usb_serial_deregister+0x6c/0x78 [usbserial]
       [<bf005330>] usb_serial_deregister_drivers+0x2c/0x4c [usbserial]
       [<bf016618>] usb_serial_module_exit+0x14/0x1c [sierra]
       [<c009d6cc>] SyS_delete_module+0x184/0x210
       [<c000f880>] ret_fast_syscall+0x0/0x48

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(table_lock);
                               lock(s_active#4);
                               lock(table_lock);
  lock(s_active#4);

 *** DEADLOCK ***

1 lock held by modprobe/190:
 #0:  (table_lock){+.+.+.}, at: [<bf004d84>] usb_serial_deregister+0x3c/0x78 [usbserial]

stack backtrace:
CPU: 0 PID: 190 Comm: modprobe Tainted: G        W     3.15.0-rc2 #123
[<c0015e10>] (unwind_backtrace) from [<c0013728>] (show_stack+0x20/0x24)
[<c0013728>] (show_stack) from [<c03a9a54>] (dump_stack+0x24/0x28)
[<c03a9a54>] (dump_stack) from [<c03a7cac>] (print_circular_bug+0x2ec/0x2f8)
[<c03a7cac>] (print_circular_bug) from [<c0076218>] (__lock_acquire+0x1928/0x1ce4)
[<c0076218>] (__lock_acquire) from [<c0076de8>] (lock_acquire+0xb4/0x154)
[<c0076de8>] (lock_acquire) from [<c0166b70>] (__kernfs_remove+0x254/0x310)
[<c0166b70>] (__kernfs_remove) from [<c0167aa0>] (kernfs_remove_by_name_ns+0x4c/0x94)
[<c0167aa0>] (kernfs_remove_by_name_ns) from [<c0169fb8>] (remove_files.isra.1+0x48/0x84)
[<c0169fb8>] (remove_files.isra.1) from [<c016a2fc>] (sysfs_remove_group+0x58/0xac)
[<c016a2fc>] (sysfs_remove_group) from [<c016a414>] (sysfs_remove_groups+0x34/0x44)
[<c016a414>] (sysfs_remove_groups) from [<c02623b8>] (driver_remove_groups+0x1c/0x20)
[<c02623b8>] (driver_remove_groups) from [<c0260e9c>] (bus_remove_driver+0x3c/0xe4)
[<c0260e9c>] (bus_remove_driver) from [<c026235c>] (driver_unregister+0x38/0x58)
[<c026235c>] (driver_unregister) from [<bf007fb4>] (usb_serial_bus_deregister+0x84/0x88 [usbserial])
[<bf007fb4>] (usb_serial_bus_deregister [usbserial]) from [<bf004db4>] (usb_serial_deregister+0x6c/0x78 [usbserial])
[<bf004db4>] (usb_serial_deregister [usbserial]) from [<bf005330>] (usb_serial_deregister_drivers+0x2c/0x4c [usbserial])
[<bf005330>] (usb_serial_deregister_drivers [usbserial]) from [<bf016618>] (usb_serial_module_exit+0x14/0x1c [sierra])
[<bf016618>] (usb_serial_module_exit [sierra]) from [<c009d6cc>] (SyS_delete_module+0x184/0x210)
[<c009d6cc>] (SyS_delete_module) from [<c000f880>] (ret_fast_syscall+0x0/0x48)
Signed-off-by: NJohan Hovold <jhovold@gmail.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      10164c2a
    • T
      usb: wusbcore: fix panic in wusbhc_chid_set · bd130ada
      由 Thomas Pugliese 提交于 
      If no valid CHID value has previously been set on an HWA, writing a
value of all zeros will cause a kernel panic in uwb_radio_stop because
wusbhc->uwb_rc has not been set.  This patch skips the call to
uwb_radio_stop if wusbhc->uwb_rc has not been initialized.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      bd130ada
    • T
      usb: wusbcore: convert nested lock to use spin_lock instead of spin_lock_irq · 7584f2eb
      由 Thomas Pugliese 提交于 
      Nesting a spin_lock_irq/unlock_irq inside a lock that has already
disabled interrupts will enable interrupts before we are ready when
spin_unlock_irq is called.  This patch converts the inner lock to use
spin_lock and spin_unlock instead.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      7584f2eb
    • T
      uwb: don't call spin_unlock_irq in a USB completion handler · c996b937
      由 Thomas Pugliese 提交于 
      This patch converts the use of spin_lock_irq/spin_unlock_irq to
spin_lock_irqsave/spin_unlock_irqrestore in uwb_rc_set_drp_cmd_done
which is called from a USB completion handler.  There are also
whitespace cleanups to make checkpatch.pl happy.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      c996b937
    • P
      usb: chipidea: coordinate usb phy initialization for different phy type · cd84f009
      由 Peter Chen 提交于 
      For internal PHY (like UTMI), the phy clock may from internal pll,
it is on/off on the fly, the access PORTSC.PTS will hang without
phy clock. So, the usb_phy_init which will open phy clock needs to
be called before hw_phymode_configure.
See: http://marc.info/?l=linux-arm-kernel&m=139350618732108&w=2

For external PHY (like ulpi), it needs to configure portsc.pts before
visit viewport, or the viewport can't be visited. so phy_phymode_configure
needs to be called before usb_phy_init.
See: cd0b42c2

It may not the best solution, but it can work for all situations.

Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: Chris Ruehl <chris.ruehl@gtsys.com.hk>
Cc: shc_work@mail.ru
Cc: denis@eukrea.com
Cc: festevam@gmail.com
Cc: stable <stable@vger.kernel.org> # 3.14
Signed-off-by: NPeter Chen <peter.chen@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      cd84f009
    • R
      ACPI / notify: Do not block unknown type notifications in root handler · f66abe92
      由 Rafael J. Wysocki 提交于 
      Commit 1a699476 "ACPI / hotplug / PCI: Hotplug notifications from
acpi_bus_notify()" changed the root notify handler, acpi_bus_notify(),
to block unknown type norifications, but it overlooked the fact that
they might be propagated to drivers via the ->notify() callback.

Fix the problem by allowing drivers to receive unknown type
notifications via ->notify() as before.

Fixes: 1a699476 (ACPI / hotplug / PCI: Hotplug notifications from acpi_bus_notify())
Reported-and-tested-by: NMantas Mikulėnas <grawity@gmail.com>
Reported-and-tested-by: NSitsofe Wheeler <sitsofe@yahoo.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
      f66abe92
    • P
      power/reset: vexpress: Fix restart/power off operation · d08b8037
      由 Pawel Moll 提交于 
      The restart/power off implementation in the vexpress driver
used to obtain the config function when necessary. This was
wrong in two respects:

1. It required memory allocation with disabled interrupts
(it worked, but lockdep - when enabled - reported warnings).

2. Used jiffies-based timeout, while jiffies are not running
at this stage of system shutdown (therefore a config
transaction error - if happened - would have never be reported).

Fixed by pre-allocating the config function per device
and using mdelay for timeout.
Signed-off-by: NPawel Moll <pawel.moll@arm.com>
      d08b8037
  11. 24 4月, 2014 2 次提交