BMC/Intel-BMC/linux.git/arch/xtensa/platforms/iss/console.c, branch dev

xtensa: ISS: fix timer_lock usage in rs_open

2013-05-09T08:07:11+00:00

This fixes the following lockdep splat: [ 66.460000] ================================= [ 66.460000] [ INFO: inconsistent lock state ] [ 66.460000] 3.9.0-rc5-00161-ga48dd49 #4 Not tainted [ 66.460000] --------------------------------- [ 66.460000] inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage. [ 66.460000] swapper/1 [HC0[0]:SC1[1]:HE1:SE0] takes: [ 66.460000] (timer_lock){+.?...}, at: [] rs_poll+0x12/0xdc [ 66.460000] {SOFTIRQ-ON-W} state was registered at: [ 66.460000] [] lock_acquire+0xec/0x13c [ 66.460000] [] _raw_spin_lock+0x3a/0x84 [ 66.460000] [] rs_open+0x18/0x58 [ 66.460000] [] tty_open+0x262/0x3cc [ 66.460000] [] chrdev_open+0x8c/0xe0 [ 66.460000] [] do_dentry_open$isra$16+0x10e/0x190 [ 66.460000] [] finish_open+0x39/0x48 [ 66.460000] [] do_last$isra$34+0x6c4/0x824 [ 66.460000] [] path_openat+0x66/0x310 [ 66.460000] [] do_filp_open+0x16/0x44 [ 66.460000] [] do_sys_open+0xd5/0x13c [ 66.460000] [] sys_open+0x12/0x18 [ 66.460000] [] kernel_init_freeable+0xe4/0x12c [ 66.460000] [] kernel_init+0xc/0x9c [ 66.460000] [] ret_from_kernel_thread+0x8/0xc [ 66.460000] irq event stamp: 132542 [ 66.460000] hardirqs last enabled at (132542): [] _raw_spin_unlock_irq+0x30/0x44 [ 66.460000] hardirqs last disabled at (132541): [] _raw_spin_lock_irq+0xe/0x8c [ 66.460000] softirqs last enabled at (132234): [] __do_softirq+0x216/0x2a4 [ 66.460000] softirqs last disabled at (132539): [] irq_exit+0x38/0x40 [ 66.460000] [ 66.460000] other info that might help us debug this: [ 66.460000] Possible unsafe locking scenario: [ 66.460000] [ 66.460000] CPU0 [ 66.460000] ---- [ 66.460000] lock(timer_lock); [ 66.460000] [ 66.460000] lock(timer_lock); [ 66.460000] [ 66.460000] *** DEADLOCK *** [ 66.460000] [ 66.460000] 1 lock held by swapper/1: [ 66.460000] #0: (((&serial_timer))){+.-...}, at: [] call_timer_fn+0x0/0x1f0 [ 66.460000] Stack: d7c2fac0 00000018 00000004 00000001 d7c2faa0 00000004 00000006 d7c2fa90 9003e87c d7c2fae0 d7c30000 d025a87c 00000001 0000000f 00000000 d7c2fac0 9004005d d7c2fb10 d7c30000 d7c30338 00000001 00000001 00000000 d7c30338 [ 66.460000] Call Trace: [ 66.460000] [] print_usage_bug$part$26+0x1c3/0x1c8 [ 66.460000] [] mark_lock+0x2b4/0x440 [ 66.460000] [] __lock_acquire+0x54d/0x16c4 [ 66.460000] [] lock_acquire+0xec/0x13c [ 66.460000] [] _raw_spin_lock+0x3a/0x84 [ 66.460000] [] rs_poll+0x12/0xdc [ 66.460000] [] call_timer_fn+0xbe/0x1f0 [ 66.460000] [] run_timer_softirq+0x198/0x1f4 [ 66.460000] [] __do_softirq+0x114/0x2a4 [ 66.460000] [] irq_exit+0x38/0x40 [ 66.460000] [] do_IRQ+0x44/0x48 [ 66.460000] [] do_interrupt+0x4c/0x54 [ 66.460000] [] common_exception_return+0x0/0x5c [ 66.460000] [] free_pcppages_bulk+0x254/0x308 [ 66.460000] Signed-off-by: Max Filippov Signed-off-by: Chris Zankel

xtensa: don't use a7 in simcalls

2013-05-09T08:07:10+00:00

To support FRAME_POINTER avoid using a7 in __simc (none of the existing simcalls needs it). Replace calls to __simc with more specific simc_read, simc_write and simc_lseek calls. Signed-off-by: Max Filippov Signed-off-by: Chris Zankel

TTY: switch tty_flip_buffer_push

2013-01-16T06:30:15+00:00

Now, we start converting tty buffer functions to actually use tty_port. This will allow us to get rid of the need of tty in many call sites. Only tty_port will needed and hence no more tty_port_tty_get in those paths. Now, the one where most of tty_port_tty_get gets removed: tty_flip_buffer_push. IOW we also closed all the races in drivers not using tty_port_tty_get at all yet. Also we move tty_flip_buffer_push declaration from include/linux/tty.h to include/linux/tty_flip.h to all others while we are changing it anyway. Signed-off-by: Jiri Slaby Signed-off-by: Greg Kroah-Hartman

TTY: switch tty_insert_flip_char

2013-01-16T06:21:36+00:00

Now, we start converting tty buffer functions to actually use tty_port. This will allow us to get rid of the need of tty in many call sites. Only tty_port will needed and hence no more tty_port_tty_get in those paths. tty_insert_flip_char is the next one to proceed. This one is used all over the code, so the patch is huge. Signed-off-by: Jiri Slaby Signed-off-by: Greg Kroah-Hartman

TTY: call tty_port_destroy in the rest of drivers

2012-11-16T01:20:58+00:00

After commit "TTY: move tty buffers to tty_port", the tty buffers are not freed in some drivers. This is because tty_port_destructor is not called whenever a tty_port is freed. This was an assumption I counted with but was unfortunately untrue. So fix the drivers to fulfil this assumption. To be sure, the TTY buffers (and later some stuff) are gone along with the tty_port, we have to call tty_port_destroy at tear-down places. This is mostly where the structure containing a tty_port is freed. This patch does exactly that -- put tty_port_destroy at those places. Signed-off-by: Jiri Slaby Signed-off-by: Greg Kroah-Hartman

xtensa: ISS: fix rs_put_char

2012-10-16T04:39:20+00:00

ISS serial console prints garbage instead of symbols printed via rs_put_char. gcc optimizes away putting prined symbol into memory buffer because there's no evidence that the buffer is used afterwards. Make rs_put_char and rs_write use simc_write that has explicit wmb. Signed-off-by: Max Filippov Signed-off-by: Chris Zankel

Merge tag 'xtensa-next-20121008' of git://github.com/czankel/xtensa-linux

2012-10-09T07:11:46+00:00

Pull Xtensa patchset from Chris Zankel: "The Xtensa tree has been broken for some time now, and this patchset brings it back to life. It has been part of the linux-next tree for some time. Most changes are inside the xtensa subdirectory; the other changes mostly add another rule to already existing #ifdefs to exclude Xtensa, where required. The only 'common' change is to add two more sections ('.xt.prop' and '.xt.lit') to the white list in modpost." * tag 'xtensa-next-20121008' of git://github.com/czankel/xtensa-linux: (27 commits) xtensa: Setup CROSS_COMPILE at the top xtensa: drop CONFIG_EMBEDDED_RAMDISK xtensa: fix TIOCGSERIAL and TIOCSSERIAL definitions xtensa: provide dummy gcc intrinsics xtensa: add missing symbol exports parport: disable for xtensa arch xtensa: rename MISC SR definition to avoid name clashes hisax: disable build for big-endian xtensa xtensa: fix CODA build xtensa: fix parallel make xtensa: ISS: drop unused io.c xtensa: ISS: exit simulator in case of halt or poweroff xtensa: ISS: change keyboard polling rate xtensa: ISS: add platform_pcibios_init xtensa: ISS: add dummy serial.h for ISS platform xtensa: change default platform clock frequency to 10MHz xtensa: add ARCH_WANT_OPTIONAL_GPIOLIB to xtensa config xtensa: set NO_IOPORT to 'n' by default xtensa: adopt generic io routines xtensa: fix ioremap ...

xtensa: ISS: change keyboard polling rate

2012-10-03T22:12:14+00:00

Polling rate of once per 20 seconds seems to be too low. Increase it to 10 times per second. Signed-off-by: Max Filippov Signed-off-by: Chris Zankel

xtensa: ISS: fix __simc implementation

2012-10-03T22:11:25+00:00

Do not rely on compiler implicitly placing function arguments into specifc registers, place them explicitly. Move __simc and related helpers to platform/simcall.h Signed-off-by: Max Filippov Signed-off-by: Chris Zankel

TTY: use tty_port_link_device

2012-08-13T23:50:19+00:00

So now for those drivers that can use neither tty_port_install nor tty_port_register_driver but still have tty_port available before tty_register_driver we use newly added tty_port_link_device. The rest of the drivers that still do not provide tty_struct <-> tty_port link will have to be converted to implement tty->ops->install. Signed-off-by: Jiri Slaby Acked-by: Alan Cox Signed-off-by: Greg Kroah-Hartman