Skip to content

GitLab

Switch branch/tag
  1. 25 Dec, 2008 10 commits
  3. 27 Nov, 2008 3 commits
    • Heiko Carstens's avatar
      [S390] Fix alignment of initial kernel stack. · 0778dc3a
      Heiko Carstens authored 
      

We need an alignment of 16384 bytes for the initial kernel stack if
the kernel is configured for 16384 bytes stacks but the linker script
currently guarantees only an alignment of 8192 bytes.

So fix this and simply use THREAD_SIZE as alignment value which will
always do the right thing.
Signed-off-by: default avatarHeiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: default avatarMartin Schwidefsky <schwidefsky@de.ibm.com>
      0778dc3a
    • Christian Borntraeger's avatar
      [S390] fix/cleanup sched_clock · 8107d829
      Christian Borntraeger authored 
      

CONFIG_PRINTK_TIME reveals that sched_clock has a wrong offset during boot:
..
[    0.000000]   Movable zone: 0 pages used for memmap
[    0.000000] Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 775679
[    0.000000] Kernel command line: dasd=4b6c root=/dev/dasda1 ro noinitrd
[    0.000000] PID hash table entries: 4096 (order: 12, 32768 bytes)
[6920575.975232] console [ttyS0] enabled
[6920575.987586] Dentry cache hash table entries: 524288 (order: 10, 4194304 bytes)
[6920575.991404] Inode-cache hash table entries: 262144 (order: 9, 2097152 bytes)
..

The s390 implementation of sched_clock uses the store clock instruction and
subtracts jiffies_timer_cc.
jiffies_timer_cc is a local variable in arch/s390/kernel/time.c and only used
for sched_clock and monotonic clock. For historical reasons there is an offset
on that value. With todays code this offset is unnecessary. By removing that
offset we can get a sched_clock which returns the nanoseconds after time_init.
This improves CONFIG_PRINTK_TIME.

Since sched_clock is the only user, I have also renamed jiffies_timer_cc to
sched_clock_base_cc. In addition, the local variable init_timer_cc is redundant
and can be romved as well.
Signed-off-by: default avatarChristian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: default avatarMartin Schwidefsky <schwidefsky@de.ibm.com>
      8107d829
    • Martin Schwidefsky's avatar
      [S390] fix system call parameter functions. · 59da2139
      Martin Schwidefsky authored 
      

syscall_get_nr() currently returns a valid result only if the call
chain of the traced process includes do_syscall_trace_enter(). But
collect_syscall() can be called for any sleeping task, the result of
syscall_get_nr() in general is completely bogus.

To make syscall_get_nr() work for any sleeping task the traps field
in pt_regs is replace with svcnr - the system call number the process
is executing. If svcnr == 0 the process is not on a system call path.

The syscall_get_arguments and syscall_set_arguments use regs->gprs[2]
for the first system call parameter. This is incorrect since gprs[2]
may have been overwritten with the system call number if the call
chain includes do_syscall_trace_enter. Use regs->orig_gprs2 instead.
Signed-off-by: default avatarMartin Schwidefsky <schwidefsky@de.ibm.com>
      59da2139
  5. 14 Nov, 2008 6 commits
  7. 28 Oct, 2008 1 commit
    • Heiko Carstens's avatar
      [S390] Fix sysdev class file creation. · da5aae70
      Heiko Carstens authored 
      

Use sysdev_class_create_file() to create create sysdev class attributes
instead of sysfs_create_file(). Using sysfs_create_file() wasn't a very
good idea since the show and store functions have a different amount of
parameters for sysfs files and sysdev class files.
In particular the pointer to the buffer is the last argument and
therefore accesses to random memory regions happened.
Still worked surprisingly well until we got a kernel panic.

Cc: stable@kernel.org
Signed-off-by: default avatarHeiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: default avatarMartin Schwidefsky <schwidefsky@de.ibm.com>
      da5aae70
  9. 16 Oct, 2008 5 commits
  11. 10 Oct, 2008 3 commits
  13. 03 Oct, 2008 1 commit
    • Heiko Carstens's avatar
      [S390] nohz: Fix __udelay. · d3d238c7
      Heiko Carstens authored 
      This fixes a regression that came with 934b2857


("[S390] nohz/sclp: disable timer on synchronous waits.").
If udelay() gets called from a disabled context it sets the clock comparator
to a value where it expects the next interrupt. When the interrupt happens
the clock comparator gets not reset and therefore the interrupt condition
doesn't get cleared. The result is an endless timer interrupt loop.

In addition this patch fixes also the following:

rcutorture reveals that our __udelay implementation is still buggy,
since it might schedule tasklets, but prevents their execution:

NOHZ: local_softirq_pending 42
NOHZ: local_softirq_pending 02
NOHZ: local_softirq_pending 142
NOHZ: local_softirq_pending 02

To fix this we make sure that only the clock comparator interrupt
is enabled when the enabled wait psw is loaded.
Also no code gets called anymore which might schedule tasklets.
Signed-off-by: default avatarHeiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: default avatarMartin Schwidefsky <schwidefsky@de.ibm.com>
      d3d238c7
  15. 09 Sep, 2008 1 commit
  17. 08 Sep, 2008 1 commit
    • Manfred Spraul's avatar
      kernel/cpu.c: create a CPU_STARTING cpu_chain notifier · e545a614
      Manfred Spraul authored 
      

Right now, there is no notifier that is called on a new cpu, before the new
cpu begins processing interrupts/softirqs.
Various kernel function would need that notification, e.g. kvm works around
by calling smp_call_function_single(), rcu polls cpu_online_map.

The patch adds a CPU_STARTING notification. It also adds a helper function
that sends the message to all cpu_chain handlers.

Tested on x86-64.
All other archs are untested. Especially on sparc, I'm not sure if I got
it right.
Signed-off-by: default avatarManfred Spraul <manfred@colorfullife.com>
Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      e545a614
  19. 25 Aug, 2008 1 commit
  21. 21 Aug, 2008 2 commits
  23. 01 Aug, 2008 4 commits
  25. 27 Jul, 2008 1 commit
  27. 25 Jul, 2008 1 commit