linux-toradex.git/arch/s390/kernel, branch v3.4.78

s390: critical section cleanup vs. machine checks

2013-03-20T20:05:00+00:00

commit 6551fbdfd8b85d1ab5822ac98abb4fb449bcfae0 upstream.

The current machine check code uses the registers stored by the machine
in the lowcore at __LC_GPREGS_SAVE_AREA as the registers of the interrupted
context. The registers 0-7 of a user process can get clobbered if a machine
checks interrupts the execution of a critical section in entry[64].S.

The reason is that the critical section cleanup code may need to modify
the PSW and the registers for the previous context to get to the end of a
critical section. If registers 0-7 have to be replaced the relevant copy
will be in the registers, which invalidates the copy in the lowcore. The
machine check handler needs to explicitly store registers 0-7 to the stack.

Signed-off-by: Martin Schwidefsky 
Signed-off-by: Greg Kroah-Hartman

s390/timer: avoid overflow when programming clock comparator

2013-02-17T18:49:25+00:00

commit d911e03d097bdc01363df5d81c43f69432eb785c upstream.

Since ed4f209 "s390/time: fix sched_clock() overflow" a new helper function
is used to avoid overflows when converting TOD format values to nanosecond
values.
The kvm interrupt code formerly however only worked by accident because of
an overflow. It tried to program a timer that would expire in more than ~29
years. Because of the old TOD-to-nanoseconds overflow bug the real expiry
value however was much smaller, but now it isn't anymore.
This however triggers yet another bug in the function that programs the clock
comparator s390_next_ktime(): if the absolute "expires" value is after 2042
this will result in an overflow and the programmed value is lower than the
current TOD value which immediatly triggers a clock comparator (= timer)
interrupt.
Since the timer isn't expired it will be programmed immediately again and so
on... the result is a dead system.
To fix this simply program the maximum possible value if an overflow is
detected.

Reported-by: Christian Borntraeger 
Tested-by: Christian Borntraeger 
Signed-off-by: Heiko Carstens 
Signed-off-by: Greg Kroah-Hartman

s390/time: fix sched_clock() overflow

2013-01-21T19:45:15+00:00

commit ed4f20943cd4c7b55105c04daedf8d63ab6d499c upstream.

Converting a 64 Bit TOD format value to nanoseconds means that the value
must be divided by 4.096. In order to achieve that we multiply with 125
and divide by 512.
When used within sched_clock() this triggers an overflow after appr.
417 days. Resulting in a sched_clock() return value that is much smaller
than previously and therefore may cause all sort of weird things in
subsystems that rely on a monotonic sched_clock() behaviour.

To fix this implement a tod_to_ns() helper function which converts TOD
values without overflow and call this function from both places that
open coded the conversion: sched_clock() and kvm_s390_handle_wait().

Reviewed-by: Martin Schwidefsky 
Signed-off-by: Heiko Carstens 
Signed-off-by: Martin Schwidefsky 
Signed-off-by: Greg Kroah-Hartman

s390/signal: set correct address space control

2012-11-26T19:38:01+00:00

commit fa968ee215c0ca91e4a9c3a69ac2405aae6e5d2f upstream.

If user space is running in primary mode it can switch to secondary
or access register mode, this is used e.g. in the clock_gettime code
of the vdso. If a signal is delivered to the user space process while
it has been running in access register mode the signal handler is
executed in access register mode as well which will result in a crash
most of the time.

Set the address space control bits in the PSW to the default for the
execution of the signal handler and make sure that the previous
address space control is restored on signal return. Take care
that user space can not switch to the kernel address space by
modifying the registers in the signal frame.

Signed-off-by: Martin Schwidefsky 
Signed-off-by: Greg Kroah-Hartman

s390: fix linker script for 31 bit builds

2012-10-28T17:14:13+00:00

commit c985cb37f1b39c2c8035af741a2a0b79f1fbaca7 upstream.

Because of a change in the s390 arch backend of binutils (commit 23ecd77
"Pick the default arch depending on the target size" in binutils repo)
31 bit builds will fail since the linker would now try to create 64 bit
binary output.
Fix this by setting OUTPUT_ARCH to s390:31-bit instead of s390.
Thanks to Andreas Krebbel for figuring out the issue.

Fixes this build error:

  LD      init/built-in.o
s390x-4.7.2-ld: s390:31-bit architecture of input file
 `arch/s390/kernel/head.o' is incompatible with s390:64-bit output

Cc: Andreas Krebbel 
Signed-off-by: Heiko Carstens 
Signed-off-by: Martin Schwidefsky 
Signed-off-by: Greg Kroah-Hartman

s390/compat: fix mmap compat system calls

2012-08-26T22:00:39+00:00

commit e85871218513c54f7dfdb6009043cb638f2fecbe upstream.

The native 31 bit and the compat behaviour for the mmap system calls differ:

In native 31 bit mode the passed in address for the mmap system call will be
unmodified passed to sys_mmap_pgoff().
In compat mode however the passed in address will be modified with
compat_ptr() which masks out the most significant bit.

The result is that in native 31 bit mode each mmap request (with MAP_FIXED)
will fail where the most significat bit is set, while in compat mode it
may succeed.

This odd behaviour was introduced with d3815898 "[S390] mmap: add missing
compat_ptr conversion to both mmap compat syscalls".

To restore a consistent behaviour accross native and compat mode this
patch functionally reverts the above mentioned commit.

Signed-off-by: Heiko Carstens 
Signed-off-by: Martin Schwidefsky 
Signed-off-by: Greg Kroah-Hartman

s390/compat: fix compat wrappers for process_vm system calls

2012-08-26T22:00:39+00:00

commit 82aabdb6f1eb61e0034ec23901480f5dd23db7c4 upstream.

The compat wrappers incorrectly called the non compat versions of
the system process_vm system calls.

Signed-off-by: Heiko Carstens 
Signed-off-by: Martin Schwidefsky 
Signed-off-by: Greg Kroah-Hartman

s390/idle: fix sequence handling vs cpu hotplug

2012-08-09T15:31:30+00:00

commit 0008204ffe85d23382d6fd0f971f3f0fbe70bae2 upstream.

The s390 idle accounting code uses a sequence counter which gets used
when the per cpu idle statistics get updated and read.

One assumption on read access is that only when the sequence counter is
even and did not change while reading all values the result is valid.
On cpu hotplug however the per cpu data structure gets initialized via
a cpu hotplug notifier on CPU_ONLINE.
CPU_ONLINE however is too late, since the onlined cpu is already running
and might access the per cpu data. Worst case is that the data structure
gets initialized while an idle thread is updating its idle statistics.
This will result in an uneven sequence counter after an update.

As a result user space tools like top, which access /proc/stat in order
to get idle stats, will busy loop waiting for the sequence counter to
become even again, which will never happen until the queried cpu will
update its idle statistics again. And even then the sequence counter
will only have an even value for a couple of cpu cycles.

Fix this by moving the initialization of the per cpu idle statistics
to cpu_init(). I prefer that solution in favor of changing the
notifier to CPU_UP_PREPARE, which would be a different solution to
the problem.

Signed-off-by: Heiko Carstens 
Signed-off-by: Martin Schwidefsky 
Signed-off-by: Greg Kroah-Hartman

[S390] Fix stfle() lowcore protection problem

2012-04-11T12:28:26+00:00

The stfle() function writes into lowcore memory when stfl_fac_list
is initialized with "S390_lowcore.stfl_fac_list = 0". For older
compilers this triggers a lowcore exception. With newer compilers
and "-OXX" compile option the bug does not show up because
the "S390_lowcore.stfl_fac_list" initialization is removed by the
compiler. The reason for thatis the incorrect "=m"
(S390_lowcore.stfl_fac_list) constraint in the stfl inline assembly.

The following shows the disassembly of the stfle() optimized code
that is inlined in the lgr_info_get() function:

000000000011325c :
  11325c:       eb 9f f0 60 00 24       stmg    %r9,%r15,96(%r15)
  113262:       c0 d0 00 29 0e 47       larl    %r13,634ef0 
  113268:       a7 f1 3f c0             tml     %r15,16320
  11326c:       b9 04 00 ef             lgr     %r14,%r15
  113270:       a7 84 00 01             je      113272 
  113274:       a7 fb ff c0             aghi    %r15,-64
  113278:       b9 04 00 c2             lgr     %r12,%r2
  11327c:       a7 29 00 01             lghi    %r2,1
  113280:       e3 e0 f0 98 00 24       stg     %r14,152(%r15)
  113286:       d7 97 c0 00 c0 00       xc      0(152,%r12),0(%r12)
  11328c:       c0 e5 00 28 db 4c       brasl   %r14,62e924 
  113292:       b2 b1 00 00             stfl    0

To fix the problem we now clear the S390_lowcore.stfl_fac_list at
startup in "head.S" for all machine types before lowcore protection
is enabled.

In addition to that the "=m" constraint is replaced by "+m".

Signed-off-by: Michael Holzheu 
Signed-off-by: Martin Schwidefsky

[S390] cpum_cf: get rid of compile warnings

2012-04-11T12:28:25+00:00

Fix these:

arch/s390/kernel/perf_cpum_cf.c:180:3: warning: format '%lx'
   expects argument of type 'long unsigned int',
   but argument 2 has type 'int' [-Wformat]
arch/s390/kernel/perf_cpum_cf.c: In function 'cpumf_pmu_disable':
arch/s390/kernel/perf_cpum_cf.c:205:3: warning: format '%lx'
   expects argument of type 'long unsigned int',
   but argument 2 has type 'int' [-Wformat]

Signed-off-by: Heiko Carstens 
Signed-off-by: Martin Schwidefsky