28 Apr, 2007
1 commit
-
* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6: (448 commits)
[IPV4] nl_fib_lookup: Initialise res.r before fib_res_put(&res)
[IPV6]: Fix thinko in ipv6_rthdr_rcv() changes.
[IPV4]: Add multipath cached to feature-removal-schedule.txt
[WIRELESS] cfg80211: Clarify locking comment.
[WIRELESS] cfg80211: Fix locking in wiphy_new.
[WEXT] net_device: Don't include wext bits if not required.
[WEXT]: Misc code cleanups.
[WEXT]: Reduce inline abuse.
[WEXT]: Move EXPORT_SYMBOL statements where they belong.
[WEXT]: Cleanup early ioctl call path.
[WEXT]: Remove options.
[WEXT]: Remove dead debug code.
[WEXT]: Clean up how wext is called.
[WEXT]: Move to net/wireless
[AFS]: Eliminate cmpxchg() usage in vlocation code.
[RXRPC]: Fix pointers passed to bitops.
[RXRPC]: Remove bogus atomic_* overrides.
[AFS]: Fix u64 printing in debug logging.
[AFS]: Add "directory write" support.
[AFS]: Implement the CB.InitCallBackState3 operation.
...
27 Apr, 2007
9 commits
-
Signed-off-by: Martin Schwidefsky
Signed-off-by: Heiko Carstens -
The page_test_and_clear_dirty primitive really consists of two
operations, page_test_dirty and the page_clear_dirty. The combination
of the two is not an atomic operation, so it makes more sense to have
two separate operations instead of one.
In addition to the improved readability of the s390 version of
SetPageUptodate, it now avoids the page_test_dirty operation which is
an insert-storage-key-extended (iske) instruction which is an expensive
operation.Signed-off-by: Martin Schwidefsky
-
Generate uevents for all cpus if cpu capability changes. This can
happen e.g. because the cpus are overheating. The cpu capability can
be read via /sys/devices/system/cpu/cpuN/capability.Signed-off-by: Martin Schwidefsky
Signed-off-by: Heiko Carstens -
s390 machines provide hardware support for creating Linux dumps on SCSI
disks. For creating a dump a special purpose dump Linux is used. The first
32 MB of memory are saved by the hardware before the dump Linux is
booted. Via an SCLP interface, the saved memory can be accessed from
Linux. This patch exports memory and registers of the crashed Linux to
userspace via a debugfs file. For more information refer to
Documentation/s390/zfcpdump.txt, which is included in this patch.Signed-off-by: Michael Holzheu
Signed-off-by: Martin Schwidefsky
Signed-off-by: Heiko Carstens -
Generic bug implementation for s390. Will increase the value of the
console output on BUG() statements since registers r0-r5,r14 will
not be clobbered by a printk() call that was previously done before
the illegal instruction of BUG() was hit.
Also implements an architecture specific WARN_ON(). Output of that
could be increased but requires common code change.Signed-off-by: Martin Schwidefsky
Signed-off-by: Heiko Carstens -
This patch adds two improvements to the oops output. First it adds an
additional line after the PSW which decodes the different fields of it.
Second a disassembler is added that decodes the instructions surrounding
the faulting PSW. The output of a test oops now looks like this:kernel BUG at init/main.c:419
illegal operation: 0001 [#1]
CPU: 0 Not tainted
Process swapper (pid: 0, task: 0000000000464968, ksp: 00000000004be000)
Krnl PSW : 0700000180000000 00000000000120b6 (rest_init+0x36/0x38)
R:0 T:1 IO:1 EX:1 Key:0 M:0 W:0 P:0 AS:0 CC:0 PM:0 EA:3
Krnl GPRS: 0000000000000003 00000000004ba017 0000000000000022 0000000000000001
000000000003a5f6 0000000000000000 00000000004be6a8 0000000000000000
0000000000000000 00000000004b8200 0000000000003a50 0000000000008000
0000000000516368 000000000033d008 00000000000120b2 00000000004bdee0
Krnl Code: 00000000000120a6: e3e0f0980024 stg %r14,152(%r15)
00000000000120ac: c0e500014296 brasl %r14,3a5d8
00000000000120b2: a7f40001 brc 15,120b4
>00000000000120b6: 0707 bcr 0,%r7
00000000000120b8: eb7ff0500024 stmg %r7,%r15,80(%r15)
00000000000120be: c0d000195825 larl %r13,33d108
00000000000120c4: a7f13f00 tmll %r15,16128
00000000000120c8: a7840001 brc 8,120ca
Call Trace:
([] rest_init+0x32/0x38)
[] start_kernel+0x37c/0x410
[] _ehead+0x20/0x80Signed-off-by: Martin Schwidefsky
Signed-off-by: Heiko Carstens -
Introduce a mutex for struct ccwgroup to prevent simuntaneous
register/unregister on the same ccwgroup device.Signed-off-by: Cornelia Huck
Signed-off-by: Martin Schwidefsky -
Add a new attribute to the channel-path sysfs directory through which
channel-path configure operations can be triggered. Also listen for
hardware events requesting channel-path configure operations and
process them accordingly.Signed-off-by: Peter Oberparleiter
Signed-off-by: Martin Schwidefsky
Signed-off-by: Heiko Carstens -
Clean interface between cio and ipl code, so Peter stops complaining.
Signed-off-by: Martin Schwidefsky
Signed-off-by: Heiko Carstens
26 Apr, 2007
2 commits
-
Now that network timestamps use ktime_t infrastructure, we can add a new
SOL_SOCKET sockopt SO_TIMESTAMPNS.This command is similar to SO_TIMESTAMP, but permits transmission of
a 'timespec struct' instead of a 'timeval struct' control message.
(nanosecond resolution instead of microsecond)Control message is labelled SCM_TIMESTAMPNS instead of SCM_TIMESTAMP
A socket cannot mix SO_TIMESTAMP and SO_TIMESTAMPNS : the two modes are
mutually exclusive.sock_recv_timestamp() became too big to be fully inlined so I added a
__sock_recv_timestamp() helper function.Signed-off-by: Eric Dumazet
CC: linux-arch@vger.kernel.org
Signed-off-by: David S. Miller -
Now network timestamps use ktime_t infrastructure, we can add a new
ioctl() SIOCGSTAMPNS command to get timestamps in 'struct timespec'.
User programs can thus access to nanosecond resolution.Signed-off-by: Eric Dumazet
CC: Stephen Hemminger
Signed-off-by: David S. Miller
27 Mar, 2007
1 commit
-
git commit f994aae1bd8e4813d59a2ed64d17585fe42d03fc changed the
function declaration of csum_tcpudp_nofold. Argument types were
changed from unsigned long to __be32 (unsigned int). Therefore we
lost the implicit type conversion that zeroed the upper half of the
registers that are used to pass parameters. Since the inline assembly
relied on this we ended up adding random values and wrong checksums
were created.
Showed only up on machines with more than 4GB since gcc produced code
where the registers that are used to pass 'saddr' and 'daddr' previously
contained addresses before calling this function.
Fix this by using 32 bit arithmetics and convert code to C, since gcc
produces better code than these hand-optimized versions.Cc: Martin Schwidefsky
Signed-off-by: Heiko Carstens
19 Mar, 2007
2 commits
-
Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky -
We used wrong length values for ipl and dump hardware structures.
Since z/VM checks the ipl parameters more accurately than LPAR,
the operations fail there.Signed-off-by: Michael Holzheu
Signed-off-by: Martin Schwidefsky
06 Mar, 2007
2 commits
-
Don't have functions in header files unless they are inline.
Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky -
Reipl doesn't work on older machines were s390_reset_machine() gets
called. The reason is that the text section is read-only but the
variable dump_prefix_page is there. Since s390_reset_machine() writes
to it we get a protection exception.
Therefore move dump_prefix_page to the bss section.Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky
21 Feb, 2007
5 commits
-
Signed-off-by: Mathieu Desnoyers
Cc: Heiko Carstens
Signed-off-by: Andrew Morton
Signed-off-by: Martin Schwidefsky -
Signed-off-by: Mathieu Desnoyers
Cc: Heiko Carstens
Signed-off-by: Andrew Morton
Signed-off-by: Martin Schwidefsky -
With CONFIG_SHARED_KERNEL the kernel text segment that might be in a
read only memory sections starts at 1MB. Memory between 0x12000 and
0x100000 is unused then. Free this, so we have appr. an extra MB
of memory available.Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky -
Setup.h has been misused for ipl related stuff in the past. We now move
everything, which has to do with ipl and reipl to a new header file named
"ipl.h".Signed-off-by: Michael Holzheu
Signed-off-by: Martin Schwidefsky -
Replace two stidp inline assemblies with one global implementation.
Signed-off-by: Michael Holzheu
Signed-off-by: Martin Schwidefsky
12 Feb, 2007
4 commits
-
Signed-off-by: Martin Schwidefsky
-
The current ccw_device_set_options() sets a specified mask of options
and clears those not specified, but there is no way to find out which
options have already been set.In order to fix this up, introduce the following interface changes:
ccw_device_set_options() now only sets the specified bits, but does
not clear those that are not specified.ccw_device_clear_options() clears the specified bits.
ccw_device_set_options_mask() provides the old semantics (setting only
the specified bits and clearing the others).Device drivers now work as expected. qdio has been adapted.
Signed-off-by: Cornelia Huck
Signed-off-by: Martin Schwidefsky -
Martin Schwidefsky wrote:
"s390 does not even need (in|out)b(_p|). I wondered what else from
io.h do we not need. The answer is: almost nothing. With the devres
patch from Al and the dma-mapping patch from Heiko we can get rid of
iomem and all associated definitions."So we'll just need to replace NO_IOPORT with NO_IOMEM in Kconfig and
kill arch/s390/mm/ioremap.c.BTW, there's an annoying bit of junk in there - IO_SPACE_LIMIT. We
only need it for /proc/ioports, which AFAICS shouldn't even be there
on s390 (or uml). OTOH, removing that thing would mean a user-visible
change - we go from "empty file in /proc" to "no such file in /proc"...Signed-off-by: Al Viro
Signed-off-by: Linus Torvalds -
The line discipline numbers N_* are currently defined for each architecture
individually, but (except for a seeming mistake) identically, in
asm/termios.h. There is no obvious reason why these numbers should be
architecture specific, nor any apparent relationship with the termios
structure. The total number of these, NR_LDISCS, is defined in linux/tty.h
anyway. So I propose the following patch which moves the definitions of
the individual line disciplines to linux/tty.h too.Three of these numbers (N_MASC, N_PROFIBUS_FDL, and N_SMSBLOCK) are unused
in the current kernel, but the patch still keeps the complete set in case
there are plans to use them yet.Signed-off-by: Tilman Schmidt
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
06 Feb, 2007
12 commits
-
Set read-only flag in the page table entries for the kernel image text
section. This will catch all instruction caused corruptions withing the
text section.
Instruction replacement via kprobes still works, since it bypasses now
dynamic address translation.Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky -
Hopefully this will make it more maintainable and less error prone.
Code makes use of search_exception_tables(). Since it calls this
function before the kernel exeception table is sorted, there is an
early call to sort_main_extable().This way it's easy to use the already present infrastructure of fixup
sections. Also this would allows to easily convert the rest of
head[31|64].S into C code.Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky -
Preset the bogomips number to the cpu capacity value reported by
store system information in SYSIB 1.2.2. This value is constant
for a particular machine model and can be used to determine
relative performance differences between machines.Signed-off-by: Martin Schwidefsky
-
3592 tape devices are able to write data encrpyted on tape mediums.
This z/Linux device driver support includes the following functions:
* ioctl to switch on/off encryption
* ioctl to query encryption status of drive
* ioctls to set and query key encrypting keys (kekls)
* long busy interrupt handlingSigned-off-by: Michael Holzheu
Signed-off-by: Martin Schwidefsky -
Add support to boot from a named saved segment (NSS).
Signed-off-by: Hongjie Yang
Signed-off-by: Martin Schwidefsky -
Starting with the z9 the CPU Cryptographic Assist Facility comes with
an integrated Pseudo Random Number Generator. The generator creates
random numbers by an algorithm similar to the ANSI X9.17 standard.
The pseudo-random numbers can be accessed via a character device driver
node called /dev/prandom. Similar to /dev/urandom any amount of bytes
can be read from the device without blocking.Signed-off-by: Jan Glauber
Signed-off-by: Martin Schwidefsky -
This patch adds support for clock synchronization to an external time
reference (ETR). The external time reference sends an oscillator
signal and a synchronization signal every 2^20 microseconds to keep
the TOD clocks of all connected servers in sync. For availability
two ETR units can be connected to a machine. If the clock deviates
for more than the sync-check tolerance all cpus get a machine check
that indicates that the clock is out of sync. For the lovely details
how to get the clock back in sync see the code below.Signed-off-by: Martin Schwidefsky
-
This provides a noexec protection on s390 hardware. Our hardware does
not have any bits left in the pte for a hw noexec bit, so this is a
different approach using shadow page tables and a special addressing
mode that allows separate address spaces for code and data.As a special feature of our "secondary-space" addressing mode, separate
page tables can be specified for the translation of data addresses
(storage operands) and instruction addresses. The shadow page table is
used for the instruction addresses and the standard page table for the
data addresses.
The shadow page table is linked to the standard page table by a pointer
in page->lru.next of the struct page corresponding to the page that
contains the standard page table (since page->private is not really
private with the pte_lock and the page table pages are not in the LRU
list).
Depending on the software bits of a pte, it is either inserted into
both page tables or just into the standard (data) page table. Pages of
a vma that does not have the VM_EXEC bit set get mapped only in the
data address space. Any try to execute code on such a page will cause a
page translation exception. The standard reaction to this is a SIGSEGV
with two exceptions: the two system call opcodes 0x0a77 (sys_sigreturn)
and 0x0aad (sys_rt_sigreturn) are allowed. They are stored by the
kernel to the signal stack frame. Unfortunately, the signal return
mechanism cannot be modified to use an SA_RESTORER because the
exception unwinding code depends on the system call opcode stored
behind the signal stack frame.This feature requires that user space is executed in secondary-space
mode and the kernel in home-space mode, which means that the addressing
modes need to be switched and that the noexec protection only works
for user space.
After switching the addressing modes, we cannot use the mvcp/mvcs
instructions anymore to copy between kernel and user space. A new
mvcos instruction has been added to the z9 EC/BC hardware which allows
to copy between arbitrary address spaces, but on older hardware the
page tables need to be walked manually.Signed-off-by: Gerald Schaefer
Signed-off-by: Martin Schwidefsky -
Currently works anyway since search_binary_handler has a
set_fs(USER_DS). But start_thread() is the place where this should be
done. Following all other architectures...Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky -
cpu_relax() has barrier() semantics hence there is no need to use both
of them in conjunction in sclp_sync_wait(). Also change cpu_relax()
so it's more obvious that it has barrier semantics.Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky -
No need to use lrag in 64 bit addressing mode since lra will do the
same.Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky -
Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky
12 Jan, 2007
1 commit
-
unionfs managed to hit this on s390. Some architectures use __ptr_t in their
FD_ZERO implementation. We don't have a __ptr_t. Switch them over to plain
old void*.Cc: Richard Henderson
Cc: Ivan Kokshaysky
Cc: Al Viro
Cc: Paul Mackerras
Cc: Benjamin Herrenschmidt
Cc: "Luck, Tony"
Cc: Martin Schwidefsky
Cc: Heiko Carstens
Cc: Josef 'Jeff' Sipek
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
09 Jan, 2007
1 commit
-
There are several places in the futex code where a spin_lock is held
and still uaccesses happen. Deadlocks are avoided by increasing the
preempt count. The pagefault handler will then not take any locks
but will immediately search the fixup tables.Signed-off-by: Heiko Carstens
Signed-off-by: Martin Schwidefsky