- 11 May, 2016 1 commit
-
-
Heiko Carstens authored
The identity mapping is suboptimal for the last 2GB frame. The mapping will be established with a mix of 4KB and 1MB mappings instead of a single 2GB mapping. This happens because of a off-by-one bug introduced with commit 50be6345 ("s390/mm: Convert bootmem to memblock"). Currently the identity mapping looks like this: 0x0000000080000000-0x0000000180000000 4G PUD RW 0x0000000180000000-0x00000001fff00000 2047M PMD RW 0x00000001fff00000-0x0000000200000000 1M PTE RW With the bug fixed it looks like this: 0x0000000080000000-0x0000000200000000 6G PUD RW Fixes: 50be6345 ("s390/mm: Convert bootmem to memblock") Signed-off-by:
Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by:
Martin Schwidefsky <schwidefsky@de.ibm.com>
-
- 10 May, 2016 7 commits
-
-
Heiko Carstens authored
arch_mmap_rnd, cpu_have_feature, and arch_randomize_brk are all defined as globally visible variables. However the files they are defined in do not include the header files with the declaration. To avoid a possible mismatch add the missing include statements so we have proper type checking in place. Signed-off-by:
-
Heiko Carstens authored
arch_dup_task_struct and the per cpu variable mt_cycles are globally visible, but do not have any header file with a declaration. Therefore add it so we have proper type checking in place. Signed-off-by:
-
Heiko Carstens authored
copy_oldmem_user() and ap_jumptable are private to the files they are being used in. Therefore make them static. Signed-off-by:
-
Heiko Carstens authored
With "s390/cpuinfo: simplify locking and skip offline cpus early" we prevent already that cpus will go away. The additional get_online_cpus() / put_online_cpus() within show_cacheinfo() is not needed anymore. Signed-off-by:
-
Heiko Carstens authored
Move the get_online_cpus() and put_online_cpus() to the start and stop operation of the seqfile ops. This way there is no need to lock cpu hotplug again and again for each single cpu. This way we can also skip offline cpus early if we simply use cpumask_next() within the next operation. Signed-off-by:
-
Martin Schwidefsky authored
The disconnect and later reconnect of a 3270 terminal does not do a tty hangup. The current session is resumed which is a security issue. Do a tty hangup after a unit check has been received. Signed-off-by: -
Martin Schwidefsky authored
If an existing 3270 terminal disconnects and reconnects with a different size, the 3270 driver still works with the old size. If the new dimensions are larger the output merely looks funny, if the new dimensions are smaller the terminal is unusable. To fix this restart the size detection after a unit check has been received. Signed-off-by:
-
- 04 May, 2016 5 commits
-
-
Martin Schwidefsky authored
If a 3270 terminal is disconnected while the tty view is active the 3270 driver goes into an endless loop of failed I/O requests until the terminal is connected again. Add code to the raw3270 interrupt handler to check for unit checks due to failed I/O requests and put the device to sleep with the RAW3270_FLAGS_BUSY flag until a unsolicited device end interrupt indicates that the device can be used again. while we are at it simplify the 3270 irq handling and remove unnecessary code. Signed-off-by: -
Martin Schwidefsky authored
The tty3270_update function tries to optimize the output stream by skipping the TO_SBA command to position the output offset if the TO_RA command of the previous line ended at the same offset. But some lines do not have the TO_RA at the end, namely those with a length equal to the number of columns. Make sure there is a TO_RA at the end of the previous line to avoid funny looking output. Signed-off-by: -
Martin Schwidefsky authored
The tty3270_close function set tty->driver_data to NULL after the open count dropped to zero but tty3270_cleanup still needs the pointer to the tty3270 view to drop the reference count. Move the clearing of tty->driver_data to tty3270_cleanup and drop the unnecessary tty_port_tty_set calls in tty3270_install. Signed-off-by: -
Martin Schwidefsky authored
The tty3270_resize_work functions acquires a reference to the tty to call tty_do_resize but fails to release the reference again. Add the missing tty_kref_put. Signed-off-by: -
Heiko Carstens authored
Implement return_address() and use it instead of __builtin_return_address(n). __builtin_return_address(n) is not guaranteed to work for n > 0, therefore implement a private return_address() function which walks the stack frames and returns the proper return address. This way we get also rid of a compile warning which gcc 6.1 emits and look like all other architectures. Signed-off-by:
-
- 03 May, 2016 2 commits
-
-
Anna-Maria Gleixner authored
Since commit 3b9d6da6 ("cpu/hotplug: Fix rollback during error-out in __cpu_disable()") it is ensured that callbacks of CPU_ONLINE and CPU_DOWN_PREPARE are processed on the hotplugged CPU. Due to this SMP function calls are no longer required. Replace smp_call_function_single() with a direct call of setup_pmc_cpu(). To keep the calling convention, interrupts are explicitly disabled around the call. Cc: linux-s390@vger.kernel.org Signed-off-by:
Anna-Maria Gleixner <anna-maria@linutronix.de> Acked-by:
-
Anna-Maria Gleixner authored
Since commit 3b9d6da6 ("cpu/hotplug: Fix rollback during error-out in __cpu_disable()") it is ensured that callbacks of CPU_ONLINE and CPU_DOWN_PREPARE are processed on the hotplugged CPU. Due to this SMP function calls are no longer required. Replace smp_call_function_single() with a direct call of setup_pmc_cpu(). To keep the calling convention, interrupts are explicitly disabled around the call. Cc: linux-s390@vger.kernel.org Signed-off-by:
-
- 21 Apr, 2016 3 commits
-
-
Martin Schwidefsky authored
The current default processor type is z900. The BPF jit compiler depends on PACK_STACK && HAVE_MARCH_Z196_FEATURES. To have the BPF jit code included in compiles with 'make allmodconfig' set the default processor type to z196. Signed-off-by: -
Martin Schwidefsky authored
Fix the initialization of a local variable to remove this warning: All warnings (new ones prefixed by >>): drivers/s390/char/sclp_pci.c: In function 'sclp_pci_report': >> drivers/s390/char/sclp_pci.c:127:9: warning: missing braces around initializer [-Wmissing-braces] struct sclp_req req = {0}; ^ drivers/s390/char/sclp_pci.c:127:9: warning: (near initialization for 'req.list') [-Wmissing-braces] Fixes: 12283a40 "s390/sclp: add error notification command" Reported-by:kbuild test robot <fengguang.wu@intel.com> Signed-off-by:
-
Martin Schwidefsky authored
Analog to git commit 0c8c0f03 "x86/fpu, sched: Dynamically allocate 'struct fpu'" move the struct fpu to the end of the struct thread_struct, set CONFIG_ARCH_WANTS_DYNAMIC_TASK_STRUCT and add the setup_task_size() function to calculate the correct size fo the task struct. For the performance_defconfig this increases the size of struct task_struct from 7424 bytes to 7936 bytes (MACHINE_HAS_VX==1) or 7552 bytes (MACHINE_HAS_VX==0). The dynamic allocation of the struct fpu is removed. The slab cache uses an 8KB block for the task struct in all cases, there is enough room for the struct fpu. For MACHINE_HAS_VX==1 each task now needs 512 bytes less memory. Signed-off-by:
-
- 15 Apr, 2016 13 commits
-
-
Martin Schwidefsky authored
The CPACF instructions are going be used in KVM as well, move the defines and the inline functions from arch/s390/crypt/crypt_s390.h to arch/s390/include/asm. Rename the header to cpacf.h and replace the crypt_s390_xxx names with cpacf_xxx. While we are at it, cleanup the header as well. The encoding for the CPACF operations is odd, there is an enum for each of the CPACF instructions with the hardware function code in the lower 8 bits of each entry and a software defined number for the CPACF instruction in the upper 8 bits. Remove the superfluous software number and replace the enums with simple defines. The crypt_s390_func_available() function tests for the presence of a specific CPACF operations. The new name of the function is cpacf_query and it works slightly different than before. It gets passed an opcode of an CPACF instruction and a function code for this instruction. The facility_mask parameter is gone, the opcode is used to find the correct MSA facility bit to check if the CPACF instruction itself is available. If it is the query function of the given instruction is used to test if the requested CPACF operation is present. Acked-by:
David Hildenbrand <dahi@linux.vnet.ibm.com> Signed-off-by:
-
Sebastian Ott authored
Implement the function type specific function measurement block used in new machines. Signed-off-by:
Sebastian Ott <sebott@linux.vnet.ibm.com> Signed-off-by:
-
Jan Höppner authored
Implement new DASD IOCTL BIODASDCHECKFMT to check a range of tracks on a DASD volume for correct formatting. The following characteristics are checked: - Block size - ECKD key length - ECKD record ID - Number of records per track Signed-off-by:
Jan Höppner <hoeppner@linux.vnet.ibm.com> Signed-off-by:
-
Sebastian Ott authored
Sort the sclp event type defines and use a macro to create the corresponding event type masks. In addition to that one unused type/mask pair is removed and another previously unused define is used now (it was probably unused/unknown because it didn't follow the EVTYP_X EVTYP_X_MASK convention). Signed-off-by:
-
Sebastian Ott authored
Provide an report_error attribute to send an adapter-error notification associated with a PCI function. Signed-off-by:
-
Sebastian Ott authored
Add SCLP event 24 "Adapter-error notification". Signed-off-by:
-
Sebastian Ott authored
sclp commands only used by the PCI code shouldn't be build for !CONFIG_PCI. Instead of adding more ifdefs to sclp_cmd.c just move them to a new file. Signed-off-by:
-
Sascha Silbe authored
Commit 6c2ff84e979a ("s390: add SMT support") added the nosmt and smt early kernel parameters for restricting the use of symmetric multithreading (SMT). They are quite useful for debugging and testing, so document them. Signed-off-by:Sascha Silbe <silbe@linux.vnet.ibm.com> Signed-off-by:
-
Peter Zijlstra authored
While looking at set_task_state() users I stumbled over the s390 pfault interrupt code. Since Heiko provided a great explanation on how it worked, I figured we ought to preserve this. Also make a few little tweaks to the code to aid in readability and explicitly comment the unusual blocking scheme. Based-on-text-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by:
-
Stefan Haberland authored
With this feature, applications can query if a DASD volume is online to another operating system instances by checking the online status of all attached hosts from the storage server. Reviewed-by:
Stefan Haberland <stefan.haberland@de.ibm.com> Signed-off-by:
-
Heiko Carstens authored
Make sure that s390 appears to be a big endian machine by defining this config option. Without this s390 appears to be little endian as seen by e.g. the recordmount script: "perl ./scripts/recordmcount.pl "s390" "little" "64"" This has no practical impact within the script since the endian variable is only evaluated for mips. However there are already a couple of common code places which evaluate this config option. None of them is relevant for s390 currently though. To avoid any issues in the future (and fix the recordmcount oddity) add the new config option. Signed-off-by:
-
Heiko Carstens authored
arch_spin_lock_wait_flags() checks if a spinlock is not held before trying a compare and swap instruction. If the lock is unlocked it tries the compare and swap instruction, however if a different cpu grabbed the lock in the meantime the instruction will fail as expected. Subsequently the arch_spin_lock_wait_flags() incorrectly tries to figure out if the cpu that holds the lock is running. However it is using the wrong cpu number for this (-1) and then will also yield the current cpu to the wrong cpu. Fix this by adding a missing continue statement. Fixes: 470ada6b ("s390/spinlock: refactor arch_spin_lock_wait[_flags]") Signed-off-by:
-
Gerald Schaefer authored
dcssblk_remove_store() holds the dcssblk_devices_sem semaphore while calling device_unregister(), which in turn tries to acquire the kernfs kn->dev_map rwsem for the device sysfs subtree. The same rwsem is also acquired when using the per-device sysfs attributes in the device sub-tree, and the attribute handlers then also acquire the dcssblk_devices_sem. This can lead to a deadlock when removing a DCSS while concurrently reading from / writing to one of its sysfs attributes. The following lockdep warning hinted towards the issue (CPU0 = dcssblk_remove_store, CPU1 = dcssblk_shared_store): [ 76.496047] Possible unsafe locking scenario: [ 76.496054] CPU0 CPU1 [ 76.496059] ---- ---- [ 76.496087] lock(&dcssblk_devices_sem); [ 76.496090] lock(s_active#175); [ 76.496106] lock(&dcssblk_devices_sem); [ 76.496110] lock(s_active#175); [ 76.496115] *** DEADLOCK *** Fix this by releasing the dcssblk_devices_sem semaphore, which only protects internal DCSS data, before calling device_unregister(). Signed-off-by:
Gerald Schaefer <gerald.schaefer@de.ibm.com> Signed-off-by:
-
- 01 Apr, 2016 9 commits
-
-
Sudip Mukherjee authored
Fixes this build error on linux-next: kernel/seccomp.c: In function '__secure_computing_strict': kernel/seccomp.c:526:3: error: implicit declaration of function 'get_compat_mode1_syscalls' The retrieval of compat syscall numbers were moved into inline function defined in asm-generic header but the asm-generic header is not being used by s390. [heiko.carstens@de.ibm.com]: even though the build error will trigger only in the next merge window it makes sense to include the generic header file already now. Fixes: ("seccomp: Get compat syscalls from asm-generic header") Cc: Matt Redfearn <matt.redfearn@imgtec.com> Signed-off-by:Sudip Mukherjee <sudip.mukherjee@codethink.co.uk> Signed-off-by:
-
Sebastian Ott authored
Newer machines might use a different (larger) format for function measurement blocks. To ensure that we comply with the alignment requirement on these machines and prevent memory corruption (when firmware writes more data than we expect) add 16 padding bytes at the end of the fmb. Cc: stable@vger.kernel.org # v4.1+ Signed-off-by:
-
Sebastian Ott authored
When we refuse a non REQ_TYPE_FS request in the build request function we already hold the queue lock. Thus we must not call blk_end_request_all but __blk_end_request_all. Reported-by:
Peter Oberparleiter <oberpar@linux.vnet.ibm.com> Fixes: de9587a2 ('s390/scm_blk: fix endless loop for requests != REQ_TYPE_FS') Signed-off-by:
-
git://people.freedesktop.org/~airlied/linuxLinus Torvalds authored
Pull drm fixes from Dave Airlie: "Nothing too crazy in here: a bunch of AMD fixes/quirks, two msm fixes, some rockchip fixes, and a udl warning fix, along with one locking fix for displayport that seems to fix some dodgy monitors" * 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: drm/udl: Use unlocked gem unreferencing drm/dp: move hw_mutex up the call stack drm/amdgpu: Don't move pinned BOs drm/radeon: Don't move pinned BOs drm/radeon: add a dpm quirk for all R7 370 parts drm/radeon: add another R7 370 quirk drm/rockchip: dw_hdmi: Don't call platform_set_drvdata() drm/rockchip: vop: Fix vop crtc cleanup drm/rockchip: dw_hdmi: Call drm_encoder_cleanup() in error path drm/rockchip: vop: Disable planes when disabling CRTC drm/rockchip: vop: Don't reject empty modesets drm/rockchip: cancel pending vblanks on close drm/rockchip: vop: fix crtc size in plane check drm/radeon: add a dpm quirk for sapphire Dual-X R7 370 2G D5 drm/amd: Beef up ACP Kconfig menu text drm/msm: fix typo in the !COMMON_CLK case drm/msm: fix bug after preclose removal
-
git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linuxLinus Torvalds authored
Pull powerpc fixes from Michael Ellerman: - Fixup preempt underflow with huge pages from Sebastian Siewior - Fix altivec SPR not being saved from Oliver O'Halloran - Correct used_vsr comment from Simon Guo * tag 'powerpc-4.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: powerpc: Correct used_vsr comment powerpc/process: Fix altivec SPR not being saved powerpc/mm: Fixup preempt underflow with huge pages
-
git://git.kernel.org/pub/scm/linux/kernel/git/s390/linuxLinus Torvalds authored
Pull s390 fixes from Martin Schwidefsky: - A proper fix for the locking issue in the dasd driver - Wire up the new preadv2 nad pwritev2 system calls - Add the mark_rodata_ro function and set DEBUG_RODATA=y - A few more bug fixes. * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: s390: wire up preadv2/pwritev2 syscalls s390/pci: PCI function group 0 is valid for clp_query_pci_fn s390/crypto: provide correct file mode at device register. s390/mm: handle PTE-mapped tail pages in fast gup s390: add DEBUG_RODATA support s390: disable postinit-readonly for now s390/dasd: reorder lcu and device lock s390/cpum_sf: Fix cpu hotplug notifier transitions s390/cpum_cf: Fix missing cpu hotplug notifier transition
-
Heiko Carstens authored
Signed-off-by:
-
Pierre Morel authored
The PCI function group 0 is a valid function group, it is wrong to reject it. Let's accept PCI function group 0. Signed-off-by:
Pierre Morel <pmorel@linux.vnet.ibm.com> Acked-by:
-
Daniel Vetter authored
For drm_gem_object_unreference callers are required to hold dev->struct_mutex, which these paths don't. Enforcing this requirement has become a bit more strict with commit ef4c6270 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Thu Oct 15 09:36:25 2015 +0200 drm/gem: Check locking in drm_gem_object_unreference Cc: stable@vger.kernel.org Signed-off-by:
Daniel Vetter <daniel.vetter@intel.com> Signed-off-by:
Dave Airlie <airlied@redhat.com>
-
