- 20 Feb, 2017 3 commits
-
-
Martin Schwidefsky authored
Fix PER tracing of system calls after git commit 34525e1f "s390: store breaking event address only for program checks" broke it. Signed-off-by:
Martin Schwidefsky <schwidefsky@de.ibm.com>
-
Paul Gortmaker authored
The Makefile in drivers/s390 has: obj-y += cio/ block/ char/ crypto/ net/ scsi/ virtio/ and the Makefile in crypto/ has: ap-objs := ap_bus.o ap_card.o ap_queue.o meaning that it currently is not being built as a module by anyone. Lets remove the modular code that is essentially orphaned, so that when reading the driver there is no doubt it is builtin-only. Since module_init translates to device_initcall in the non-modular case, the init ordering remains unchanged with this commit. Also note that MODULE_ALIAS is a no-op for non-module builds. We also delete the MODULE_LICENSE tag etc. since all that information is already contained at the top of the file in the comments. We replace module.h with moduleparam.h since the file does declare some module parameters even though it is not modular itself. Signed-off-by:
Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by:
Harald Freudenberger <freude@linux.vnet.ibm.com> Signed-off-by:
-
Harald Freudenberger authored
The ap bus code and the zcrypt api had invocations to the debug feature debugfs_create_dir() call but never populated these directories in any way. Removed this unneeded code. Signed-off-by:
-
- 17 Feb, 2017 9 commits
-
-
Heiko Carstens authored
Signed-off-by:
Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by:
-
Heiko Carstens authored
Walking kernel page tables within the kernel page table dumper may potentially take a lot of time. This may lead to soft lockup warning messages. To avoid this add a cond_resched call for each pgd_level iteration. This is the same as "x86/mm/ptdump: Fix soft lockup in page table walker" for x86. Signed-off-by:
-
Heiko Carstens authored
Both MACHINE_HAS_PFMF and MACHINE_HAS_HPAGE are just an alias for MACHINE_HAS_EDAT1. So simply use MACHINE_HAS_EDAT1 instead. Signed-off-by:
-
Heiko Carstens authored
Fix this compile error for !MEMORY_HOTPLUG && NUMA: arch/s390/built-in.o: In function `emu_setup_size_adjust': arch/s390/numa/mode_emu.c:477: undefined reference to `memory_block_size_bytes' Signed-off-by:
-
Christian Borntraeger authored
Remove the last places of ACCESS_ONCE in s390 code. Signed-off-by:
Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by:
-
Paul Gortmaker authored
Historically a lot of these existed because we did not have a distinction between what was modular code and what was providing support to modules via EXPORT_SYMBOL and friends. That changed when we forked out support for the latter into the export.h file. This means we should be able to reduce the usage of module.h in code that is obj-y Makefile or bool Kconfig. The advantage in doing so is that module.h itself sources about 15 other headers; adding significantly to what we feed cpp, and it can obscure what headers we are effectively using. Since module.h was the source for init.h (for __init) and for export.h (for EXPORT_SYMBOL) we consider each change instance for the presence of either and replace as needed. An instance where module_param was used without moduleparam.h was also fixed, as well as implicit use of ptrace.h and string.h headers. Signed-off-by:
-
Paul Gortmaker authored
Historically a lot of these existed because we did not have a distinction between what was modular code and what was providing support to modules via EXPORT_SYMBOL and friends. That changed when we forked out support for the latter into the export.h file. This means we should be able to reduce the usage of module.h in code that is obj-y Makefile or bool Kconfig. The advantage in doing so is that module.h itself sources about 15 other headers; adding significantly to what we feed cpp, and it can obscure what headers we are effectively using. Since module.h was the source for init.h (for __init) and for export.h (for EXPORT_SYMBOL) we consider each change instance for the presence of either and replace as needed. An instance where module_param was used without moduleparam.h was also fixed, as well as an implict use of asm/elf.h header. Signed-off-by:
-
Paul Gortmaker authored
Historically a lot of these existed because we did not have a distinction between what was modular code and what was providing support to modules via EXPORT_SYMBOL and friends. That changed when we forked out support for the latter into the export.h file. This means we should be able to reduce the usage of module.h in code that is obj-y Makefile or bool Kconfig. The advantage in doing so is that module.h itself sources about 15 other headers; adding significantly to what we feed cpp, and it can obscure what headers we are effectively using. Since module.h was the source for init.h (for __init) and for export.h (for EXPORT_SYMBOL) we consider each change instance for the presence of either and replace as needed. Build testing revealed some implicit header usage that was fixed up accordingly. Signed-off-by:
-
Michael Holzheu authored
In binutils/libbfd (bfd/elf.c) it is enforced that all s390 specific ELF notes like e.g. NT_S390_PREFIX or NT_S390_CTRS have "LINUX" specified as note name. Otherwise the notes are ignored. For /proc/vmcore we currently use "CORE" for these notes. Up to now this has not been a real problem because the dump analysis tool "crash" does not check the note name. But it will break all programs that use libbfd for processing ELF notes. So fix this and use "LINUX" for all s390 specific notes to comply with libbfd. Cc: stable@vger.kernel.org # v4.4+ Reported-by:
Philipp Rudo <prudo@linux.vnet.ibm.com> Reviewed-by:
Michael Holzheu <holzheu@linux.vnet.ibm.com> Signed-off-by:
-
- 08 Feb, 2017 8 commits
-
-
Martin Schwidefsky authored
Bit 0x100 of a page table, segment table of region table entry can be used to disallow code execution for the virtual addresses associated with the entry. There is one tricky bit, the system call to return from a signal is part of the signal frame written to the user stack. With a non-executable stack this would stop working. To avoid breaking things the protection fault handler checks the opcode that caused the fault for 0x0a77 (sys_sigreturn) and 0x0aad (sys_rt_sigreturn) and injects a system call. This is preferable to the alternative solution with a stub function in the vdso because it works for vdso=off and statically linked binaries as well. Signed-off-by: -
Martin Schwidefsky authored
Add hardware capability bits and feature tags to /proc/cpuinfo for the "Vector Packed Decimal Facility" (tag "vxd" / hwcap bit 2^12) and the "Vector Enhancements Facility 1" (tag "vxe" / hwcap bit 2^13). Signed-off-by: -
Heiko Carstens authored
The current implementation of setup_randomness uses the stack address and therefore the pointer to the SYSIB 3.2.2 block as input data address. Furthermore the length of the input data is the number of virtual-machine description blocks which is typically one. This means that typically a single zero byte is fed to add_device_randomness. Fix both of these and use the address of the first virtual machine description block as input data address and also use the correct length. Fixes: bcfcbb6b ("s390: add system information as device randomness") Signed-off-by:
-
Heiko Carstens authored
Get rid of common response code handling. Each command requires its own response code handling anyway. Also the retry in case of -EBUSY does not work and can be simply removed. Reviewed-by:
Peter Oberparleiter <oberpar@linux.vnet.ibm.com> Signed-off-by:
-
Heiko Carstens authored
The early vt220 sclp printk code added an extra new line to each printed multi-line text. If used for the early sclp console this will lead to numerous extra new lines. Therefore get rid of this semantic and require that each to be printed string contains a line feed character if a new line is wanted. Reviewed-by:
-
Heiko Carstens authored
This patch - unifies the old sclp early code and the sclp early printk code, so they can use common functions - makes sure all sclp early functions and variables have the same "sclp_early" prefix - converts the sclp early printk code into readable code by using existing data structures instead of hard coded magic arrays - splits the early sclp code into two files: sclp_early.c and sclp_early_core.c. The core file contains everything that is required by the kernel decompressor and may not call functions not contained within the core file. Otherwise the result would be a link error. - changes interrupt handling to be completely synchronous. The old early sclp code had a small window which allowed to receive several interrupts instead of exactly the single expected interrupt. This did hide a subtle potential bug, which is fixed with this large rework. - contains a couple of small cleanups. Reviewed-by:
-
Heiko Carstens authored
Make sure the early sclp code does not generate any sclp requests anymore as soon as the base sclp driver is active. Otherwise both drivers may see unexpected requests or may miss expected interrupts. Reviewed-by:
-
Heiko Carstens authored
Move the early sclp printk code to the drivers folder where also the rest of the sclp code can be found. This way it is possible to use the sclp private header files for further cleanups. Reviewed-by:
-
- 07 Feb, 2017 3 commits
-
-
Martin Schwidefsky authored
The debug features currently uses absolute TOD time stamps for the debug events. Given that the TOD clock can jump forward and backward due to STP sync checks the order of debug events can get obfuscated. Replace the absolute TOD time stamps with a delta to the IPL time stamp. On a STP sync check the TOD clock correction is added to the IPL time stamp as well to make the deltas unaffected by STP sync check. The readout of the debug feature entries will convert the deltas back to absolute time stamps based on the Unix epoch. Signed-off-by: -
Martin Schwidefsky authored
The data stored by the STSI instruction can be up to a page in size but the memblock_virt_alloc allocation for tl_info only specifies 16 bytes. The memory after the short allocation is overwritten every time arch_update_cpu_topology is called. Fixes: 8c910580 "s390/numa: establish cpu to node mapping early" Signed-off-by:
-
Heiko Carstens authored
Commit bcfcbb6b ("s390: add system information as device randomness") intended to add some virtual machine specific information to the randomness pool. Unfortunately it uses the page allocator before it is ready to use. In result the page allocator always returns NULL and the setup_randomness function never adds anything to the randomness pool. To fix this use memblock_alloc and memblock_free instead. Fixes: bcfcbb6b ("s390: add system information as device randomness") Signed-off-by:
-
- 03 Feb, 2017 5 commits
-
-
Julian Wiedmann authored
With multiple input queues, these DBFs turned out to be not very helpful... Signed-off-by:
Julian Wiedmann <jwi@linux.vnet.ibm.com> Signed-off-by:
-
Julian Wiedmann authored
Missed in commit f4eae94f ("s390/airq: simplify adapter interrupt code") Signed-off-by:
Steffen Maier <maier@linux.vnet.ibm.com> Signed-off-by:
-
Julian Wiedmann authored
In tiqdio_call_inq_handlers(), we're looping over all input queues on the *same* irq. So instead of using the queues' back pointer, we can just access the irq directly. No functional change. Signed-off-by:
-
Julian Wiedmann authored
For devices with multiple input queues, tiqdio_call_inq_handlers() iterates over all input queues and clears the device's DSCI during each iteration. If the DSCI is re-armed during one of the later iterations, we therefore do not scan the previous queues again. The re-arming also raises a new adapter interrupt. But its handler does not trigger a rescan for the device, as the DSCI has already been erroneously cleared. This can result in queue stalls on devices with multiple input queues. Fix it by clearing the DSCI just once, prior to scanning the queues. As the code is moved in front of the loop, we also need to access the DSCI directly (ie irq->dsci) instead of going via each queue's parent pointer to the same irq. This is not a functional change, and a follow-up patch will clean up the other users. In practice, this bug only affects CQ-enabled HiperSockets devices, ie. devices with sysfs-attribute "hsuid" set. Setting a hsuid is needed for AF_IUCV socket applications that use HiperSockets communication. Fixes: 104ea556 ("qdio: support asynchronous delivery of storage blocks") Cc: <stable@vger.kernel.org> # v3.2+ Reviewed-by:
Ursula Braun <ubraun@linux.vnet.ibm.com> Signed-off-by:
-
Gerald Schaefer authored
Since commit dd22f551 "block: Change direct_access calling convention", the device size calculation in dcssblk_direct_access() is off-by-one. This results in bdev_direct_access() always returning -ENXIO because the returned value is not page aligned. Fix this by adding 1 to the dev_sz calculation. Fixes: dd22f551 ("block: Change direct_access calling convention") Cc: <stable@vger.kernel.org> # 4.0+ Signed-off-by:
Gerald Schaefer <gerald.schaefer@de.ibm.com> Signed-off-by:
-
- 31 Jan, 2017 11 commits
-
-
Stefan Haberland authored
Signed-off-by:
Stefan Haberland <sth@linux.vnet.ibm.com> Signed-off-by:
-
Stefan Haberland authored
If safe offline is called for a DASD alias device a null pointer is passed to fsync_bdev. So check for existence of the blockdevice before calling fsync_bdev. Should not be a real world problem since safe offline for an alias device does not make sense and fsync_bdev can deal with a NULL pointer which it gets after successful NULL pointer dereferencing on s390. Signed-off-by:
-
Stefan Haberland authored
Check if the device pointer is valid. Just a sanity check since we already are in the int handler of the device. Signed-off-by:
-
Stefan Haberland authored
Allow 0 as valid input for the path_threshold attribute to deactivate the IFCC/CCC error handling. Signed-off-by:
-
Martin Schwidefsky authored
The principles of operations specifies that the breaking event address is stored to the address 0x110 in the prefix page only for program checks. The last branch in user space is lost as soon as a branch in kernel space is executed after e.g. an svc. This makes it impossible to accurately maintain the breaking event address for a user space process. Simplify the code, just copy the current breaking event address from 0x110 to the task structure for program checks from user space. Signed-off-by: -
Jan Höppner authored
The function dasd_busid() still uses simple_strtoul() to convert a string to an integer value. This function is obsolete for quite some time already and should be replaced. The whole parameter parsing semantic still relies somewhat on the fact, that simple_strtoul() parses a string containing literals without complains and just returns the parsed integer value plus the residual string. kstrtoint(), however, would return -EINVAL in such a case. Since we want to get rid of simple_strtoul() and now have a nice dasd[] containing only single elements, we can clean up and simplify a few things. Replace simple_strtoul() with kstrtouint(), improve and simplify the overall parameter parsing by the following: - instead of residual strings return proper error codes - remove dasd_parse_next_element() and decide directly what sort of element is being parsed - if we parse a device or a range of devices, split that element into separate bits with a new function - remove warning about invalid ending as it doesn't apply anymore - annotate all parsing functions and data that can be freed after initialisation with __init and __initdata respectively - clean up bits and pieces while at it Reviewed-by:
Sebastian Ott <sebott@linux.vnet.ibm.com> Signed-off-by:
Jan Höppner <hoeppner@linux.vnet.ibm.com> Signed-off-by:
-
Jan Höppner authored
When the DASD driver is built into the kernel, the entire comma separated parameter list is stored as one single element in the dasd[] array, opposed to the module build where each element is stored separately in dasd[]. There is no point in doing so. Therefore, store each part of the list as single elements in dasd[] as well when built into the kernel. Also, create a define for the maximum of 256 parameters. Reviewed-by:
-
Harald Freudenberger authored
During tests the Kernel complained about inconsistend lock state: inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage. Now all the queue locks use spin_lock_bh/spin_unlock_bh. Signed-off-by: -
Harald Freudenberger authored
The generate_entropy function used a sha256 for compacting together 256 bits of entropy into 32 bytes hash. However, it is questionable if a sha256 can really be used here, as potential collisions may reduce the max entropy fitting into a 32 byte hash value. So this batch introduces the use of sha512 instead and the required buffer adjustments for the calling functions. Further more the working buffer for the generate_entropy function has been widened from one page to two pages. So now 1024 stckf invocations are used to gather 256 bits of entropy. This has been done to be on the save side if the jitters of stckf values isn't as good as supposed. Signed-off-by:
-
Harald Freudenberger authored
In fips mode only xts keys with 128 bit or 125 bit are allowed. This fix extends the xts_aes_set_key function to check for these valid key lengths in fips mode. Signed-off-by:
-
Matthew Rosato authored
Triple-DES implementations will soon be required to check for uniqueness of keys with fips mode enabled. Add checks to ensure none of the 3 keys match. Signed-off-by:
Matthew Rosato <mjrosato@linux.vnet.ibm.com> Signed-off-by:
-
- 16 Jan, 2017 1 commit
-
-
Daniel Borkmann authored
After we already allocated the jit.prg_buf image via bpf_jit_binary_alloc() and filled it out with instructions, jit.prg_buf cannot be NULL anymore. Thus, remove the unnecessary check. Tested on s390x with test_bpf module. Signed-off-by:
Daniel Borkmann <daniel@iogearbox.net> Cc: Michael Holzheu <holzheu@linux.vnet.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by:
-
