- 29 Aug, 2024 10 commits
-
-
Harald Freudenberger authored
As a preparation step for introducing a common function API between the pkey API module and the handlers (that is the cca, ep11 and pckmo code) this patch unifies the functions signatures exposed by the handlers and reworks all the invocation code of these functions. Signed-off-by:
Harald Freudenberger <freude@linux.ibm.com> Reviewed-by:
Holger Dengler <dengler@linux.ibm.com> Signed-off-by:
Vasily Gorbik <gor@linux.ibm.com>
-
Harald Freudenberger authored
This is a huge rework of all the pkey kernel module code. The goal is to split the code into individual parts with a dedicated calling interface: - move all the sysfs related code into pkey_sysfs.c - all the CCA related code goes to pkey_cca.c - the EP11 stuff has been moved to pkey_ep11.c - the PCKMO related code is now in pkey_pckmo.c The CCA, EP11 and PCKMO code may be seen as "handlers" with a similar calling interface. The new header file pkey_base.h declares this calling interface. The remaining code in pkey_api.c handles the ioctl, the pkey module things and the "handler" independent code on top of the calling interface invoking the handlers. This regrouping of the code will be the base for a real pkey kernel module split into a pkey base module which acts as a dispatcher and handler modules providing their service. Signed-off-by:
-
Harald Freudenberger authored
Split the very huge ioctl handling function pkey_unlocked_ioctl() into individual functions per each IOCTL command. There is no change in functional code coming with this patch. The work is a simple copy-and-paste with the goal to have the functionality absolutely untouched. Signed-off-by:
-
Holger Dengler authored
Add new shash exploiting the HMAC hardware accelerations for SHA224, SHA256, SHA384 and SHA512 introduced with message-security assist extension 11. Reviewed-by:
-
Holger Dengler authored
Add new cipher exploiting the full AES-XTS hardware acceleration introduced with message-security assist extension 10. The full AES-XTS cipher is registered as preferred cipher in addition to the discrete AES-XTS variant. Reviewed-by:
-
Mete Durlu authored
Remove leftover dentry variable after hypfs refactoring. Before 2fcb3686, hypfs_diag.c and other hypfs files were using debugfs_create_file() explicitly for creating debugfs files and were storing the returned pointer. After the refactor, common debugfs file operations and also the related dentry pointers have been moved into hypfs_dbfs.c and redefined as new common mechanisms. Therefore the dentry variable and the debugfs_remove() function calls in hypfs_diag.c are now redundant. Current code is not effected since the dentry pointer in hypfs_diag is implicitly assigned to NULL and debugfs_remove() returns without an error if the passed pointer is NULL. Acked-by:
Alexander Gordeev <agordeev@linux.ibm.com> Signed-off-by:
Mete Durlu <meted@linux.ibm.com> Signed-off-by:
-
Vasily Gorbik authored
Add more instructions to the kernel disassembler. Reviewed-by:
Jens Remus <jremus@linux.ibm.com> Reviewed-by:
Heiko Carstens <hca@linux.ibm.com> Signed-off-by:
-
Vasily Gorbik authored
Since commit ba05b39d ("s390/expoline: Make modules use kernel expolines"), there is no longer any reason not to use CONFIG_EXPOLINE_EXTERN when supported by the compiler. On the positive side: - there is only a single set of expolines generated and used by both the kernel code and modules, - it eliminates expolines "comdat" sections, which can confuse tools like kpatch. Always enable EXPOLINE_EXTERN if supported by the compiler. Suggested-by:
Sumanth Korikkar <sumanthk@linux.ibm.com> Reviewed-by:
-
Jens Remus authored
Over the course of CPU generations a few instructions got extended, changing their base mnemonic, while keeping the former as an extended mnemonic. Update the instruction mnemonics in the disassembler to their latest base mnemonic as documented in the latest IBM z/Architecture Principles of Operation specification [1]. With the IBM z14 the base mnemonics of the following vector instructions have been changed: - Vector FP Load Lengthened (VFLL) - Vector FP Load Rounded (VFLR) With Message-Security-Assist Extension 5 Perform Pseudorandom Number Operation (PPNO) has been renamed to Perform Random Number Operation (PRNO). With Vector Enhancements Facility 2 the base mnemonics of the following vector instructions have been changed: - Vector FP Convert from Fixed (VCFPS) - Vector FP Convert from Logical (VCFPL) - Vector FP Convert to Fixed (VCSFP) - Vector FP Convert to Logical (VCLFP) [1] IBM z/Architecture Principles of Operation, SA22-7832-13, IBM z16, https://publibfp.dhe.ibm.com/epubs/pdf/a227832d.pdfAcked-by: -
Jens Remus authored
Treat register numbers as unsigned. Treat signed operand values as signed. This resolves multiple instances of the Cppcheck warning: warning: %i in format string (no. 1) requires 'int' but the argument type is 'unsigned int'. [invalidPrintfArgType_sint] Acked-by:
-
- 27 Aug, 2024 4 commits
-
-
Vasily Gorbik authored
ftrace_return_address() is called extremely often from performance-critical code paths when debugging features like CONFIG_TRACE_IRQFLAGS are enabled. For example, with debug_defconfig, ftrace selftests on my LPAR currently execute ftrace_return_address() as follows: ftrace_return_address(0) - 0 times (common code uses __builtin_return_address(0) instead) ftrace_return_address(1) - 2,986,805,401 times (with this patch applied) ftrace_return_address(2) - 140 times ftrace_return_address(>2) - 0 times The use of __builtin_return_address(n) was replaced by return_address() with an unwinder call by commit cae74ba8 ("s390/ftrace: Use unwinder instead of __builtin_return_address()") because __builtin_return_address(n) simply walks the stack backchain and doesn't check for reaching the stack top. For shallow stacks with fewer than "n" frames, this results in reads at low addresses and random memory accesses. While calling the fully functional unwinder "works", it is very slow for this purpose. Moreover, potentially following stack switches and walking past IRQ context is simply wrong thing to do for ftrace_return_address(). Reimplement return_address() to essentially be __builtin_return_address(n) with checks for reaching the stack top. Since the ftrace_return_address(n) argument is always a constant, keep the implementation in the header, allowing both GCC and Clang to unroll the loop and optimize it to the bare minimum. Fixes: cae74ba8 ("s390/ftrace: Use unwinder instead of __builtin_return_address()") Cc: stable@vger.kernel.org Reported-by:
-
Jens Remus authored
Since commit 778666df ("s390: compile relocatable kernel without -fPIE") the kernel vmlinux ELF file is linked with --emit-relocs to preserve all relocations, so that all absolute relocations can be extracted using the 'relocs' tool to adjust them during boot. Port and adapt Petr Pavlu's x86 commit 9d9173e9 ("x86/build: Avoid relocation information in final vmlinux") to s390 to strip all relocations from the final vmlinux ELF file to optimize its size. Following is his original commit message with minor adaptions for s390: The Linux build process on s390 roughly consists of compiling all input files, statically linking them into a vmlinux ELF file, and then taking and turning this file into an actual bzImage bootable file. vmlinux has in this process two main purposes: 1) It is an intermediate build target on the way to produce the final bootable image. 2) It is a file that is expected to be used by debuggers and standard ELF tooling to work with the built kernel. For the second purpose, a vmlinux file is typically collected by various package build recipes, such as distribution spec files, including the kernel's own tar-pkg target. When building the kernel vmlinux contains also relocation information produced by using the --emit-relocs linker option. This is utilized by subsequent build steps to create relocs.S and produce a relocatable image. However, the information is not needed by debuggers and other standard ELF tooling. The issue is then that the collected vmlinux file and hence distribution packages end up unnecessarily large because of this extra data. The following is a size comparison of vmlinux v6.10 with and without the relocation information: | Configuration | With relocs | Stripped relocs | | defconfig | 696 MB | 320 MB | | -CONFIG_DEBUG_INFO | 48 MB | 32 MB | Optimize a resulting vmlinux by adding a postlink step that splits the relocation information into relocs.S and then strips it from the vmlinux binary. Reviewed-by:
-
Vasily Gorbik authored
Now that both the kernel modules area and the kernel image itself are located within 4 GB, there is no longer a need to maintain a separate ftrace_plt trampoline. Use the existing trampoline in the kernel. Reviewed-by:
Ilya Leoshkevich <iii@linux.ibm.com> Reviewed-by:
-
Vasily Gorbik authored
Unused since commit b860b934 ("s390/ftrace: remove dead code"). Reviewed-by:
-
- 22 Aug, 2024 4 commits
-
-
Heiko Carstens authored
If the early program check handler cannot resolve a program check dump register contents and a call trace to the console before loading a disabled wait psw. This makes debugging much easier. Emit an extra message with early_printk() for cases where regular printk() via the early console is not yet working so that at least some information is available. Signed-off-by:
-
Heiko Carstens authored
__do_early_pgm_check() is a function which is only needed during early setup code. Mark it __init in order to save a few bytes. Signed-off-by:
-
Thomas Richter authored
Remove WARN_ON_ONCE statements. These have not triggered in the past. Signed-off-by:
Thomas Richter <tmricht@linux.ibm.com> Acked-by:
-
Thomas Richter authored
Rework debug messages: - Remove most of the debug_sprintf_event() invocations. - Do not split string format statements - Remove colon after function name. Signed-off-by:
-
- 21 Aug, 2024 7 commits
-
-
Thomas Richter authored
qsi() executes the instruction qsi (query sample information) and stores the result of the query in a sample information block pointed to by the function argument. The instruction does not change the condition code register. The return code is always zero. No need to check for errors. Remove now unreferenced macros PMC_FAILURE and RS_INIT_FAILURE_QSI. Signed-off-by:
-
Thomas Richter authored
sf_disable() returns the condition code of instruction lsctl (load sampling controls). However the parameter to lsctl() in sf_disable() is a sample control block containing all zeroes. This invocation of lsctl() does not fail and returns always zero even when there is no authorization for sampling on the machine. In short, sampling can be always turned off. Ignore the return code of sf_disable() and change the function return to void. Signed-off-by:
-
Holger Dengler authored
The dynamic debugging provides function names on request. So remove all explicit function strings. Reviewed-by:
-
Holger Dengler authored
The dynamic debugging provides function names on request. So remove all explicit function strings. Reviewed-by:
-
Holger Dengler authored
The dynamic debugging provides function names on request. So remove all explicit function strings. Reviewed-by:
-
Holger Dengler authored
The dynamic debugging provides function names on request. So remove all explicit function strings. Reviewed-by:
-
Holger Dengler authored
The dynamic debugging provides function names on request. So remove all explicit function strings. Reviewed-by:
-
- 07 Aug, 2024 14 commits
-
-
Heiko Carstens authored
Save some bytes and move early_pgm_check_handler() to init text section. Reviewed-by:
-
Heiko Carstens authored
Add missing warning handling to the early program check handler. This way a warning is printed to the console as soon as the early console is setup, and the kernel continues to boot. Before this change a disabled wait psw was loaded instead and the machine was silently stopped without giving an idea about what happened. Reviewed-by:
Sven Schnelle <svens@linux.ibm.com> Signed-off-by:
-
Heiko Carstens authored
Add the missing pieces so the early program check handler also works with a relocated lowcore. Right now the result of an early program check in case of a relocated lowcore would be a program check loop. Fixes: 8f1e70ad ("s390/boot: Add cmdline option to relocate lowcore") Reviewed-by:
-
Heiko Carstens authored
Have all program check handlers in one file to make future changes easy. Reviewed-by:
-
Heiko Carstens authored
Generate the address marker array dynamically instead of modifying a large static array at kernel startup. Each marker is added twice to the array: with and without a "start" indicator. This way the code and logic stays similar to other architectures. Acked-by:
-
Thomas Richter authored
All functions but setup_pmc_cpu() use a local variable named cpuhw to refer to struct cpu_hw_sf. In setup_pmc_cpu() rename variable cpusf to cpuhw. This makes the naming scheme consistent with all other functions. No functional change. Signed-off-by:
-
Thomas Richter authored
Define and initialize a variable in one place. Remove space between cast and variable. No functional change. Signed-off-by:
-
Thomas Richter authored
In hw_perf_event_update() and cpumsf_pmu_enable() use variable hwc consistently to access event's hardware related data. No functional change. Signed-off-by:
-
Thomas Richter authored
The macros PERF_CPUM_CF_MAX_CTR and PERF_EVENT_CPUM_CF_DIAG are used in only one source file arch/s390/kernel/perf_cpum_cf.c. Move these defines from the header file arch/s390/include/asm/perf_event.h to the only user. No functional change. Signed-off-by:
-
Thomas Richter authored
Some defines in common header file arch/s390/include/asm/perf_event.h are only used in one source file arch/s390/kernel/perf_cpum_sf.c. Move these defines from header to source file. No functional change. Signed-off-by:
-
Thomas Richter authored
Rename macro SAMPLE_FREQ_MODE to SAMPL_FREQ_MODE to make its prefix consistent with all other macro starting with prefix SAMPL_XXX. No functional change. Signed-off-by:
-
Thomas Richter authored
Member hw_perf_event::reg.reg is set but never used, so remove it. Defines REG_NONE and REG_OVERFLOW are not referenced anymore. The initialization to zero takes place in function perf_event_alloc() where ... event = kmem_cache_alloc_node(perf_event_cache, GFP_KERNEL | __GFP_ZERO, node); ... makes sure memory allocated for the event is zero'ed. This is done in the kernel's common code in kernel/events/core.c The struct perf_event contains member hw_perf_event as in struct perf_event { .... struct hw_perf_event hw; .... }; This contained sub-structure is also initialized to zero. No functional change. Signed-off-by: -
Thomas Richter authored
Remove unreferenced marco PERF_CPUM_SF_MODE_MASK. No functional change. Signed-off-by:
-
Thomas Richter authored
Replace atomic_t by refcount_t for reference counting of events. Signed-off-by:
-
- 04 Aug, 2024 1 commit
-
-
Linus Torvalds authored
-
