- 07 Oct, 2020 8 commits
-
-
Harald Freudenberger authored
Revisit the ap queue error handling: Based on discussions and evaluatios with the firmware folk here is now a rework of the response code handling for all the AP instructions. The idea is to distinguish between failures because of some kind of invalid request where a retry does not make any sense and a failure where another attempt to send the very same request may succeed. The first case is handled by returning EINVAL to the userspace application. The second case results in retries within the zcrypt API controlled by a per message retry counter. Revisit the zcrpyt error handling: Similar here, based on discussions with the firmware people here comes a rework of the handling of all the reply codes. Main point here is that there are only very few cases left, where a zcrypt device queue is switched to offline. It should never be the case that an AP reply message is 'unknown' to the device driver as it indicates a total mismatch between device driver and crypto card firmware. In all other cases, the code distinguishes between failure because of invalid message (see above - EINVAL) or failures of the infrastructure (see above - EAGAIN). Signed-off-by:
Harald Freudenberger <freude@linux.ibm.com> Signed-off-by:
Vasily Gorbik <gor@linux.ibm.com>
-
Harald Freudenberger authored
Support SCLP AP adapter config and deconfig operations: The sysfs deconfig attribute /sys/devices/ap/cardxx/deconfig for each AP card is now read-write. Writing in a '1' triggers a synchronous SCLP request to configure the adapter, writing in a '0' sends a synchronous SCLP deconfigure request. Signed-off-by:
-
Harald Freudenberger authored
Add support for AP bus adapter config and deconfig to the sclp core code. The code is statically build into the kernel when ZCRYPT is configured either as module or with static support. This is the base functionality for having configure/deconfigure support in the AP bus and card code. Another patch will exploit this soon. Signed-off-by:
Pierre Morel <pmorel@linux.ibm.com> Signed-off-by:
-
Harald Freudenberger authored
This patch adds a new config state to the ap card and queue devices. This state reflects the response code 0x03 "AP deconfigured" on TQAP invocation and is tracked with every ap bus scan. Together with this new state now a card/queue device which is 'deconfigured' is not disposed any more. However, for backward compatibility the online state now needs to take this state into account. So a card/queue is offline when the device is not configured. Furthermore a device can't get switched from offline to online state when not configured. The config state is shown in sysfs at /sys/devices/ap/cardxx/config for the card and /sys/devices/ap/cardxx/xx.yyyy/config for each queue within each card. It is a read-only attribute reflecting the negation of the 'AP deconfig' state as it is noted in the AP documents. Signed-off-by:
-
Harald Freudenberger authored
On AP instruction failures the last response code is now kept in the struct ap_queue. There is also a new sysfs attribute showing this field (enabled only on debug kernels). Also slight rework of the AP_DBF macros to get some more content into one debug feature message line. Signed-off-by:
-
Harald Freudenberger authored
The state machine for each ap queue covered a mixture of device states and state machine (firmware queue state) states. This patch splits the device states and the state machine states into two different enums and variables. The major state is the device state with currently these values: AP_DEV_STATE_UNINITIATED - fresh and virgin, not touched AP_DEV_STATE_OPERATING - queue dev is working normal AP_DEV_STATE_SHUTDOWN - remove/unbind/shutdown in progress AP_DEV_STATE_ERROR - device is in error state only when the device state is > UNINITIATED the state machine is run. The state machine represents the states of the firmware queue: AP_SM_STATE_RESET_START - starting point, reset (RAPQ) ap queue AP_SM_STATE_RESET_WAIT - reset triggered, waiting to be finished if irqs enabled, set up irq (AQIC) AP_SM_STATE_SETIRQ_WAIT - enable irq triggered, waiting to be finished, then go to IDLE AP_SM_STATE_IDLE - queue is operational but empty AP_SM_STATE_WORKING - queue is operational, requests are stored and replies may wait for getting fetched AP_SM_STATE_QUEUE_FULL - firmware queue is full, so only replies can get fetched For debugging each ap queue shows a sysfs attribute 'states' which displays the device and state machine state and is only available when the kernel is build with CONFIG_ZCRYPT_DEBUG enabled. Signed-off-by:
-
Harald Freudenberger authored
Introduce a new config switch CONFIG_ZCRYPT_DEBUG which will be used to enable some features for debugging the zcrypt device driver and ap bus system: Another patch will use this for displaying ap card and ap queue state information via sysfs attribute. A furher patch will use this to enable some special treatment for some fields of an crypto request to be able to inject failures and so help debugging with regards to handling of failures. Signed-off-by:
-
Harald Freudenberger authored
Introduce a new internal struct zcrypt_track with an retry counter field and a last return code field. Fill and update these fields at certain points during processing of an request/reply. This tracking info is then used to - avoid trying to resend the message forever. Now each message is tried to be send TRACK_AGAIN_MAX (currently 10) times and then the ioctl returns to userspace with errno EAGAIN. - avoid trying to resend the message on the very same card/domain. If possible (more than one APQN with same quality) don't use the very same qid as the previous attempt when again scheduling the request. This is done by adding penalty weight values when the dispatching takes place. There is a penalty TRACK_AGAIN_CARD_WEIGHT_PENALTY for using the same card as previously and another penalty define TRACK_AGAIN_QUEUE_WEIGHT_PENALTY to be considered when the same qid as the previous sent attempt is calculated. Both values make it harder to choose the very same card/domain but not impossible. For example when only one APQN is available a resend can only address the very same APQN. There are some more ideas for the future to extend the use of this tracking information. For example the last response code at NQAP and DQAP could be stored there, giving the possibility to extended tracing and debugging about requests failing to get processed properly. Signed-off-by:
-
- 02 Oct, 2020 7 commits
-
-
Vasily Gorbik authored
Early sclp console messages are printed in line mode on z/VM and LPAR, but under kvm newlines matter. Add a missing newline between "kernel version" and "Kernel fault". Signed-off-by: -
Vasily Gorbik authored
arch/s390/kernel/entry.h: suspend_zero_pages - only declaration left after commit 39421627 ("s390: remove broken hibernate / power management support") arch/s390/include/asm/setup.h: vmhalt_cmd - only declaration left after commit 99ca4e58 ("[S390] kernel: Shutdown Actions Interface") arch/s390/include/asm/setup.h: vmpoff_cmd - only declaration left after commit 99ca4e58 ("[S390] kernel: Shutdown Actions Interface") Signed-off-by:
-
Vasily Gorbik authored
Since commit 998f5bbe ("s390/kasan: fix early pgm check handler execution") early pgm check handler is executed with DAT on if Kasan is enabled. Still there is a window between setup_lowcore_dat_off() and setup_lowcore_dat_on() when int handlers could be executed with DAT off under Kasan. If this happens the kernel ends up in pgm check loop due to Kasan shadow memory access attempts. With Kasan enabled paging is initialized much earlier and DAT flag has to be on at all times instrumented code is executed. Make sure int handlers are set up to be called with DAT on right away in this case. Signed-off-by:
-
Gerald Schaefer authored
Re-IPL for nvme is currently done by using diag 308 with the "Load Clear" subcode, which means that all memory will be cleared. This can increase re-IPL duration considerably on very large machines. For list-directed IPL like nvme or fcp IPL, a "Load Normal" subcode was introduced with z14. The "Load Normal" diag 308 subcode allows to re-IPL without clearing memory. This patch adds a new "clear" sysfs attribute to /sys/firmware/reipl/nvme, which can be set to either "0" or "1" to disable or enable re-IPL with memory clearing. The default value is "0", which disables memory clearing. Signed-off-by:
Gerald Schaefer <gerald.schaefer@de.ibm.com> Reviewed-by:
Alexander Egorenkov <egorenar@linux.ibm.com> Signed-off-by:
-
Alexander Egorenkov authored
From the kernel perspective NVMe dump works exactly like zFCP dump. Therefore, adapt all places where code explicitly tests only for IPL of type FCP DUMP. And also set the memory end correctly in this case. Signed-off-by:
Philipp Rudo <prudo@linux.ibm.com> Signed-off-by:
-
Alexander Egorenkov authored
Enable extracting of extra kernel command-line parameters from the NVMe IPL block passed by the firmware to the kernel at boot. Signed-off-by:
-
Jason J. Herne authored
Add the nvme dump ipl type, associated data, and sysfs entries. This allows booting into a stand alone dump environment that resides on an nvme device. Signed-off-by:
Jason J. Herne <jjherne@linux.ibm.com> Signed-off-by:
-
- 30 Sep, 2020 8 commits
-
-
Vasily Gorbik authored
arch/s390/pci/pci_bus.h: zpci_bus_init - only declaration left after commit 05bc1be6 ("s390/pci: create zPCI bus") arch/s390/include/asm/gmap.h: gmap_pte_notify - only declaration left after commit 4be130a0 ("s390/mm: add shadow gmap support") arch/s390/include/asm/pgalloc.h: rcu_table_freelist_finish - only declaration left after commit 36409f63 ("[S390] use generic RCU page-table freeing code") arch/s390/include/asm/tlbflush.h: smp_ptlb_all - only declaration left after commit 5a79859a ("s390: remove 31 bit support") arch/s390/include/asm/vtimer.h: init_cpu_vtimer - only declaration left after commit b5f87f15 ("s390/idle: consolidate idle functions and definitions") arch/s390/include/asm/pci.h: zpci_debug_info - only declaration left after commit 386aa051 ("s390/pci: remove per device debug attribute") arch/s390/include/asm/vdso.h: vdso_alloc_boot_cpu - only declaration left after commit 4bff8cb5 ("s390: convert to GENERIC_VDSO") arch/s390/include/asm/smp.h: smp_vcpu_scheduled - only declaration left after commit 67626fad ("s390: enforce CONFIG_SMP") arch/s390/kernel/entry.h: restart_call_handler - only declaration left after commit 8b646bd7 ("[S390] rework smp code") arch/s390/kernel/entry.h: startup_init_nobss - only declaration left after commit 2e83e0eb ("s390: clean .bss before running uncompressed kernel") arch/s390/kernel/entry.h: s390_early_resume - only declaration left after commit 39421627 ("s390: remove broken hibernate / power management support") drivers/s390/char/raw3270.h: raw3270_request_alloc_bootmem - only declaration left after commit 33403dcf ("[S390] 3270 console: convert from bootmem to slab") drivers/s390/cio/device.h: ccw_device_schedule_sch_unregister - only declaration left after commit 37de53bb ("[S390] cio: introduce ccw device todos") drivers/s390/char/tape.h: tape_hotplug_event - has only declaration since recorded git history. drivers/s390/char/tape.h: tape_oper_handler - has only declaration since recorded git history. drivers/s390/char/tape.h: tape_noper_handler - has only declaration since recorded git history. drivers/s390/char/tape_std.h: tape_std_check_locate - only declaration left after commit 161beff8 ("s390/tape: remove tape block leftovers") drivers/s390/char/tape_std.h: tape_std_default_handler - has only declaration since recorded git history. drivers/s390/char/tape_std.h: tape_std_unexpect_uchk_handler - has only declaration since recorded git history. drivers/s390/char/tape_std.h: tape_std_irq - has only declaration since recorded git history. drivers/s390/char/tape_std.h: tape_std_error_recovery - has only declaration since recorded git history. drivers/s390/char/tape_std.h: tape_std_error_recovery_has_failed - has only declaration since recorded git history. drivers/s390/char/tape_std.h: tape_std_error_recovery_succeded - has only declaration since recorded git history. drivers/s390/char/tape_std.h: tape_std_error_recovery_do_retry - has only declaration since recorded git history. drivers/s390/char/tape_std.h: tape_std_error_recovery_read_opposite - has only declaration since recorded git history. drivers/s390/char/tape_std.h: tape_std_error_recovery_HWBUG - has only declaration since recorded git history. Reviewed-by:
Sven Schnelle <svens@linux.ibm.com> Signed-off-by:
-
Vasily Gorbik authored
startup pgm check handler is active since the very beginning of kernel code execution until uncompressed kernel sets up s390_base_pgm_handler. It is useful not just for the decompressor debugging itself, but also for early code of uncompressed kernel, in particular Kasan initialization. But since there is no stack trace or symbolic representation of failing psw address it is impossible to figure out faulty code location without knowing Kaslr kernel base. So, let's add it to the startup pgm check info printed as well. Reviewed-by:
-
Vasily Gorbik authored
sclp_set_columns and sclp_set_htab are leftovers since commit 095761d2 ("[S390] sclp_tty: remove ioctl interface."), remove them as a dead code. Reviewed-by:
-
Vasily Gorbik authored
sclp_sdias cannot be built as a module, CRASH_DUMP option is a bool not a tristate. zcore_exit() has already been removed with commit cbe62fac ("s390: char: make zcore explicitly non-modular"). Remove orphaned sclp_sdias_exit for consistency as well. Reviewed-by:
-
Vasily Gorbik authored
Currently if just "dfltcc" is passed as a kernel command line option "val" going to be NULL, this leads to reading at address 0 in strcmp(val, "off") Fix that by making sure "val" is not NULL. This does not affect option handling logic. Reviewed-by:
-
Vasily Gorbik authored
Remove couple of declarations which are unused since commit 4bff8cb5 ("s390: convert to GENERIC_VDSO"). Acked-by:
-
Vasily Gorbik authored
Added with commit 77e844b9 ("s390/hibernate: add early resume function") unused since commit 39421627 ("s390: remove broken hibernate / power management support"). Reviewed-by:
Vineeth Vijayan <vneethv@linux.ibm.com> Signed-off-by:
-
Sven Schnelle authored
remove the cad command line option as the instruction was never published and never used by userspace. Signed-off-by:
Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by:
-
- 29 Sep, 2020 5 commits
-
-
Vasily Gorbik authored
Currently we overflow save_area_sync and write over save_area_async. Although this is not a real problem make startup_pgm_check_handler consistent with late pgm check handler and store [%r0,%r7] directly into gpregs_save_area. Reviewed-by:
-
Vasily Gorbik authored
Since commit 39421627 ("s390: remove broken hibernate / power management support") _swsusp_reset_dma is unused and could be safely removed. Reviewed-by:
-
Vasily Gorbik authored
Currently there are several minor problems with randomization base generation code: 1. It might misbehave in low memory conditions. In particular there might be enough space for the kernel on [0, block_sum] but after if (base < safe_addr) base = safe_addr; it might not be enough anymore. 2. It does not correctly handle minimal address constraint. In condition if (base < safe_addr) base = safe_addr; a synthetic value is compared with an address. If we have a memory setup with memory holes due to offline memory regions, and safe_addr is close to the end of the first online memory block - we might position the kernel in invalid memory. 3. block_sum calculation logic contains off-by-one error. Let's say we have a memory block in which the kernel fits perfectly (end - start == kernel_size). In this case: if (end - start < kernel_size) continue; block_sum += end - start - kernel_size; block_sum is not increased, while it is a valid kernel position. So, address problems listed and explain algorithm used. Besides that restructuring the code makes it possible to extend kernel positioning algorithm further. Currently we pick position in between single [min, max] range (min = safe_addr, max = memory_limit). In future we can do that for multiple ranges as well (by calling count_valid_kernel_positions for each range). Reviewed-by:
-
Vasily Gorbik authored
0 is a valid random value. To avoid mixing it with error code 0 as an return code make get_random() take extra argument to output random value and return an error code. Reviewed-by:
-
Qinglang Miao authored
Simplify the return expression. Link: https://lkml.kernel.org/r/20200921131101.93037-1-miaoqinglang@huawei.comSigned-off-by:
Qinglang Miao <miaoqinglang@huawei.com> Signed-off-by:
-
- 26 Sep, 2020 6 commits
-
-
Sven Schnelle authored
No need to have two mutexes, and while at it rename it to stp_mutex. Signed-off-by:
-
Sven Schnelle authored
This patch introduces /sys/devices/system/stp/scheduled_leap_seconds, which will contain either 0,0 if no leap second is scheduled, or the UTC timestamp + leap second offset. Signed-off-by:
-
Sven Schnelle authored
In the current implementation, leap seconds are only synchronized during the bootup process when the STP clock is synced. If the Leap second offset (LSO) changes the machine must be rebooted, which is not desired. This patch adds the required code to handle Leap second changes during runtime. If the Leap second changes, a Configuration change machine check is triggered. The STP code than schedules a Leap second insertion/deletion with do_adjtimex(). Signed-off-by:
-
Sven Schnelle authored
In hardware-dependent headers using u32 is easier to read and less error-prone. Signed-off-by:
-
Sven Schnelle authored
Use __packed instead of __attribute__((packed)) Signed-off-by:
-
Sven Schnelle authored
The sysfs function might race with stp_work_fn. To prevent that, add the required locking. Another issue is that the sysfs functions are checking the stp_online flag, but this flag just holds the user setting whether STP is enabled. Add a flag to clock_sync_flag whether stp_info holds valid data and use that instead. Cc: stable@vger.kernel.org Signed-off-by:
-
- 24 Sep, 2020 2 commits
-
-
Harald Freudenberger authored
This patch extends the pkey kernel module to support CCA and EP11 secure ECC (private) keys as source for deriving ECC protected (private) keys. There is yet another new ioctl to support this: PKEY_KBLOB2PROTK3 can handle all the old keys plus CCA and EP11 secure ECC keys. For details see ioctl description in pkey.h. The CPACF unit currently only supports a subset of 5 different ECC curves (P-256, P-384, P-521, ED25519, ED448) and so only keys of this curve type can be transformed into protected keys. However, the pkey and the cca/ep11 low level functions do not check this but simple pass-through the key blob to the firmware onto the crypto cards. So most likely the failure will be a response carrying an error code resulting in user space errno value EIO instead of EINVAL. Deriving a protected key from an EP11 ECC secure key requires a CEX7 in EP11 mode. Deriving a protected key from an CCA ECC secure key requires a CEX7 in CCA mode. Together with this new ioctl the ioctls for querying lists of apqns (PKEY_APQNS4K and PKEY_APQNS4KT) have been extended to support EP11 and CCA ECC secure key type and key blobs. Together with this ioctl there comes a new struct ep11kblob_header which is to be prepended onto the EP11 key blob. See details in pkey.h for the fields in there. The older EP11 AES key blob with some info stored in the (unused) session field is also supported with this new ioctl. Signed-off-by:
Ingo Franzki <ifranzki@linux.ibm.com> Signed-off-by:
-
Harald Freudenberger authored
Support for CCA APKA (used for CCA ECC keys) master keys. The existing mkvps sysfs attribute for each queue for cards in CCA mode is extended to show the APKA master key register states and verification pattern: Improve the mkvps sysfs attribute to display the APKA master key verification patterns for old, current and new master key registers. The APKA master key is used to encrypt CCA ECC secure keys. The syntax is analog to the existing AES mk verification patterns: APKA NEW: <new_apka_mk_state> <new_apka_mk_mkvp> APKA CUR: <cur_apka_mk_state> <cur_apka_mk_mkvp> APKA OLD: <old_apka_mk_state> <old_apka_mk_mkvp> with <new_apka_mk_state>: 'empty' or 'partial' or 'full' <cur_apka_mk_state>: 'valid' or 'invalid' <old_apka_mk_state>: 'valid' or 'invalid' <new_apka_mk_mkvp>, <cur_apka_mk_mkvp>, <old_apka_mk_mkvp> 8 byte hex string with leading 0x MKVP means Master Key Verification Pattern and is a folded hash over the key value. Only the states 'full' and 'valid' result in displaying a useful mkvp, otherwise a mkvp of all bytes zero is shown. If for any reason the FQ fails and the (cached) information is not available, the state '-' will be shown with the mkvp value also '-'. The values shown here are the very same as the cca panel tools displays. The internal function cca_findcard2() also supports to match against the APKA master key verification patterns and the pkey kernel module which uses this function needed compatible rewrite of these invocations. Signed-off-by:
-
- 21 Sep, 2020 3 commits
-
-
Vasily Gorbik authored
This reverts commit 55a5542a ("s390/hibernate: fix error handling when suspend cpu != resume cpu"). It added sclp_early_printk_force() which is no longer used since commit 39421627 ("s390: remove broken hibernate / power management support"). No hibernate - no problem. Signed-off-by:
-
Vasily Gorbik authored
Since commit 980d5f9a ("s390/boot: enable .bss section for compressed kernel") .bss section usage is no longer restricted. .bss section is a part of the decompressor's image and is zeroed by the linker. For that reason clean up now unneeded .data section usage. Reviewed-by:
-
Vasily Gorbik authored
.bss section is a part of the decompressor's image now, linker fills it with zeros already. No need do it with memset additionally. Reviewed-by:
-
- 17 Sep, 2020 1 commit
-
-
Qinglang Miao authored
The spinlock ap_poll_timer_lock is initialized statically. It is unnecessary to initialize by spin_lock_init(). Signed-off-by:
Heiko Carstens <hca@linux.ibm.com> Signed-off-by:
-
