brintos

brintos / linux-shallow public Read only

0
0
Text · 6.9 KiB · 2eabef3 Raw
178 lines · plain
1Ramoops oops/panic logger2=========================3 4Sergiu Iordache <sergiu@chromium.org>5 6Updated: 10 Feb 20217 8Introduction9------------10 11Ramoops is an oops/panic logger that writes its logs to RAM before the system12crashes. It works by logging oopses and panics in a circular buffer. Ramoops13needs a system with persistent RAM so that the content of that area can14survive after a restart.15 16Ramoops concepts17----------------18 19Ramoops uses a predefined memory area to store the dump. The start and size20and type of the memory area are set using three variables:21 22  * ``mem_address`` for the start23  * ``mem_size`` for the size. The memory size will be rounded down to a24    power of two.25  * ``mem_type`` to specify if the memory type (default is pgprot_writecombine).26  * ``mem_name`` to specify a memory region defined by ``reserve_mem`` command27    line parameter.28 29Typically the default value of ``mem_type=0`` should be used as that sets the pstore30mapping to pgprot_writecombine. Setting ``mem_type=1`` attempts to use31``pgprot_noncached``, which only works on some platforms. This is because pstore32depends on atomic operations. At least on ARM, pgprot_noncached causes the33memory to be mapped strongly ordered, and atomic operations on strongly ordered34memory are implementation defined, and won't work on many ARMs such as omaps.35Setting ``mem_type=2`` attempts to treat the memory region as normal memory,36which enables full cache on it. This can improve the performance.37 38The memory area is divided into ``record_size`` chunks (also rounded down to39power of two) and each kmesg dump writes a ``record_size`` chunk of40information.41 42Limiting which kinds of kmsg dumps are stored can be controlled via43the ``max_reason`` value, as defined in include/linux/kmsg_dump.h's44``enum kmsg_dump_reason``. For example, to store both Oopses and Panics,45``max_reason`` should be set to 2 (KMSG_DUMP_OOPS), to store only Panics46``max_reason`` should be set to 1 (KMSG_DUMP_PANIC). Setting this to 047(KMSG_DUMP_UNDEF), means the reason filtering will be controlled by the48``printk.always_kmsg_dump`` boot param: if unset, it'll be KMSG_DUMP_OOPS,49otherwise KMSG_DUMP_MAX.50 51The module uses a counter to record multiple dumps but the counter gets reset52on restart (i.e. new dumps after the restart will overwrite old ones).53 54Ramoops also supports software ECC protection of persistent memory regions.55This might be useful when a hardware reset was used to bring the machine back56to life (i.e. a watchdog triggered). In such cases, RAM may be somewhat57corrupt, but usually it is restorable.58 59Setting the parameters60----------------------61 62Setting the ramoops parameters can be done in several different manners:63 64 A. Use the module parameters (which have the names of the variables described65 as before). For quick debugging, you can also reserve parts of memory during66 boot and then use the reserved memory for ramoops. For example, assuming a67 machine with > 128 MB of memory, the following kernel command line will tell68 the kernel to use only the first 128 MB of memory, and place ECC-protected69 ramoops region at 128 MB boundary::70 71	mem=128M ramoops.mem_address=0x8000000 ramoops.ecc=172 73 B. Use Device Tree bindings, as described in74 ``Documentation/devicetree/bindings/reserved-memory/ramoops.yaml``.75 For example::76 77	reserved-memory {78		#address-cells = <2>;79		#size-cells = <2>;80		ranges;81 82		ramoops@8f000000 {83			compatible = "ramoops";84			reg = <0 0x8f000000 0 0x100000>;85			record-size = <0x4000>;86			console-size = <0x4000>;87		};88	};89 90 C. Use a platform device and set the platform data. The parameters can then91 be set through that platform data. An example of doing that is:92 93 .. code-block:: c94 95  #include <linux/pstore_ram.h>96  [...]97 98  static struct ramoops_platform_data ramoops_data = {99        .mem_size               = <...>,100        .mem_address            = <...>,101        .mem_type               = <...>,102        .record_size            = <...>,103        .max_reason             = <...>,104        .ecc                    = <...>,105  };106 107  static struct platform_device ramoops_dev = {108        .name = "ramoops",109        .dev = {110                .platform_data = &ramoops_data,111        },112  };113 114  [... inside a function ...]115  int ret;116 117  ret = platform_device_register(&ramoops_dev);118  if (ret) {119	printk(KERN_ERR "unable to register platform device\n");120	return ret;121  }122 123 D. Using a region of memory reserved via ``reserve_mem`` command line124    parameter. The address and size will be defined by the ``reserve_mem``125    parameter. Note, that ``reserve_mem`` may not always allocate memory126    in the same location, and cannot be relied upon. Testing will need127    to be done, and it may not work on every machine, nor every kernel.128    Consider this a "best effort" approach. The ``reserve_mem`` option129    takes a size, alignment and name as arguments. The name is used130    to map the memory to a label that can be retrieved by ramoops.131 132	reserve_mem=2M:4096:oops  ramoops.mem_name=oops133 134You can specify either RAM memory or peripheral devices' memory. However, when135specifying RAM, be sure to reserve the memory by issuing memblock_reserve()136very early in the architecture code, e.g.::137 138	#include <linux/memblock.h>139 140	memblock_reserve(ramoops_data.mem_address, ramoops_data.mem_size);141 142Dump format143-----------144 145The data dump begins with a header, currently defined as ``====`` followed by a146timestamp and a new line. The dump then continues with the actual data.147 148Reading the data149----------------150 151The dump data can be read from the pstore filesystem. The format for these152files is ``dmesg-ramoops-N``, where N is the record number in memory. To delete153a stored record from RAM, simply unlink the respective pstore file.154 155Persistent function tracing156---------------------------157 158Persistent function tracing might be useful for debugging software or hardware159related hangs. The functions call chain log is stored in a ``ftrace-ramoops``160file. Here is an example of usage::161 162 # mount -t debugfs debugfs /sys/kernel/debug/163 # echo 1 > /sys/kernel/debug/pstore/record_ftrace164 # reboot -f165 [...]166 # mount -t pstore pstore /mnt/167 # tail /mnt/ftrace-ramoops168 0 ffffffff8101ea64  ffffffff8101bcda  native_apic_mem_read <- disconnect_bsp_APIC+0x6a/0xc0169 0 ffffffff8101ea44  ffffffff8101bcf6  native_apic_mem_write <- disconnect_bsp_APIC+0x86/0xc0170 0 ffffffff81020084  ffffffff8101a4b5  hpet_disable <- native_machine_shutdown+0x75/0x90171 0 ffffffff81005f94  ffffffff8101a4bb  iommu_shutdown_noop <- native_machine_shutdown+0x7b/0x90172 0 ffffffff8101a6a1  ffffffff8101a437  native_machine_emergency_restart <- native_machine_restart+0x37/0x40173 0 ffffffff811f9876  ffffffff8101a73a  acpi_reboot <- native_machine_emergency_restart+0xaa/0x1e0174 0 ffffffff8101a514  ffffffff8101a772  mach_reboot_fixups <- native_machine_emergency_restart+0xe2/0x1e0175 0 ffffffff811d9c54  ffffffff8101a7a0  __const_udelay <- native_machine_emergency_restart+0x110/0x1e0176 0 ffffffff811d9c34  ffffffff811d9c80  __delay <- __const_udelay+0x30/0x40177 0 ffffffff811d9d14  ffffffff811d9c3f  delay_tsc <- __delay+0xf/0x20178