brintos

brintos / linux-shallow public Read only

0
0
Text · 11.3 KiB · 5c3c0df Raw
419 lines · c
1/* SPDX-License-Identifier: GPL-2.0-only */2/* Copyright(c) 2022 Intel Corporation. */3 4#ifndef _IFS_H_5#define _IFS_H_6 7/**8 * DOC: In-Field Scan9 *10 * =============11 * In-Field Scan12 * =============13 *14 * Introduction15 * ------------16 *17 * In Field Scan (IFS) is a hardware feature to run circuit level tests on18 * a CPU core to detect problems that are not caught by parity or ECC checks.19 * Future CPUs will support more than one type of test which will show up20 * with a new platform-device instance-id.21 *22 *23 * IFS Image24 * ---------25 *26 * Intel provides a firmware file containing the scan tests via27 * github [#f1]_.  Similar to microcode there is a separate file for each28 * family-model-stepping. IFS Images are not applicable for some test types.29 * Wherever applicable the sysfs directory would provide a "current_batch" file30 * (see below) for loading the image.31 *32 *33 * IFS Image Loading34 * -----------------35 *36 * The driver loads the tests into memory reserved BIOS local to each CPU37 * socket in a two step process using writes to MSRs to first load the38 * SHA hashes for the test. Then the tests themselves. Status MSRs provide39 * feedback on the success/failure of these steps.40 *41 * The test files are kept in a fixed location: /lib/firmware/intel/ifs_<n>/42 * For e.g if there are 3 test files, they would be named in the following43 * fashion:44 * ff-mm-ss-01.scan45 * ff-mm-ss-02.scan46 * ff-mm-ss-03.scan47 * (where ff refers to family, mm indicates model and ss indicates stepping)48 *49 * A different test file can be loaded by writing the numerical portion50 * (e.g 1, 2 or 3 in the above scenario) into the curent_batch file.51 * To load ff-mm-ss-02.scan, the following command can be used::52 *53 *   # echo 2 > /sys/devices/virtual/misc/intel_ifs_<n>/current_batch54 *55 * The above file can also be read to know the currently loaded image.56 *57 * Running tests58 * -------------59 *60 * Tests are run by the driver synchronizing execution of all threads on a61 * core and then writing to the ACTIVATE_SCAN MSR on all threads. Instruction62 * execution continues when:63 *64 * 1) All tests have completed.65 * 2) Execution was interrupted.66 * 3) A test detected a problem.67 *68 * Note that ALL THREADS ON THE CORE ARE EFFECTIVELY OFFLINE FOR THE69 * DURATION OF THE TEST. This can be up to 200 milliseconds. If the system70 * is running latency sensitive applications that cannot tolerate an71 * interruption of this magnitude, the system administrator must arrange72 * to migrate those applications to other cores before running a core test.73 * It may also be necessary to redirect interrupts to other CPUs.74 *75 * In all cases reading the corresponding test's STATUS MSR provides details on what76 * happened. The driver makes the value of this MSR visible to applications77 * via the "details" file (see below). Interrupted tests may be restarted.78 *79 * The IFS driver provides sysfs interfaces via /sys/devices/virtual/misc/intel_ifs_<n>/80 * to control execution:81 *82 * Test a specific core::83 *84 *   # echo <cpu#> > /sys/devices/virtual/misc/intel_ifs_<n>/run_test85 *86 * when HT is enabled any of the sibling cpu# can be specified to test87 * its corresponding physical core. Since the tests are per physical core,88 * the result of testing any thread is same. All siblings must be online89 * to run a core test. It is only necessary to test one thread.90 *91 * For e.g. to test core corresponding to cpu592 *93 *   # echo 5 > /sys/devices/virtual/misc/intel_ifs_<n>/run_test94 *95 * Results of the last test is provided in /sys::96 *97 *   $ cat /sys/devices/virtual/misc/intel_ifs_<n>/status98 *   pass99 *100 * Status can be one of pass, fail, untested101 *102 * Additional details of the last test is provided by the details file::103 *104 *   $ cat /sys/devices/virtual/misc/intel_ifs_<n>/details105 *   0x8081106 *107 * The details file reports the hex value of the test specific status MSR.108 * Hardware defined error codes are documented in volume 4 of the Intel109 * Software Developer's Manual but the error_code field may contain one of110 * the following driver defined software codes:111 *112 * +------+--------------------+113 * | 0xFD | Software timeout   |114 * +------+--------------------+115 * | 0xFE | Partial completion |116 * +------+--------------------+117 *118 * Driver design choices119 * ---------------------120 *121 * 1) The ACTIVATE_SCAN MSR allows for running any consecutive subrange of122 * available tests. But the driver always tries to run all tests and only123 * uses the subrange feature to restart an interrupted test.124 *125 * 2) Hardware allows for some number of cores to be tested in parallel.126 * The driver does not make use of this, it only tests one core at a time.127 *128 * .. [#f1] https://github.com/intel/TBD129 *130 *131 * Structural Based Functional Test at Field (SBAF):132 * -------------------------------------------------133 *134 * SBAF is a new type of testing that provides comprehensive core test135 * coverage complementing Scan at Field (SAF) testing. SBAF mimics the136 * manufacturing screening environment and leverages the same test suite.137 * It makes use of Design For Test (DFT) observation sites and features138 * to maximize coverage in minimum time.139 *140 * Similar to the SAF test, SBAF isolates the core under test from the141 * rest of the system during execution. Upon completion, the core142 * seamlessly resets to its pre-test state and resumes normal operation.143 * Any machine checks or hangs encountered during the test are confined to144 * the isolated core, preventing disruption to the overall system.145 *146 * Like the SAF test, the SBAF test is also divided into multiple batches,147 * and each batch test can take hundreds of milliseconds (100-200 ms) to148 * complete. If such a lengthy interruption is undesirable, it is149 * recommended to relocate the time-sensitive applications to other cores.150 */151#include <linux/device.h>152#include <linux/miscdevice.h>153 154#define MSR_ARRAY_BIST				0x00000105155 156#define MSR_COPY_SBAF_HASHES			0x000002b8157#define MSR_SBAF_HASHES_STATUS			0x000002b9158#define MSR_AUTHENTICATE_AND_COPY_SBAF_CHUNK	0x000002ba159#define MSR_SBAF_CHUNKS_AUTHENTICATION_STATUS	0x000002bb160#define MSR_ACTIVATE_SBAF			0x000002bc161#define MSR_SBAF_STATUS				0x000002bd162 163#define MSR_COPY_SCAN_HASHES			0x000002c2164#define MSR_SCAN_HASHES_STATUS			0x000002c3165#define MSR_AUTHENTICATE_AND_COPY_CHUNK		0x000002c4166#define MSR_CHUNKS_AUTHENTICATION_STATUS	0x000002c5167#define MSR_ACTIVATE_SCAN			0x000002c6168#define MSR_SCAN_STATUS				0x000002c7169#define MSR_ARRAY_TRIGGER			0x000002d6170#define MSR_ARRAY_STATUS			0x000002d7171#define MSR_SAF_CTRL				0x000004f0172#define MSR_SBAF_CTRL				0x000004f8173 174#define SCAN_NOT_TESTED				0175#define SCAN_TEST_PASS				1176#define SCAN_TEST_FAIL				2177 178#define IFS_TYPE_SAF			0179#define IFS_TYPE_ARRAY_BIST		1180#define IFS_TYPE_SBAF			2181 182#define ARRAY_GEN0			0183#define ARRAY_GEN1			1184 185/* MSR_SCAN_HASHES_STATUS bit fields */186union ifs_scan_hashes_status {187	u64	data;188	struct {189		u32	chunk_size	:16;190		u32	num_chunks	:8;191		u32	rsvd1		:8;192		u32	error_code	:8;193		u32	rsvd2		:11;194		u32	max_core_limit	:12;195		u32	valid		:1;196	};197};198 199union ifs_scan_hashes_status_gen2 {200	u64	data;201	struct {202		u16	chunk_size;203		u16	num_chunks;204		u32	error_code		:8;205		u32	chunks_in_stride	:9;206		u32	rsvd			:2;207		u32	max_core_limit		:12;208		u32	valid			:1;209	};210};211 212/* MSR_CHUNKS_AUTH_STATUS bit fields */213union ifs_chunks_auth_status {214	u64	data;215	struct {216		u32	valid_chunks	:8;217		u32	total_chunks	:8;218		u32	rsvd1		:16;219		u32	error_code	:8;220		u32	rsvd2		:24;221	};222};223 224union ifs_chunks_auth_status_gen2 {225	u64	data;226	struct {227		u16	valid_chunks;228		u16	total_chunks;229		u32	error_code	:8;230		u32	rsvd2		:8;231		u32	max_bundle	:16;232	};233};234 235/* MSR_ACTIVATE_SCAN bit fields */236union ifs_scan {237	u64	data;238	struct {239		union {240			struct {241				u8	start;242				u8	stop;243				u16	rsvd;244			} gen0;245			struct {246				u16	start;247				u16	stop;248			} gen2;249		};250		u32	delay	:31;251		u32	sigmce	:1;252	};253};254 255/* MSR_SCAN_STATUS bit fields */256union ifs_status {257	u64	data;258	struct {259		union {260			struct {261				u8	chunk_num;262				u8	chunk_stop_index;263				u16	rsvd1;264			} gen0;265			struct {266				u16	chunk_num;267				u16	chunk_stop_index;268			} gen2;269		};270		u32	error_code		:8;271		u32	rsvd2			:22;272		u32	control_error		:1;273		u32	signature_error		:1;274	};275};276 277/* MSR_ARRAY_BIST bit fields */278union ifs_array {279	u64	data;280	struct {281		u32	array_bitmask;282		u16	array_bank;283		u16	rsvd			:15;284		u16	ctrl_result		:1;285	};286};287 288/* MSR_ACTIVATE_SBAF bit fields */289union ifs_sbaf {290	u64	data;291	struct {292		u32	bundle_idx	:9;293		u32	rsvd1		:5;294		u32	pgm_idx		:2;295		u32	rsvd2		:16;296		u32	delay		:31;297		u32	sigmce		:1;298	};299};300 301/* MSR_SBAF_STATUS bit fields */302union ifs_sbaf_status {303	u64	data;304	struct {305		u32	bundle_idx	:9;306		u32	rsvd1		:5;307		u32	pgm_idx		:2;308		u32	rsvd2		:16;309		u32	error_code	:8;310		u32	rsvd3		:21;311		u32	test_fail	:1;312		u32	sbaf_status	:2;313	};314};315 316/*317 * Driver populated error-codes318 * 0xFD: Test timed out before completing all the chunks.319 * 0xFE: not all scan chunks were executed. Maximum forward progress retries exceeded.320 */321#define IFS_SW_TIMEOUT				0xFD322#define IFS_SW_PARTIAL_COMPLETION		0xFE323 324#define IFS_SUFFIX_SZ		5325 326struct ifs_test_caps {327	int	integrity_cap_bit;328	int	test_num;329	char	image_suffix[IFS_SUFFIX_SZ];330};331 332/**333 * struct ifs_test_msrs - MSRs used in IFS tests334 * @copy_hashes: Copy test hash data335 * @copy_hashes_status: Status of copied test hash data336 * @copy_chunks: Copy chunks of the test data337 * @copy_chunks_status: Status of the copied test data chunks338 * @test_ctrl: Control the test attributes339 */340struct ifs_test_msrs {341	u32	copy_hashes;342	u32	copy_hashes_status;343	u32	copy_chunks;344	u32	copy_chunks_status;345	u32	test_ctrl;346};347 348/**349 * struct ifs_data - attributes related to intel IFS driver350 * @loaded_version: stores the currently loaded ifs image version.351 * @loaded: If a valid test binary has been loaded into the memory352 * @loading_error: Error occurred on another CPU while loading image353 * @valid_chunks: number of chunks which could be validated.354 * @status: it holds simple status pass/fail/untested355 * @scan_details: opaque scan status code from h/w356 * @cur_batch: number indicating the currently loaded test file357 * @generation: IFS test generation enumerated by hardware358 * @chunk_size: size of a test chunk359 * @array_gen: test generation of array test360 * @max_bundle: maximum bundle index361 */362struct ifs_data {363	int	loaded_version;364	bool	loaded;365	bool	loading_error;366	int	valid_chunks;367	int	status;368	u64	scan_details;369	u32	cur_batch;370	u32	generation;371	u32	chunk_size;372	u32	array_gen;373	u32	max_bundle;374};375 376struct ifs_work {377	struct work_struct w;378	struct device *dev;379};380 381struct ifs_device {382	const struct ifs_test_caps *test_caps;383	const struct ifs_test_msrs *test_msrs;384	struct ifs_data rw_data;385	struct miscdevice misc;386};387 388static inline struct ifs_data *ifs_get_data(struct device *dev)389{390	struct miscdevice *m = dev_get_drvdata(dev);391	struct ifs_device *d = container_of(m, struct ifs_device, misc);392 393	return &d->rw_data;394}395 396static inline const struct ifs_test_caps *ifs_get_test_caps(struct device *dev)397{398	struct miscdevice *m = dev_get_drvdata(dev);399	struct ifs_device *d = container_of(m, struct ifs_device, misc);400 401	return d->test_caps;402}403 404static inline const struct ifs_test_msrs *ifs_get_test_msrs(struct device *dev)405{406	struct miscdevice *m = dev_get_drvdata(dev);407	struct ifs_device *d = container_of(m, struct ifs_device, misc);408 409	return d->test_msrs;410}411 412extern bool *ifs_pkg_auth;413int ifs_load_firmware(struct device *dev);414int do_core_test(int cpu, struct device *dev);415extern struct attribute *plat_ifs_attrs[];416extern struct attribute *plat_ifs_array_attrs[];417 418#endif419