diff --git a/.crash.metadata b/.crash.metadata index 7b42f11..44210de 100644 --- a/.crash.metadata +++ b/.crash.metadata @@ -1,2 +1,2 @@ -d154d57291952baea7929b706d76a91edd07552e SOURCES/crash-8.0.3.tar.gz +335ab5dfe04f5265cf5f7bb5a44d6ee0afad1bdc SOURCES/crash-8.0.4.tar.gz 6bf5ee7877a4740835745ed97ce525a00bb2232c SOURCES/gdb-10.2.tar.gz diff --git a/.gitignore b/.gitignore index 6232913..665cca0 100644 --- a/.gitignore +++ b/.gitignore @@ -1,2 +1,2 @@ -SOURCES/crash-8.0.3.tar.gz +SOURCES/crash-8.0.4.tar.gz SOURCES/gdb-10.2.tar.gz diff --git a/SOURCES/0001-Fix-rd-command-for-zram-data-display-in-Linux-6.2-an.patch b/SOURCES/0001-Fix-rd-command-for-zram-data-display-in-Linux-6.2-an.patch new file mode 100644 index 0000000..56d264d --- /dev/null +++ b/SOURCES/0001-Fix-rd-command-for-zram-data-display-in-Linux-6.2-an.patch @@ -0,0 +1,142 @@ +From 38acd02c7fc09843ffb10fc2d695cccdd10cc7f6 Mon Sep 17 00:00:00 2001 +From: Chengen Du +Date: Fri, 17 Nov 2023 11:45:33 +0800 +Subject: [PATCH 01/14] Fix "rd" command for zram data display in Linux 6.2 and + later + +Kernel commit 7ac07a26dea7 ("zram: preparation for multi-zcomp support") +replaced "compressor" member with "comp_algs" in the zram struct. +Without the patch, the "rd" command can triggers the following error: + + rd: WARNING: Some pages are swapped out to zram. Please run mod -s zram. + rd: invalid user virtual address: ffff7d23f010 type: "64-bit UVADDR" + +Related kernel commit: + 84b33bf78889 ("zram: introduce recompress sysfs knob") + +Signed-off-by: Chengen Du +Signed-off-by: Kazuhito Hagio +Signed-off-by: Lianbo Jiang +--- + defs.h | 1 + + diskdump.c | 47 ++++++++++++++++++++++++++++++----------------- + 2 files changed, 31 insertions(+), 17 deletions(-) + +diff --git a/defs.h b/defs.h +index 788f63ada739..2cae5b61e589 100644 +--- a/defs.h ++++ b/defs.h +@@ -2227,6 +2227,7 @@ struct offset_table { /* stash of commonly-used offsets */ + long module_memory_size; + long irq_data_irq; + long zspage_huge; ++ long zram_comp_algs; + }; + + struct size_table { /* stash of commonly-used sizes */ +diff --git a/diskdump.c b/diskdump.c +index 0fe46f4644d0..25054d96313e 100644 +--- a/diskdump.c ++++ b/diskdump.c +@@ -2757,6 +2757,8 @@ diskdump_device_dump_info(FILE *ofp) + + static ulong ZRAM_FLAG_SHIFT; + static ulong ZRAM_FLAG_SAME_BIT; ++static ulong ZRAM_COMP_PRIORITY_BIT1; ++static ulong ZRAM_COMP_PRIORITY_MASK; + + static void + zram_init(void) +@@ -2765,6 +2767,8 @@ zram_init(void) + + MEMBER_OFFSET_INIT(zram_mempoll, "zram", "mem_pool"); + MEMBER_OFFSET_INIT(zram_compressor, "zram", "compressor"); ++ if (INVALID_MEMBER(zram_compressor)) ++ MEMBER_OFFSET_INIT(zram_comp_algs, "zram", "comp_algs"); + MEMBER_OFFSET_INIT(zram_table_flag, "zram_table_entry", "flags"); + if (INVALID_MEMBER(zram_table_flag)) + MEMBER_OFFSET_INIT(zram_table_flag, "zram_table_entry", "value"); +@@ -2782,6 +2786,8 @@ zram_init(void) + + ZRAM_FLAG_SHIFT = 1 << zram_flag_shift; + ZRAM_FLAG_SAME_BIT = 1 << (zram_flag_shift+1); ++ ZRAM_COMP_PRIORITY_BIT1 = ZRAM_FLAG_SHIFT + 7; ++ ZRAM_COMP_PRIORITY_MASK = 0x3; + + if (CRASHDEBUG(1)) + fprintf(fp, "zram_flag_shift: %ld\n", zram_flag_shift); +@@ -2981,9 +2987,9 @@ try_zram_decompress(ulonglong pte_val, unsigned char *buf, ulong len, ulonglong + ulong zram, zram_table_entry, sector, index, entry, flags, size, + outsize, off; + +- if (INVALID_MEMBER(zram_compressor)) { ++ if (INVALID_MEMBER(zram_mempoll)) { + zram_init(); +- if (INVALID_MEMBER(zram_compressor)) { ++ if (INVALID_MEMBER(zram_mempoll)) { + error(WARNING, + "Some pages are swapped out to zram. " + "Please run mod -s zram.\n"); +@@ -2997,8 +3003,28 @@ try_zram_decompress(ulonglong pte_val, unsigned char *buf, ulong len, ulonglong + if (!get_disk_name_private_data(pte_val, vaddr, NULL, &zram)) + return 0; + +- readmem(zram + OFFSET(zram_compressor), KVADDR, name, +- sizeof(name), "zram compressor", FAULT_ON_ERROR); ++ if (THIS_KERNEL_VERSION >= LINUX(2, 6, 0)) ++ swp_offset = (ulonglong)__swp_offset(pte_val); ++ else ++ swp_offset = (ulonglong)SWP_OFFSET(pte_val); ++ ++ sector = swp_offset << (PAGESHIFT() - 9); ++ index = sector >> SECTORS_PER_PAGE_SHIFT; ++ readmem(zram, KVADDR, &zram_table_entry, ++ sizeof(void *), "zram_table_entry", FAULT_ON_ERROR); ++ zram_table_entry += (index * SIZE(zram_table_entry)); ++ readmem(zram_table_entry + OFFSET(zram_table_flag), KVADDR, &flags, ++ sizeof(void *), "zram_table_flag", FAULT_ON_ERROR); ++ if (VALID_MEMBER(zram_compressor)) ++ readmem(zram + OFFSET(zram_compressor), KVADDR, name, sizeof(name), ++ "zram compressor", FAULT_ON_ERROR); ++ else { ++ ulong comp_alg_addr; ++ uint32_t prio = (flags >> ZRAM_COMP_PRIORITY_BIT1) & ZRAM_COMP_PRIORITY_MASK; ++ readmem(zram + OFFSET(zram_comp_algs) + sizeof(const char *) * prio, KVADDR, ++ &comp_alg_addr, sizeof(comp_alg_addr), "zram comp_algs", FAULT_ON_ERROR); ++ read_string(comp_alg_addr, name, sizeof(name)); ++ } + if (STREQ(name, "lzo")) { + #ifdef LZO + if (!(dd->flags & LZO_SUPPORTED)) { +@@ -3019,12 +3045,6 @@ try_zram_decompress(ulonglong pte_val, unsigned char *buf, ulong len, ulonglong + return 0; + } + +- if (THIS_KERNEL_VERSION >= LINUX(2, 6, 0)) { +- swp_offset = (ulonglong)__swp_offset(pte_val); +- } else { +- swp_offset = (ulonglong)SWP_OFFSET(pte_val); +- } +- + zram_buf = (unsigned char *)GETBUF(PAGESIZE()); + /* lookup page from swap cache */ + off = PAGEOFFSET(vaddr); +@@ -3034,15 +3054,8 @@ try_zram_decompress(ulonglong pte_val, unsigned char *buf, ulong len, ulonglong + goto out; + } + +- sector = swp_offset << (PAGESHIFT() - 9); +- index = sector >> SECTORS_PER_PAGE_SHIFT; +- readmem(zram, KVADDR, &zram_table_entry, +- sizeof(void *), "zram_table_entry", FAULT_ON_ERROR); +- zram_table_entry += (index * SIZE(zram_table_entry)); + readmem(zram_table_entry, KVADDR, &entry, + sizeof(void *), "entry of table", FAULT_ON_ERROR); +- readmem(zram_table_entry + OFFSET(zram_table_flag), KVADDR, &flags, +- sizeof(void *), "zram_table_flag", FAULT_ON_ERROR); + if (!entry || (flags & ZRAM_FLAG_SAME_BIT)) { + int count; + ulong *same_buf = (ulong *)GETBUF(PAGESIZE()); +-- +2.41.0 + diff --git a/SOURCES/0001-symbols-skip-the-module-if-the-given-address-is-not-.patch b/SOURCES/0001-symbols-skip-the-module-if-the-given-address-is-not-.patch new file mode 100644 index 0000000..bd6989b --- /dev/null +++ b/SOURCES/0001-symbols-skip-the-module-if-the-given-address-is-not-.patch @@ -0,0 +1,87 @@ +From 28891d1127542dbb2d5ba16c575e14e741ed73ef Mon Sep 17 00:00:00 2001 +From: Tao Liu +Date: Thu, 4 Jan 2024 09:20:27 +0800 +Subject: [PATCH] symbols: skip the module if the given address is not within + its address range + +Previously, to find a module symbol and its offset by an arbitrary address, +all symbols within the module will be iterated by address ascending order +until the last symbol with a smaller address been noticed. + +However if the address is not within the module address range, e.g. +the address is higher than the module's last symbol's address, then +the module can be surely skipped, because its symbol iteration is +unnecessary. This can speed up the kernel module symbols finding and improve +the overall performance. + +Without the patch: + $ time echo "bt 8993" | ~/crash-dev/crash vmcore vmlinux + crash> bt 8993 + PID: 8993 TASK: ffff927569cc2100 CPU: 2 COMMAND: "WriterPool0" + #0 [ffff927569cd76f0] __schedule at ffffffffb3db78d8 + #1 [ffff927569cd7758] schedule_preempt_disabled at ffffffffb3db8bf9 + #2 [ffff927569cd7768] __mutex_lock_slowpath at ffffffffb3db6ca7 + #3 [ffff927569cd77c0] mutex_lock at ffffffffb3db602f + #4 [ffff927569cd77d8] ucache_retrieve at ffffffffc0cf4409 [secfs2] + ...snip the stacktrace of the same module... + #11 [ffff927569cd7ba0] cskal_path_vfs_getattr_nosec at ffffffffc05cae76 [falcon_kal] + ...snip... + #13 [ffff927569cd7c40] _ZdlPv at ffffffffc086e751 [falcon_lsm_serviceable] + ...snip... + #20 [ffff927569cd7ef8] unload_network_ops_symbols at ffffffffc06f11c0 [falcon_lsm_pinned_14713] + #21 [ffff927569cd7f50] system_call_fastpath at ffffffffb3dc539a + RIP: 00007f2b28ed4023 RSP: 00007f2a45fe7f80 RFLAGS: 00000206 + RAX: 0000000000000012 RBX: 00007f2a68302e00 RCX: 00007f2a682546d8 + RDX: 0000000000000826 RSI: 00007eb57ea6a000 RDI: 00000000000000e3 + RBP: 00007eb57ea6a000 R8: 0000000000000826 R9: 00000002670bdfd2 + R10: 00000002670bdfd2 R11: 0000000000000293 R12: 00000002670bdfd2 + R13: 00007f29d501a480 R14: 0000000000000826 R15: 00000002670bdfd2 + ORIG_RAX: 0000000000000012 CS: 0033 SS: 002b + crash> + real 7m14.826s + user 7m12.502s + sys 0m1.091s + +With the patch: + $ time echo "bt 8993" | ~/crash-dev/crash vmcore vmlinux + crash> bt 8993 + PID: 8993 TASK: ffff927569cc2100 CPU: 2 COMMAND: "WriterPool0" + #0 [ffff927569cd76f0] __schedule at ffffffffb3db78d8 + #1 [ffff927569cd7758] schedule_preempt_disabled at ffffffffb3db8bf9 + ...snip the same output... + crash> + real 0m8.827s + user 0m7.896s + sys 0m0.938s + +Signed-off-by: Tao Liu +--- + symbols.c | 5 ++++- + 1 file changed, 4 insertions(+), 1 deletion(-) + +diff --git a/symbols.c b/symbols.c +index 5d91991..88a3fd1 100644 +--- a/symbols.c ++++ b/symbols.c +@@ -5561,7 +5561,7 @@ value_search_module_6_4(ulong value, ulong *offset) + sp = lm->symtable[t]; + sp_end = lm->symend[t]; + +- if (value < sp->value) ++ if (value < sp->value || value > sp_end->value) + continue; + + splast = NULL; +@@ -5646,6 +5646,9 @@ retry: + if (sp->value > value) /* invalid -- between modules */ + break; + ++ if (sp_end->value < value) /* not within the module */ ++ continue; ++ + /* + * splast will contain the last module symbol encountered. + * Note: "__insmod_"-type symbols will be set in splast only +-- +2.40.1 + diff --git a/SOURCES/0002-Fix-typos-in-offset_table-and-missing-help-o-items.patch b/SOURCES/0002-Fix-typos-in-offset_table-and-missing-help-o-items.patch new file mode 100644 index 0000000..f032387 --- /dev/null +++ b/SOURCES/0002-Fix-typos-in-offset_table-and-missing-help-o-items.patch @@ -0,0 +1,168 @@ +From d65e5d3eae0dd06a5308a5cb00c05fee60594093 Mon Sep 17 00:00:00 2001 +From: Kazuhito Hagio +Date: Mon, 20 Nov 2023 13:22:56 +0900 +Subject: [PATCH 02/14] Fix typos in offset_table and missing "help -o" items + +A few of zram related members in the offset_table have typos and +irregular naming rule, also they are not present in the "help -o" +output. Let's fix these. + +Signed-off-by: Kazuhito Hagio +Signed-off-by: Lianbo Jiang +--- + defs.h | 8 ++++---- + diskdump.c | 24 ++++++++++++------------ + memory.c | 2 +- + symbols.c | 12 ++++++++++++ + 4 files changed, 29 insertions(+), 17 deletions(-) + +diff --git a/defs.h b/defs.h +index 2cae5b61e589..5218a94fe4a4 100644 +--- a/defs.h ++++ b/defs.h +@@ -2112,13 +2112,13 @@ struct offset_table { /* stash of commonly-used offsets */ + long bpf_prog_aux_name; + long page_private; + long swap_info_struct_bdev; +- long zram_mempoll; ++ long zram_mem_pool; + long zram_compressor; +- long zram_table_flag; +- long zspoll_size_class; ++ long zram_table_entry_flags; ++ long zs_pool_size_class; + long size_class_size; + long gendisk_private_data; +- long zram_table_entry; ++ long zram_table_entry; /* unused; but cannot remove */ + long module_core_size_rw; + long module_core_size_rx; + long module_init_size_rw; +diff --git a/diskdump.c b/diskdump.c +index 25054d96313e..f20f3ac519a1 100644 +--- a/diskdump.c ++++ b/diskdump.c +@@ -2765,15 +2765,15 @@ zram_init(void) + { + long zram_flag_shift; + +- MEMBER_OFFSET_INIT(zram_mempoll, "zram", "mem_pool"); ++ MEMBER_OFFSET_INIT(zram_mem_pool, "zram", "mem_pool"); + MEMBER_OFFSET_INIT(zram_compressor, "zram", "compressor"); + if (INVALID_MEMBER(zram_compressor)) + MEMBER_OFFSET_INIT(zram_comp_algs, "zram", "comp_algs"); +- MEMBER_OFFSET_INIT(zram_table_flag, "zram_table_entry", "flags"); +- if (INVALID_MEMBER(zram_table_flag)) +- MEMBER_OFFSET_INIT(zram_table_flag, "zram_table_entry", "value"); ++ MEMBER_OFFSET_INIT(zram_table_entry_flags, "zram_table_entry", "flags"); ++ if (INVALID_MEMBER(zram_table_entry_flags)) ++ MEMBER_OFFSET_INIT(zram_table_entry_flags, "zram_table_entry", "value"); + STRUCT_SIZE_INIT(zram_table_entry, "zram_table_entry"); +- MEMBER_OFFSET_INIT(zspoll_size_class, "zs_pool", "size_class"); ++ MEMBER_OFFSET_INIT(zs_pool_size_class, "zs_pool", "size_class"); + MEMBER_OFFSET_INIT(size_class_size, "size_class", "size"); + MEMBER_OFFSET_INIT(zspage_huge, "zspage", "huge"); + +@@ -2826,7 +2826,7 @@ zram_object_addr(ulong pool, ulong handle, unsigned char *zram_buf) + if (zs_magic != ZSPAGE_MAGIC) + error(FATAL, "zspage magic incorrect: %x\n", zs_magic); + +- class = pool + OFFSET(zspoll_size_class); ++ class = pool + OFFSET(zs_pool_size_class); + class += (class_idx * sizeof(void *)); + readmem(class, KVADDR, &class, sizeof(void *), "size_class", FAULT_ON_ERROR); + readmem(class + OFFSET(size_class_size), KVADDR, +@@ -2987,9 +2987,9 @@ try_zram_decompress(ulonglong pte_val, unsigned char *buf, ulong len, ulonglong + ulong zram, zram_table_entry, sector, index, entry, flags, size, + outsize, off; + +- if (INVALID_MEMBER(zram_mempoll)) { ++ if (INVALID_MEMBER(zram_mem_pool)) { + zram_init(); +- if (INVALID_MEMBER(zram_mempoll)) { ++ if (INVALID_MEMBER(zram_mem_pool)) { + error(WARNING, + "Some pages are swapped out to zram. " + "Please run mod -s zram.\n"); +@@ -3013,8 +3013,8 @@ try_zram_decompress(ulonglong pte_val, unsigned char *buf, ulong len, ulonglong + readmem(zram, KVADDR, &zram_table_entry, + sizeof(void *), "zram_table_entry", FAULT_ON_ERROR); + zram_table_entry += (index * SIZE(zram_table_entry)); +- readmem(zram_table_entry + OFFSET(zram_table_flag), KVADDR, &flags, +- sizeof(void *), "zram_table_flag", FAULT_ON_ERROR); ++ readmem(zram_table_entry + OFFSET(zram_table_entry_flags), KVADDR, &flags, ++ sizeof(void *), "zram_table_entry.flags", FAULT_ON_ERROR); + if (VALID_MEMBER(zram_compressor)) + readmem(zram + OFFSET(zram_compressor), KVADDR, name, sizeof(name), + "zram compressor", FAULT_ON_ERROR); +@@ -3072,8 +3072,8 @@ try_zram_decompress(ulonglong pte_val, unsigned char *buf, ulong len, ulonglong + goto out; + } + +- readmem(zram + OFFSET(zram_mempoll), KVADDR, &zram, +- sizeof(void *), "zram_mempoll", FAULT_ON_ERROR); ++ readmem(zram + OFFSET(zram_mem_pool), KVADDR, &zram, ++ sizeof(void *), "zram.mem_pool", FAULT_ON_ERROR); + + obj_addr = zram_object_addr(zram, entry, zram_buf); + if (obj_addr == NULL) { +diff --git a/memory.c b/memory.c +index 86ccec5e2bac..791194a405d4 100644 +--- a/memory.c ++++ b/memory.c +@@ -519,7 +519,7 @@ vm_init(void) + "swap_info_struct", "old_block_size"); + MEMBER_OFFSET_INIT(swap_info_struct_bdev, "swap_info_struct", "bdev"); + +- MEMBER_OFFSET_INIT(zspoll_size_class, "zs_pool", "size_class"); ++ MEMBER_OFFSET_INIT(zs_pool_size_class, "zs_pool", "size_class"); + MEMBER_OFFSET_INIT(size_class_size, "size_class", "size"); + + MEMBER_OFFSET_INIT(block_device_bd_inode, "block_device", "bd_inode"); +diff --git a/symbols.c b/symbols.c +index 8e8b4c31d915..176c95026f03 100644 +--- a/symbols.c ++++ b/symbols.c +@@ -10304,6 +10304,7 @@ dump_offset_table(char *spec, ulong makestruct) + OFFSET(page_active)); + fprintf(fp, " page_compound_head: %ld\n", + OFFSET(page_compound_head)); ++ fprintf(fp, " page_private: %ld\n", OFFSET(page_private)); + + fprintf(fp, " trace_print_flags_mask: %ld\n", + OFFSET(trace_print_flags_mask)); +@@ -10330,6 +10331,7 @@ dump_offset_table(char *spec, ulong makestruct) + OFFSET(swap_info_struct_inuse_pages)); + fprintf(fp, "swap_info_struct_old_block_size: %ld\n", + OFFSET(swap_info_struct_old_block_size)); ++ fprintf(fp, " swap_info_struct_bdev: %ld\n", OFFSET(swap_info_struct_bdev)); + fprintf(fp, " block_device_bd_inode: %ld\n", + OFFSET(block_device_bd_inode)); + fprintf(fp, " block_device_bd_list: %ld\n", +@@ -11359,6 +11361,8 @@ dump_offset_table(char *spec, ulong makestruct) + OFFSET(gendisk_part0)); + fprintf(fp, " gendisk_queue: %ld\n", + OFFSET(gendisk_queue)); ++ fprintf(fp, " gendisk_private_data: %ld\n", OFFSET(gendisk_private_data)); ++ + fprintf(fp, " hd_struct_dev: %ld\n", + OFFSET(hd_struct_dev)); + fprintf(fp, " hd_struct_dkstats: %ld\n", +@@ -11765,6 +11769,14 @@ dump_offset_table(char *spec, ulong makestruct) + fprintf(fp, " maple_metadata_end: %ld\n", OFFSET(maple_metadata_end)); + fprintf(fp, " maple_metadata_gap: %ld\n", OFFSET(maple_metadata_gap)); + ++ fprintf(fp, " zram_mem_pool: %ld\n", OFFSET(zram_mem_pool)); ++ fprintf(fp, " zram_compressor: %ld\n", OFFSET(zram_compressor)); ++ fprintf(fp, " zram_comp_algs: %ld\n", OFFSET(zram_comp_algs)); ++ fprintf(fp, " zram_table_entry_flags: %ld\n", OFFSET(zram_table_entry_flags)); ++ fprintf(fp, " zs_pool_size_class: %ld\n", OFFSET(zs_pool_size_class)); ++ fprintf(fp, " size_class_size: %ld\n", OFFSET(size_class_size)); ++ fprintf(fp, " zspage_huge: %ld\n", OFFSET(zspage_huge)); ++ + fprintf(fp, "\n size_table:\n"); + fprintf(fp, " page: %ld\n", SIZE(page)); + fprintf(fp, " page_flags: %ld\n", SIZE(page_flags)); +-- +2.41.0 + diff --git a/SOURCES/0003-zram-Fixes-for-lookup_swap_cache.patch b/SOURCES/0003-zram-Fixes-for-lookup_swap_cache.patch new file mode 100644 index 0000000..74b9e41 --- /dev/null +++ b/SOURCES/0003-zram-Fixes-for-lookup_swap_cache.patch @@ -0,0 +1,78 @@ +From 582febffa8b3567339148c2bb916fc70f2fc546e Mon Sep 17 00:00:00 2001 +From: Johan Erlandsson +Date: Fri, 20 Oct 2023 19:10:52 +0200 +Subject: [PATCH 03/14] zram: Fixes for lookup_swap_cache() + +Fix the following three issues: +(1) swap cache missing page tree offset + The radix or xarray start at an offset inside struct address_space. +(2) swap cache entries are pointer to struct page + The entries in radix, xarray (swap cache) are address to struct page. +(3) exclude shadow entries from swap cache lookup + radix or xarray can contain shadow entries from previous page + entries. These should be ignored when looking for a page pointer. + +Without the patch, +- lookup_swap_cache() returns NULL since do_xarray() call returns FALSE, +- in try_zram_decompress(), since 'entry' is NULL, page is filled with 0, + if (!entry || (flags & ZRAM_FLAG_SAME_BIT)) { +and pages in swap cache will be seen to be a 'zero' page. + +Signed-off-by: Johan Erlandsson +Signed-off-by: Kazuhito Hagio +Signed-off-by: Lianbo Jiang +--- + diskdump.c | 19 ++++++++++++++----- + 1 file changed, 14 insertions(+), 5 deletions(-) + +diff --git a/diskdump.c b/diskdump.c +index f20f3ac519a1..660c25729dad 100644 +--- a/diskdump.c ++++ b/diskdump.c +@@ -27,6 +27,7 @@ + #include "diskdump.h" + #include "xen_dom0.h" + #include "vmcore.h" ++#include "maple_tree.h" + + #define BITMAP_SECT_LEN 4096 + +@@ -2877,11 +2878,16 @@ out: + return zram_buf; + } + ++static inline bool radix_tree_exceptional_entry(ulong entry) ++{ ++ return entry & RADIX_TREE_EXCEPTIONAL_ENTRY; ++} ++ + static unsigned char * + lookup_swap_cache(ulonglong pte_val, unsigned char *zram_buf) + { + ulonglong swp_offset; +- ulong swp_type, swp_space, page; ++ ulong swp_type, swp_space; + struct list_pair lp; + physaddr_t paddr; + static int is_xarray = -1; +@@ -2907,10 +2913,13 @@ lookup_swap_cache(ulonglong pte_val, unsigned char *zram_buf) + swp_space += (swp_offset >> SWAP_ADDRESS_SPACE_SHIFT) * SIZE(address_space); + + lp.index = swp_offset; +- if ((is_xarray ? do_xarray : do_radix_tree)(swp_space, RADIX_TREE_SEARCH, &lp)) { +- readmem((ulong)lp.value, KVADDR, &page, sizeof(void *), +- "swap_cache page", FAULT_ON_ERROR); +- if (!is_page_ptr(page, &paddr)) { ++ if ((is_xarray ? do_xarray : do_radix_tree) ++ (swp_space+OFFSET(address_space_page_tree), RADIX_TREE_SEARCH, &lp)) { ++ if ((is_xarray ? xa_is_value : radix_tree_exceptional_entry)((ulong)lp.value)) { ++ /* ignore shadow values */ ++ return NULL; ++ } ++ if (!is_page_ptr((ulong)lp.value, &paddr)) { + error(WARNING, "radix page: %lx: not a page pointer\n", lp.value); + return NULL; + } +-- +2.41.0 + diff --git a/SOURCES/0004-symbols-expand-all-kernel-module-symtable-if-not-all.patch b/SOURCES/0004-symbols-expand-all-kernel-module-symtable-if-not-all.patch new file mode 100644 index 0000000..2873a4a --- /dev/null +++ b/SOURCES/0004-symbols-expand-all-kernel-module-symtable-if-not-all.patch @@ -0,0 +1,171 @@ +From f2ee6fa6c841ddc37ba665909dafbc7294c34d64 Mon Sep 17 00:00:00 2001 +From: Tao Liu +Date: Fri, 17 Nov 2023 15:52:19 +0800 +Subject: [PATCH 04/14] symbols: expand all kernel module symtable if not all + expanded previously + +There is an issue that, for kernel modules, "dis -rl" fails to display +modules code line number data after execute "bt" command in crash. + +Without the patch: + crsah> mod -S + crash> bt + PID: 1500 TASK: ff2bd8b093524000 CPU: 16 COMMAND: "lpfc_worker_0" + #0 [ff2c9f725c39f9e0] machine_kexec at ffffffff8e0686d3 + ...snip... + #8 [ff2c9f725c39fcc0] __lpfc_sli_release_iocbq_s4 at ffffffffc0f2f425 [lpfc] + ...snip... + crash> dis -rl ffffffffc0f60f82 + 0xffffffffc0f60eb0 : nopl 0x0(%rax,%rax,1) [FTRACE NOP] + 0xffffffffc0f60eb5 : push %rbp + 0xffffffffc0f60eb6 : push %rbx + 0xffffffffc0f60eb7 : test %rdi,%rdi + +With the patch: + crash> mod -S + crash> bt + PID: 1500 TASK: ff2bd8b093524000 CPU: 16 COMMAND: "lpfc_worker_0" + #0 [ff2c9f725c39f9e0] machine_kexec at ffffffff8e0686d3 + ...snip... + #8 [ff2c9f725c39fcc0] __lpfc_sli_release_iocbq_s4 at ffffffffc0f2f425 [lpfc] + ...snip... + crash> dis -rl ffffffffc0f60f82 + /usr/src/debug/kernel-4.18.0-425.13.1.el8_7/linux-4.18.0-425.13.1.el8_7.x86_64/drivers/scsi/lpfc/lpfc_hbadisc.c: 6756 + 0xffffffffc0f60eb0 : nopl 0x0(%rax,%rax,1) [FTRACE NOP] + /usr/src/debug/kernel-4.18.0-425.13.1.el8_7/linux-4.18.0-425.13.1.el8_7.x86_64/drivers/scsi/lpfc/lpfc_hbadisc.c: 6759 + 0xffffffffc0f60eb5 : push %rbp + +The root cause is, after kernel module been loaded by mod command, the symtable +is not expanded in gdb side. crash bt or dis command will trigger such an +expansion. However the symtable expansion is different for the 2 commands: + +The stack trace of "dis -rl" for symtable expanding: + + #0 0x00000000008d8d9f in add_compunit_symtab_to_objfile ... + #1 0x00000000006d3293 in buildsym_compunit::end_symtab_with_blockvector ... + #2 0x00000000006d336a in buildsym_compunit::end_symtab_from_static_block ... + #3 0x000000000077e8e9 in process_full_comp_unit ... + #4 process_queue ... + #5 dw2_do_instantiate_symtab ... + #6 0x000000000077ed67 in dw2_instantiate_symtab ... + #7 0x000000000077f75e in dw2_expand_all_symtabs ... + #8 0x00000000008f254d in gdb_get_line_number ... + #9 0x00000000008f22af in gdb_command_funnel_1 ... + #10 0x00000000008f2003 in gdb_command_funnel ... + #11 0x00000000005b7f02 in gdb_interface ... + #12 0x00000000005f5bd8 in get_line_number ... + #13 0x000000000059e574 in cmd_dis ... + +The stack trace of "bt" for symtable expanding: + + #0 0x00000000008d8d9f in add_compunit_symtab_to_objfile ... + #1 0x00000000006d3293 in buildsym_compunit::end_symtab_with_blockvector ... + #2 0x00000000006d336a in buildsym_compunit::end_symtab_from_static_block ... + #3 0x000000000077e8e9 in process_full_comp_unit ... + #4 process_queue ... + #5 dw2_do_instantiate_symtab ... + #6 0x000000000077ed67 in dw2_instantiate_symtab ... + #7 0x000000000077f8ed in dw2_lookup_symbol ... + #8 0x00000000008e6d03 in lookup_symbol_via_quick_fns ... + #9 0x00000000008e7153 in lookup_symbol_in_objfile ... + #10 0x00000000008e73c6 in lookup_symbol_global_or_static_iterator_cb ... + #11 0x00000000008b99c4 in svr4_iterate_over_objfiles_in_search_order ... + #12 0x00000000008e754e in lookup_global_or_static_symbol ... + #13 0x00000000008e75da in lookup_static_symbol ... + #14 0x00000000008e632c in lookup_symbol_aux ... + #15 0x00000000008e5a7a in lookup_symbol_in_language ... + #16 0x00000000008e5b30 in lookup_symbol ... + #17 0x00000000008f2a4a in gdb_get_datatype ... + #18 0x00000000008f22c0 in gdb_command_funnel_1 ... + #19 0x00000000008f2003 in gdb_command_funnel ... + #20 0x00000000005b7f02 in gdb_interface ... + #21 0x00000000005f8a9f in datatype_info ... + #22 0x0000000000599947 in cpu_map_size ... + #23 0x00000000005a975d in get_cpus_online ... + #24 0x0000000000637a8b in diskdump_get_prstatus_percpu ... + #25 0x000000000062f0e4 in get_netdump_regs_x86_64 ... + #26 0x000000000059fe68 in back_trace ... + #27 0x00000000005ab1cb in cmd_bt ... + +For the stacktrace of "dis -rl", it calls dw2_expand_all_symtabs() to expand +all symtable of the objfile, or "*.ko.debug" in our case. However for +the stacktrace of "bt", it doesn't expand all, but only a subset of symtable +which is enough to find a symbol by dw2_lookup_symbol(). As a result, the +objfile->compunit_symtabs, which is the head of a single linked list of +struct compunit_symtab, is not NULL but didn't contain all symtables. It +will not be reinitialized in gdb_get_line_number() by "dis -rl" because +!objfile_has_full_symbols(objfile) check will fail, so it cannot display +the proper code line number data. + +Since objfile_has_full_symbols(objfile) check cannot ensure all symbols +been expanded, this patch add a new member as a flag for struct objfile +to record if all symbols have been expanded. The flag will be set only ofter +expand_all_symtabs been called. + +Signed-off-by: Tao Liu +Signed-off-by: Lianbo Jiang +--- + gdb-10.2.patch | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++ + 1 file changed, 50 insertions(+) + +diff --git a/gdb-10.2.patch b/gdb-10.2.patch +index d81030d946e8..2f7d585105aa 100644 +--- a/gdb-10.2.patch ++++ b/gdb-10.2.patch +@@ -3187,3 +3187,53 @@ exit 0 + result = stringtab + symbol_entry->_n._n_n._n_offset; + } + else ++--- gdb-10.2/gdb/objfiles.h.orig +++++ gdb-10.2/gdb/objfiles.h ++@@ -712,6 +712,8 @@ struct objfile ++ next time. If an objfile does not have the symbols, it will ++ never have them. */ ++ bool skip_jit_symbol_lookup = false; +++ +++ bool all_symtabs_expanded = false; ++ }; ++ ++ /* A deleter for objfile. */ ++--- gdb-10.2/gdb/symfile.c.orig +++++ gdb-10.2/gdb/symfile.c ++@@ -1133,8 +1133,10 @@ symbol_file_add_with_addrs (bfd *abfd, const char *name, ++ printf_filtered (_("Expanding full symbols from %ps...\n"), ++ styled_string (file_name_style.style (), name)); ++ ++- if (objfile->sf) +++ if (objfile->sf) { ++ objfile->sf->qf->expand_all_symtabs (objfile); +++ objfile->all_symtabs_expanded = true; +++ } ++ } ++ ++ /* Note that we only print a message if we have no symbols and have ++--- gdb-10.2/gdb/symtab.c.orig +++++ gdb-10.2/gdb/symtab.c ++@@ -7097,8 +7097,9 @@ gdb_get_line_number(struct gnu_request *req) ++ */ ++ if (req->lm) { ++ objfile = req->lm->loaded_objfile; ++- if (!objfile_has_full_symbols(objfile) && objfile->sf) { +++ if (!objfile->all_symtabs_expanded && objfile->sf) { ++ objfile->sf->qf->expand_all_symtabs(objfile); +++ objfile->all_symtabs_expanded = true; ++ sal = find_pc_line(pc, 0); ++ } ++ } ++@@ -7761,8 +7765,10 @@ iterate_datatypes (struct gnu_request *req) ++ { ++ for (objfile *objfile : current_program_space->objfiles ()) ++ { ++- if (objfile->sf) +++ if (objfile->sf) { ++ objfile->sf->qf->expand_all_symtabs(objfile); +++ objfile->all_symtabs_expanded = true; +++ } ++ ++ for (compunit_symtab *cust : objfile->compunits ()) ++ { +-- +2.41.0 + diff --git a/SOURCES/0005-symbols-skip-load-.init.-sections-if-module-was-succ.patch b/SOURCES/0005-symbols-skip-load-.init.-sections-if-module-was-succ.patch new file mode 100644 index 0000000..fd8d189 --- /dev/null +++ b/SOURCES/0005-symbols-skip-load-.init.-sections-if-module-was-succ.patch @@ -0,0 +1,67 @@ +From 0c5ef6a4a3a2759915ffe72b1366dce2f32f65c5 Mon Sep 17 00:00:00 2001 +From: Tao Liu +Date: Tue, 14 Nov 2023 16:32:07 +0800 +Subject: [PATCH 05/14] symbols: skip load .init.* sections if module was + successfully initialized + +There might be address overlap of one modules .init.text symbols and +another modules .text symbols. As a result, gdb fails to translate the +address to symbol name correctly: + + crash> sym -m virtio_blk | grep MODULE + ffffffffc00a4000 MODULE START: virtio_blk + ffffffffc00a86ec MODULE END: virtio_blk + crash> gdb info address floppy_module_init + Symbol "floppy_module_init" is a function at address 0xffffffffc00a4131. + +Since the .init.* sections of a module had been freed by kernel if the +module was initialized successfully, there is no need to load the .init.* +sections data from "*.ko.debug" in gdb to create such an overlap. +lm->mod_init_module_ptr is used as a flag of whether module is freed. + +Without the patch: + crash> mod -S + crash> struct blk_mq_ops 0xffffffffc00a7160 + struct blk_mq_ops { + queue_rq = 0xffffffffc00a45b0 , <-- translated from module floppy + map_queue = 0xffffffff813015c0 , + ...snip... + complete = 0xffffffffc00a4370 , + init_request = 0xffffffffc00a4260 , + ...snip... + } + +With the patch: + crash> mod -S + crash> struct blk_mq_ops 0xffffffffc00a7160 + struct blk_mq_ops { + queue_rq = 0xffffffffc00a45b0 , <-- translated from module virtio_blk + map_queue = 0xffffffff813015c0 , + ...snip... + complete = 0xffffffffc00a4370 , + init_request = 0xffffffffc00a4260 , + ...snip... + } + +Signed-off-by: Tao Liu +Signed-off-by: Lianbo Jiang +--- + symbols.c | 2 +- + 1 file changed, 1 insertion(+), 1 deletion(-) + +diff --git a/symbols.c b/symbols.c +index 176c95026f03..5d919910164e 100644 +--- a/symbols.c ++++ b/symbols.c +@@ -13295,7 +13295,7 @@ add_symbol_file_kallsyms(struct load_module *lm, struct gnu_request *req) + shift_string_right(req->buf, strlen(buf)); + BCOPY(buf, req->buf, strlen(buf)); + retval = TRUE; +- } else { ++ } else if (lm->mod_init_module_ptr || !STRNEQ(section_name, ".init.")) { + sprintf(buf, " -s %s 0x%lx", section_name, section_vaddr); + while ((len + strlen(buf)) >= buflen) { + RESIZEBUF(req->buf, buflen, buflen * 2); +-- +2.41.0 + diff --git a/SOURCES/0006-use-NR_SWAPCACHE-when-nr_swapper_spaces-isn-t-availa.patch b/SOURCES/0006-use-NR_SWAPCACHE-when-nr_swapper_spaces-isn-t-availa.patch new file mode 100644 index 0000000..0f0d86c --- /dev/null +++ b/SOURCES/0006-use-NR_SWAPCACHE-when-nr_swapper_spaces-isn-t-availa.patch @@ -0,0 +1,47 @@ +From c15da07526291a5c357010cb4aaf4bde6151e642 Mon Sep 17 00:00:00 2001 +From: Johan Erlandsson +Date: Wed, 19 Apr 2023 11:26:04 +0200 +Subject: [PATCH 06/14] use NR_SWAPCACHE when nr_swapper_spaces isn't available + +In 5.12 the following change was introduced: +b6038942480e ("mm: memcg: add swapcache stat for memcg v2") + +Then the variable 'nr_swapper_spaces' is not read (unless +CONFIG_DEBUG_VM=y). In GKI builds this variable is then optimized +out. But the same change provided a new way to obtain the same +information, using NR_SWAPCACHE. + +Reported-by: xueguolun +Signed-off-by: Johan Erlandsson +Signed-off-by: Lianbo Jiang +--- + memory.c | 6 ++++-- + 1 file changed, 4 insertions(+), 2 deletions(-) + +diff --git a/memory.c b/memory.c +index 791194a405d4..b84e974a3325 100644 +--- a/memory.c ++++ b/memory.c +@@ -8486,7 +8486,7 @@ dump_kmeminfo(void) + ulong hugetlb_total_pages, hugetlb_total_free_pages = 0; + int done_hugetlb_calc = 0; + long nr_file_pages, nr_slab; +- ulong swapper_space_nrpages; ++ long swapper_space_nrpages; + ulong pct; + uint tmp; + struct meminfo meminfo; +@@ -8609,7 +8609,9 @@ dump_kmeminfo(void) + char *swapper_space = GETBUF(SIZE(address_space)); + + swapper_space_nrpages = 0; +- if (symbol_exists("nr_swapper_spaces") && ++ if (dump_vm_stat("NR_SWAPCACHE", &swapper_space_nrpages, 0)) { ++ ; ++ } else if (symbol_exists("nr_swapper_spaces") && + (len = get_array_length("nr_swapper_spaces", + NULL, 0))) { + char *nr_swapper_space = +-- +2.41.0 + diff --git a/SOURCES/0007-Fix-identity_map_base-value-dump-on-S390.patch b/SOURCES/0007-Fix-identity_map_base-value-dump-on-S390.patch new file mode 100644 index 0000000..3f9fb2d --- /dev/null +++ b/SOURCES/0007-Fix-identity_map_base-value-dump-on-S390.patch @@ -0,0 +1,43 @@ +From 2e513114e7d77fadc88011f186ef943ccf397d35 Mon Sep 17 00:00:00 2001 +From: Alexander Gordeev +Date: Wed, 29 Nov 2023 13:47:34 +0100 +Subject: [PATCH 07/14] Fix identity_map_base value dump on S390 + +Kernel virtual base instead of identity base is printed + +Signed-off-by: Alexander Gordeev +Signed-off-by: Lianbo Jiang +--- + s390.c | 2 +- + s390x.c | 2 +- + 2 files changed, 2 insertions(+), 2 deletions(-) + +diff --git a/s390.c b/s390.c +index 42f5cc63ae52..a8b2bcca86c7 100644 +--- a/s390.c ++++ b/s390.c +@@ -183,7 +183,7 @@ s390_dump_machdep_table(ulong arg) + fprintf(fp, ")\n"); + + fprintf(fp, " kvbase: %lx\n", machdep->kvbase); +- fprintf(fp, " identity_map_base: %lx\n", machdep->kvbase); ++ fprintf(fp, " identity_map_base: %lx\n", machdep->identity_map_base); + fprintf(fp, " pagesize: %d\n", machdep->pagesize); + fprintf(fp, " pageshift: %d\n", machdep->pageshift); + fprintf(fp, " pagemask: %llx\n", machdep->pagemask); +diff --git a/s390x.c b/s390x.c +index d7ee3755fc0b..096c072186f5 100644 +--- a/s390x.c ++++ b/s390x.c +@@ -650,7 +650,7 @@ s390x_dump_machdep_table(ulong arg) + fprintf(fp, ")\n"); + + fprintf(fp, " kvbase: %lx\n", machdep->kvbase); +- fprintf(fp, " identity_map_base: %lx\n", machdep->kvbase); ++ fprintf(fp, " identity_map_base: %lx\n", machdep->identity_map_base); + fprintf(fp, " pagesize: %d\n", machdep->pagesize); + fprintf(fp, " pageshift: %d\n", machdep->pageshift); + fprintf(fp, " pagemask: %llx\n", machdep->pagemask); +-- +2.41.0 + diff --git a/SOURCES/0008-s390x-fix-virtual-vs-physical-address-confusion.patch b/SOURCES/0008-s390x-fix-virtual-vs-physical-address-confusion.patch new file mode 100644 index 0000000..9d439d4 --- /dev/null +++ b/SOURCES/0008-s390x-fix-virtual-vs-physical-address-confusion.patch @@ -0,0 +1,70 @@ +From 4c78eb4a9199631fe94845cb3fbd6376aae1251d Mon Sep 17 00:00:00 2001 +From: Alexander Gordeev +Date: Wed, 29 Nov 2023 13:47:35 +0100 +Subject: [PATCH 08/14] s390x: fix virtual vs physical address confusion + +Physical and virtual addresses are the same on S390X. +That led to missing to use PTOV and VTOP macros where +they actually expected. + +Signed-off-by: Alexander Gordeev +Signed-off-by: Lianbo Jiang +--- + s390x.c | 12 ++++++------ + 1 file changed, 6 insertions(+), 6 deletions(-) + +diff --git a/s390x.c b/s390x.c +index 096c072186f5..957b839a5fa9 100644 +--- a/s390x.c ++++ b/s390x.c +@@ -311,7 +311,7 @@ static struct s390x_cpu *s390x_cpu_get(struct bt_info *bt) + readmem(lowcore_ptr + cpu * sizeof(long), KVADDR, + &prefix, sizeof(long), "lowcore_ptr", FAULT_ON_ERROR); + for (i = 0; i < s390x_cpu_cnt; i++) { +- if (s390x_cpu_vec[i].prefix == prefix) ++ if (s390x_cpu_vec[i].prefix == VTOP(prefix)) + return &s390x_cpu_vec[i]; + } + error(FATAL, "cannot determine CPU for task: %lx\n", bt->task); +@@ -985,12 +985,12 @@ int s390x_vtop(ulong table, ulong vaddr, physaddr_t *phys_addr, int verbose) + verbose); + if (!entry) + return FALSE; +- table = entry & ~0xfffULL; ++ table = PTOV(entry & ~0xfffULL); + /* Check if this a 2GB page */ + if ((entry & 0x400ULL) && (level == 1)) { + /* Add the 2GB frame offset & return the final value. */ + table &= ~0x7fffffffULL; +- *phys_addr = table + (vaddr & 0x7fffffffULL); ++ *phys_addr = VTOP(table + (vaddr & 0x7fffffffULL)); + return TRUE; + } + len = entry & 0x3ULL; +@@ -1001,12 +1001,12 @@ int s390x_vtop(ulong table, ulong vaddr, physaddr_t *phys_addr, int verbose) + if (entry & 0x400ULL) { + /* Add the 1MB page offset and return the final value. */ + table &= ~0xfffffULL; +- *phys_addr = table + (vaddr & 0xfffffULL); ++ *phys_addr = VTOP(table + (vaddr & 0xfffffULL)); + return TRUE; + } + + /* Get the page table entry */ +- entry = _kl_pg_table_deref_s390x(vaddr, entry & ~0x7ffULL, verbose); ++ entry = _kl_pg_table_deref_s390x(vaddr, PTOV(entry & ~0x7ffULL), verbose); + if (!entry) + return FALSE; + +@@ -1033,7 +1033,7 @@ s390x_vmalloc_start(void) + { + unsigned long highmem_addr,high_memory; + highmem_addr=symbol_value("high_memory"); +- readmem(highmem_addr, PHYSADDR, &high_memory,sizeof(long), ++ readmem(highmem_addr, KVADDR, &high_memory,sizeof(long), + "highmem",FAULT_ON_ERROR); + return high_memory; + } +-- +2.41.0 + diff --git a/SOURCES/0009-s390x-uncouple-physical-and-virtual-memory-spaces.patch b/SOURCES/0009-s390x-uncouple-physical-and-virtual-memory-spaces.patch new file mode 100644 index 0000000..9515898 --- /dev/null +++ b/SOURCES/0009-s390x-uncouple-physical-and-virtual-memory-spaces.patch @@ -0,0 +1,323 @@ +From d0164e7e480ad2ffd3fe73fe53c46087e5e137a6 Mon Sep 17 00:00:00 2001 +From: Alexander Gordeev +Date: Thu, 7 Dec 2023 16:54:06 +0100 +Subject: [PATCH 09/14] s390x: uncouple physical and virtual memory spaces + +Rework VTOP and PTOV macros to reflect the future +uncoupling of physical and virtual address spaces +in kernel. Existing versions are not affected. + +Signed-off-by: Alexander Gordeev +Signed-off-by: Lianbo Jiang +--- + defs.h | 20 +++++- + s390x.c | 212 +++++++++++++++++++++++++++++++++++++++++++++++++++++++- + 2 files changed, 228 insertions(+), 4 deletions(-) + +diff --git a/defs.h b/defs.h +index 5218a94fe4a4..20237b72a10b 100644 +--- a/defs.h ++++ b/defs.h +@@ -4564,9 +4564,9 @@ struct efi_memory_desc_t { + #define _64BIT_ + #define MACHINE_TYPE "S390X" + +-#define PTOV(X) ((unsigned long)(X)+(machdep->kvbase)) +-#define VTOP(X) ((unsigned long)(X)-(machdep->kvbase)) +-#define IS_VMALLOC_ADDR(X) (vt->vmalloc_start && (ulong)(X) >= vt->vmalloc_start) ++#define PTOV(X) s390x_PTOV((ulong)(X)) ++#define VTOP(X) s390x_VTOP((ulong)(X)) ++#define IS_VMALLOC_ADDR(X) s390x_IS_VMALLOC_ADDR(X) + #define PTRS_PER_PTE 512 + #define PTRS_PER_PMD 1024 + #define PTRS_PER_PGD 2048 +@@ -6827,7 +6827,21 @@ void get_s390_panicmsg(char *); + * s390x.c + */ + #ifdef S390X ++ ++struct machine_specific ++{ ++ ulong (*virt_to_phys)(ulong vaddr); ++ ulong (*phys_to_virt)(ulong paddr); ++ int (*is_vmalloc_addr)(ulong vaddr); ++ ulong __kaslr_offset_phys; ++ ulong amode31_start; ++ ulong amode31_end; ++}; ++ + void s390x_init(int); ++ulong s390x_PTOV(ulong); ++ulong s390x_VTOP(ulong); ++int s390x_IS_VMALLOC_ADDR(ulong); + void s390x_dump_machdep_table(ulong); + #define display_idt_table() \ + error(FATAL, "-d option is not applicable to S390X architecture\n") +diff --git a/s390x.c b/s390x.c +index 957b839a5fa9..794ae825906d 100644 +--- a/s390x.c ++++ b/s390x.c +@@ -47,6 +47,7 @@ + #define S390X_PSW_MASK_PSTATE 0x0001000000000000UL + + #define S390X_LC_VMCORE_INFO 0xe0c ++#define S390X_LC_OS_INFO 0xe18 + + /* + * Flags for Region and Segment table entries. +@@ -168,6 +169,19 @@ static struct line_number_hook s390x_line_number_hooks[]; + static int s390x_is_uvaddr(ulong, struct task_context *); + static int s390x_get_kvaddr_ranges(struct vaddr_range *); + static int set_s390x_max_physmem_bits(void); ++static ulong s390x_generic_VTOP(ulong vaddr); ++static ulong s390x_generic_PTOV(ulong paddr); ++static int s390x_generic_IS_VMALLOC_ADDR(ulong vaddr); ++static ulong s390x_vr_VTOP(ulong vaddr); ++static ulong s390x_vr_PTOV(ulong paddr); ++static int s390x_vr_IS_VMALLOC_ADDR(ulong vaddr); ++static int s390x_vr_is_kvaddr(ulong); ++ ++struct machine_specific s390x_machine_specific = { ++ .virt_to_phys = s390x_generic_VTOP, ++ .phys_to_virt = s390x_generic_PTOV, ++ .is_vmalloc_addr = s390x_generic_IS_VMALLOC_ADDR, ++}; + + /* + * struct lowcore name (old: "_lowcore", new: "lowcore") +@@ -546,6 +560,191 @@ static void s390x_check_kaslr(void) + free(vmcoreinfo); + } + ++#define OS_INFO_VERSION_MAJOR 1 ++#define OS_INFO_VERSION_MINOR 1 ++ ++#define OS_INFO_VMCOREINFO 0 ++#define OS_INFO_REIPL_BLOCK 1 ++#define OS_INFO_FLAGS_ENTRY 2 ++#define OS_INFO_RESERVED 3 ++#define OS_INFO_IDENTITY_BASE 4 ++#define OS_INFO_KASLR_OFFSET 5 ++#define OS_INFO_KASLR_OFF_PHYS 6 ++#define OS_INFO_VMEMMAP 7 ++#define OS_INFO_AMODE31_START 8 ++#define OS_INFO_AMODE31_END 9 ++ ++struct os_info_entry { ++ union { ++ __u64 addr; ++ __u64 val; ++ }; ++ __u64 size; ++ __u32 csum; ++} __attribute__((packed)); ++ ++struct os_info { ++ __u64 magic; ++ __u32 csum; ++ __u16 version_major; ++ __u16 version_minor; ++ __u64 crashkernel_addr; ++ __u64 crashkernel_size; ++ struct os_info_entry entry[10]; ++ __u8 reserved[3864]; ++} __attribute__((packed)); ++ ++struct vm_info { ++ __u64 __identity_base; ++ __u64 __kaslr_offset; ++ __u64 __kaslr_offset_phys; ++ __u64 amode31_start; ++ __u64 amode31_end; ++}; ++ ++static bool ++vmcoreinfo_read_u64(const char *key, __u64 *val) ++{ ++ char *string; ++ ++ string = pc->read_vmcoreinfo(key); ++ if (string) { ++ *val = strtoul(string, NULL, 16); ++ free(string); ++ return true; ++ } ++ ++ return false; ++} ++ ++static bool vmcoreinfo_read_vm_info(struct vm_info *_vm_info) ++{ ++ struct vm_info vm_info; ++ ++ if (!vmcoreinfo_read_u64("IDENTITYBASE", &vm_info.__identity_base) || ++ !vmcoreinfo_read_u64("KERNELOFFSET", &vm_info.__kaslr_offset) || ++ !vmcoreinfo_read_u64("KERNELOFFPHYS", &vm_info.__kaslr_offset_phys) || ++ !vmcoreinfo_read_u64("SAMODE31", &vm_info.amode31_start) || ++ !vmcoreinfo_read_u64("EAMODE31", &vm_info.amode31_end)) ++ return false; ++ ++ *_vm_info = vm_info; ++ ++ return true; ++} ++ ++static bool os_info_read_vm_info(struct vm_info *vm_info) ++{ ++ struct os_info os_info; ++ ulong addr; ++ ++ if (!readmem(S390X_LC_OS_INFO, PHYSADDR, &addr, ++ sizeof(addr), "s390x os_info ptr", ++ QUIET|RETURN_ON_ERROR)) ++ return false; ++ ++ if (addr == 0) ++ return true; ++ ++ if (!readmem(addr, PHYSADDR, &os_info, ++ offsetof(struct os_info, reserved), "s390x os_info header", ++ QUIET|RETURN_ON_ERROR)) ++ return false; ++ ++ vm_info->__identity_base = os_info.entry[OS_INFO_IDENTITY_BASE].val; ++ vm_info->__kaslr_offset = os_info.entry[OS_INFO_KASLR_OFFSET].val; ++ vm_info->__kaslr_offset_phys = os_info.entry[OS_INFO_KASLR_OFF_PHYS].val; ++ vm_info->amode31_start = os_info.entry[OS_INFO_AMODE31_START].val; ++ vm_info->amode31_end = os_info.entry[OS_INFO_AMODE31_END].val; ++ ++ return true; ++} ++ ++static bool vm_info_empty(struct vm_info *vm_info) ++{ ++ return !vm_info->__kaslr_offset; ++} ++ ++static bool s390x_init_vm(void) ++{ ++ struct vm_info vm_info; ++ ++ if (pc->flags & PROC_KCORE) { ++ if (!vmcoreinfo_read_vm_info(&vm_info)) ++ return true; ++ } else { ++ if (!os_info_read_vm_info(&vm_info)) ++ return false; ++ } ++ if (vm_info_empty(&vm_info)) ++ return true; ++ ++ machdep->identity_map_base = vm_info.__identity_base; ++ machdep->kvbase = vm_info.__kaslr_offset; ++ machdep->machspec->__kaslr_offset_phys = vm_info.__kaslr_offset_phys; ++ machdep->machspec->amode31_start = vm_info.amode31_start; ++ machdep->machspec->amode31_end = vm_info.amode31_end; ++ ++ machdep->is_kvaddr = s390x_vr_is_kvaddr; ++ machdep->machspec->virt_to_phys = s390x_vr_VTOP; ++ machdep->machspec->phys_to_virt = s390x_vr_PTOV; ++ machdep->machspec->is_vmalloc_addr = s390x_vr_IS_VMALLOC_ADDR; ++ ++ return true; ++} ++ ++static ulong s390x_generic_VTOP(ulong vaddr) ++{ ++ return vaddr - machdep->kvbase; ++} ++ ++static ulong s390x_generic_PTOV(ulong paddr) ++{ ++ return paddr + machdep->kvbase; ++} ++ ++static int s390x_generic_IS_VMALLOC_ADDR(ulong vaddr) ++{ ++ return vt->vmalloc_start && vaddr >= vt->vmalloc_start; ++} ++ ++static ulong s390x_vr_VTOP(ulong vaddr) ++{ ++ if (vaddr < LOWCORE_SIZE) ++ return vaddr; ++ if ((vaddr < machdep->machspec->amode31_end) && ++ (vaddr >= machdep->machspec->amode31_start)) ++ return vaddr; ++ if (vaddr < machdep->kvbase) ++ return vaddr - machdep->identity_map_base; ++ return vaddr - machdep->kvbase + machdep->machspec->__kaslr_offset_phys; ++} ++ ++static ulong s390x_vr_PTOV(ulong paddr) ++{ ++ return paddr + machdep->identity_map_base; ++} ++ ++static int s390x_vr_IS_VMALLOC_ADDR(ulong vaddr) ++{ ++ return (vaddr >= vt->vmalloc_start && vaddr < machdep->kvbase); ++} ++ ++ulong s390x_VTOP(ulong vaddr) ++{ ++ return machdep->machspec->virt_to_phys(vaddr); ++} ++ ++ulong s390x_PTOV(ulong paddr) ++{ ++ return machdep->machspec->phys_to_virt(paddr); ++} ++ ++int s390x_IS_VMALLOC_ADDR(ulong vaddr) ++{ ++ return machdep->machspec->is_vmalloc_addr(vaddr); ++} ++ + /* + * Do all necessary machine-specific setup here. This is called several + * times during initialization. +@@ -560,6 +759,7 @@ s390x_init(int when) + machdep->process_elf_notes = s390x_process_elf_notes; + break; + case PRE_SYMTAB: ++ machdep->machspec = &s390x_machine_specific; + machdep->verify_symbol = s390x_verify_symbol; + if (pc->flags & KERNEL_DEBUG_QUERY) + return; +@@ -587,6 +787,8 @@ s390x_init(int when) + machdep->kvbase = 0; + machdep->identity_map_base = 0; + machdep->is_kvaddr = generic_is_kvaddr; ++ if (!s390x_init_vm()) ++ error(FATAL, "cannot initialize VM parameters."); + machdep->is_uvaddr = s390x_is_uvaddr; + machdep->eframe_search = s390x_eframe_search; + machdep->back_trace = s390x_back_trace_cmd; +@@ -681,7 +883,9 @@ s390x_dump_machdep_table(ulong arg) + fprintf(fp, " dis_filter: s390x_dis_filter()\n"); + fprintf(fp, " cmd_mach: s390x_cmd_mach()\n"); + fprintf(fp, " get_smp_cpus: s390x_get_smp_cpus()\n"); +- fprintf(fp, " is_kvaddr: generic_is_kvaddr()\n"); ++ fprintf(fp, " is_kvaddr: %s()\n", machdep->is_kvaddr == s390x_vr_is_kvaddr ? ++ "s390x_vr_is_kvaddr" : ++ "generic_is_kvaddr"); + fprintf(fp, " is_uvaddr: s390x_is_uvaddr()\n"); + fprintf(fp, " verify_paddr: generic_verify_paddr()\n"); + fprintf(fp, " get_kvaddr_ranges: s390x_get_kvaddr_ranges()\n"); +@@ -702,6 +906,12 @@ s390x_dump_machdep_table(ulong arg) + fprintf(fp, " machspec: %lx\n", (ulong)machdep->machspec); + } + ++static int ++s390x_vr_is_kvaddr(ulong vaddr) ++{ ++ return (vaddr < LOWCORE_SIZE) || (vaddr >= machdep->identity_map_base); ++} ++ + /* + * Check if address is in context's address space + */ +-- +2.41.0 + diff --git a/SOURCES/0010-RISCV64-Dump-NT_PRSTATUS-in-help-n.patch b/SOURCES/0010-RISCV64-Dump-NT_PRSTATUS-in-help-n.patch new file mode 100644 index 0000000..f4a1bf9 --- /dev/null +++ b/SOURCES/0010-RISCV64-Dump-NT_PRSTATUS-in-help-n.patch @@ -0,0 +1,160 @@ +From 5187a0320cc54a9cb8b326cf012e69795950a716 Mon Sep 17 00:00:00 2001 +From: Song Shuai +Date: Tue, 12 Dec 2023 18:20:50 +0800 +Subject: [PATCH 10/14] RISCV64: Dump NT_PRSTATUS in 'help -n' + +With the patch we can get full dump of "struct elf_prstatus" in 'help -n': +``` +crash> help -n + +Elf64_Nhdr: + n_namesz: 5 ("CORE") + n_descsz: 376 + n_type: 1 (NT_PRSTATUS) + si.signo: 0 si.code: 0 si.errno: 0 + cursig: 0 sigpend: 0 sighold: 0 + pid: 1 ppid: 0 pgrp: 0 sid:0 + utime: 0.000000 stime: 0.000000 + cutime: 0.000000 cstime: 0.000000 + epc: ffffffff8000a1dc ra: ffffffff800af958 sp: ff6000001fc501c0 + gp: ffffffff81515d38 tp: ff600000000d8000 t0: 6666666666663c5b + t1: ff600000000d88c8 t2: 666666666666663c s0: ff6000001fc50320 + s1: ffffffff815170d8 a0: ff6000001fc501c8 a1: c0000000ffffefff + a2: 0000000000000000 a3: 0000000000000001 a4: 0000000000000000 + a5: ff60000001782c00 a6: 000000000130e0f0 a7: 0000000000000000 + s2: ffffffff81517820 s3: ff6000001fc501c8 s4: 000000000000000f + s5: 0000000000000000 s6: ff20000000013e60 s7: 0000000000000000 + s8: ff60000000861000 s9: 00007fffc3641694 s10: 00007fffc3641690 + s11: 00005555796ed240 t3: 0000000000010297 t4: ffffffff80c17810 + t5: ffffffff8195e7b8 t6: ff6000001fc50048 + 0000000000000000 0000000000000000 + 0000000000000000 0000000000000000 + 0000000000000001 0000000000000000 + 0000000000000000 0000000000000000 + 0000000000000000 0000000000000000 + 0000000000000000 0000000000000000 + 0000000000000000 0000000000000000 + ffffffff8000a1dc ffffffff800af958 + ff6000001fc501c0 ffffffff81515d38 + ff600000000d8000 6666666666663c5b + +``` + +Signed-off-by: Song Shuai +Signed-off-by: Lianbo Jiang +--- + netdump.c | 84 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ + 1 file changed, 84 insertions(+) + +diff --git a/netdump.c b/netdump.c +index 390786364959..32586b6809d3 100644 +--- a/netdump.c ++++ b/netdump.c +@@ -2578,6 +2578,8 @@ dump_Elf64_Nhdr(Elf64_Off offset, int store) + display_ELF_note(EM_PPC64, PRSTATUS_NOTE, note, nd->ofp); + if (machine_type("ARM64") && (note->n_type == NT_PRSTATUS)) + display_ELF_note(EM_AARCH64, PRSTATUS_NOTE, note, nd->ofp); ++ if (machine_type("RISCV64") && (note->n_type == NT_PRSTATUS)) ++ display_ELF_note(EM_RISCV, PRSTATUS_NOTE, note, nd->ofp); + } + for (i = lf = 0; i < note->n_descsz/sizeof(ulonglong); i++) { + if (((i%2)==0)) { +@@ -3399,6 +3401,80 @@ display_prstatus_arm64(void *note_ptr, FILE *ofp) + space(sp), pr->pr_reg[33], pr->pr_fpvalid); + } + ++struct riscv64_elf_siginfo { ++ int si_signo; ++ int si_code; ++ int si_errno; ++}; ++ ++struct riscv64_elf_prstatus { ++ struct riscv64_elf_siginfo pr_info; ++ short pr_cursig; ++ unsigned long pr_sigpend; ++ unsigned long pr_sighold; ++ pid_t pr_pid; ++ pid_t pr_ppid; ++ pid_t pr_pgrp; ++ pid_t pr_sid; ++ struct timeval pr_utime; ++ struct timeval pr_stime; ++ struct timeval pr_cutime; ++ struct timeval pr_cstime; ++/* elf_gregset_t pr_reg; => typedef struct user_regs_struct elf_gregset_t; */ ++ unsigned long pr_reg[32]; ++ int pr_fpvalid; ++}; ++ ++static void ++display_prstatus_riscv64(void *note_ptr, FILE *ofp) ++{ ++ struct riscv64_elf_prstatus *pr; ++ Elf64_Nhdr *note; ++ int sp; ++ ++ note = (Elf64_Nhdr *)note_ptr; ++ pr = (struct riscv64_elf_prstatus *)( ++ (char *)note + sizeof(Elf64_Nhdr) + note->n_namesz); ++ pr = (struct riscv64_elf_prstatus *)roundup((ulong)pr, 4); ++ sp = nd->num_prstatus_notes ? 25 : 22; ++ ++ fprintf(ofp, ++ "%ssi.signo: %d si.code: %d si.errno: %d\n" ++ "%scursig: %d sigpend: %lx sighold: %lx\n" ++ "%spid: %d ppid: %d pgrp: %d sid:%d\n" ++ "%sutime: %01lld.%06d stime: %01lld.%06d\n" ++ "%scutime: %01lld.%06d cstime: %01lld.%06d\n", ++ space(sp), pr->pr_info.si_signo, pr->pr_info.si_code, pr->pr_info.si_errno, ++ space(sp), pr->pr_cursig, pr->pr_sigpend, pr->pr_sighold, ++ space(sp), pr->pr_pid, pr->pr_ppid, pr->pr_pgrp, pr->pr_sid, ++ space(sp), (long long)pr->pr_utime.tv_sec, (int)pr->pr_utime.tv_usec, ++ (long long)pr->pr_stime.tv_sec, (int)pr->pr_stime.tv_usec, ++ space(sp), (long long)pr->pr_cutime.tv_sec, (int)pr->pr_cutime.tv_usec, ++ (long long)pr->pr_cstime.tv_sec, (int)pr->pr_cstime.tv_usec); ++ fprintf(ofp, ++ "%sepc: %016lx ra: %016lx sp: %016lx\n" ++ "%s gp: %016lx tp: %016lx t0: %016lx\n" ++ "%s t1: %016lx t2: %016lx s0: %016lx\n" ++ "%s s1: %016lx a0: %016lx a1: %016lx\n" ++ "%s a2: %016lx a3: %016lx a4: %016lx\n" ++ "%s a5: %016lx a6: %016lx a7: %016lx\n" ++ "%s s2: %016lx s3: %016lx s4: %016lx\n" ++ "%s s5: %016lx s6: %016lx s7: %016lx\n" ++ "%s s8: %016lx s9: %016lx s10: %016lx\n" ++ "%ss11: %016lx t3: %016lx t4: %016lx\n" ++ "%s t5: %016lx t6: %016lx\n", ++ space(sp), pr->pr_reg[0], pr->pr_reg[1], pr->pr_reg[2], ++ space(sp), pr->pr_reg[3], pr->pr_reg[4], pr->pr_reg[5], ++ space(sp), pr->pr_reg[6], pr->pr_reg[7], pr->pr_reg[8], ++ space(sp), pr->pr_reg[9], pr->pr_reg[10], pr->pr_reg[11], ++ space(sp), pr->pr_reg[12], pr->pr_reg[13], pr->pr_reg[14], ++ space(sp), pr->pr_reg[15], pr->pr_reg[16], pr->pr_reg[17], ++ space(sp), pr->pr_reg[18], pr->pr_reg[19], pr->pr_reg[20], ++ space(sp), pr->pr_reg[21], pr->pr_reg[22], pr->pr_reg[23], ++ space(sp), pr->pr_reg[24], pr->pr_reg[25], pr->pr_reg[26], ++ space(sp), pr->pr_reg[27], pr->pr_reg[28], pr->pr_reg[29], ++ space(sp), pr->pr_reg[30], pr->pr_reg[31]); ++} + + void + display_ELF_note(int machine, int type, void *note, FILE *ofp) +@@ -3449,6 +3525,14 @@ display_ELF_note(int machine, int type, void *note, FILE *ofp) + break; + } + break; ++ case EM_RISCV: ++ switch (type) ++ { ++ case PRSTATUS_NOTE: ++ display_prstatus_riscv64(note, ofp); ++ break; ++ } ++ break; + + default: + return; +-- +2.41.0 + diff --git a/SOURCES/0011-RISCV64-Fix-bt-output-when-no-ra-on-the-stack-top.patch b/SOURCES/0011-RISCV64-Fix-bt-output-when-no-ra-on-the-stack-top.patch new file mode 100644 index 0000000..3e85ee6 --- /dev/null +++ b/SOURCES/0011-RISCV64-Fix-bt-output-when-no-ra-on-the-stack-top.patch @@ -0,0 +1,87 @@ +From 9b69093e623f1d54c373b1e091900d40576c059b Mon Sep 17 00:00:00 2001 +From: Song Shuai +Date: Tue, 12 Dec 2023 18:20:51 +0800 +Subject: [PATCH 11/14] RISCV64: Fix 'bt' output when no ra on the stack top + +Same as the Linux commit f766f77a74f5 ("riscv/stacktrace: Fix +stack output without ra on the stack top"). + +When a function doesn't have a callee, then it will not +push ra into the stack, such as lkdtm functions, so +correct the FP of the second frame and use pt_regs to get +the right PC of the second frame. + +Before this patch, the `bt -f` outputs only the first frame with +the wrong PC and FP of next frame: +``` +crash> bt -f +PID: 1 TASK: ff600000000e0000 CPU: 1 COMMAND: "sh" + #0 [ff20000000013cf0] lkdtm_EXCEPTION at ffffffff805303c0 + [PC: ffffffff805303c0 RA: ff20000000013d10 SP: ff20000000013cf0 SIZE: 16] <- wrong next PC + ff20000000013cf0: 0000000000000001 ff20000000013d10 <- next FP + ff20000000013d00: ff20000000013d40 +crash> +``` +After this patch, the `bt` outputs the full frames: +``` +crash> bt +PID: 1 TASK: ff600000000e0000 CPU: 1 COMMAND: "sh" + #0 [ff20000000013cf0] lkdtm_EXCEPTION at ffffffff805303c0 + #1 [ff20000000013d00] lkdtm_do_action at ffffffff8052fe36 + #2 [ff20000000013d10] direct_entry at ffffffff80530018 + #3 [ff20000000013d40] full_proxy_write at ffffffff80305044 + #4 [ff20000000013d80] vfs_write at ffffffff801b68b4 + #5 [ff20000000013e30] ksys_write at ffffffff801b6c4a + #6 [ff20000000013e80] __riscv_sys_write at ffffffff801b6cc4 + #7 [ff20000000013e90] do_trap_ecall_u at ffffffff80836798 +crash> +``` + +Acked-by: Kazuhito Hagio +Signed-off-by: Song Shuai +Signed-off-by: Lianbo Jiang +--- + riscv64.c | 15 +++++++++++++-- + 1 file changed, 13 insertions(+), 2 deletions(-) + +diff --git a/riscv64.c b/riscv64.c +index 0aaa14b2671e..872be594d72b 100644 +--- a/riscv64.c ++++ b/riscv64.c +@@ -747,11 +747,14 @@ riscv64_back_trace_cmd(struct bt_info *bt) + { + struct riscv64_unwind_frame current, previous; + struct stackframe curr_frame; ++ struct riscv64_register * regs; + int level = 0; + + if (bt->flags & BT_REGS_NOT_FOUND) + return; + ++ regs = (struct riscv64_register *) bt->machdep; ++ + current.pc = bt->instptr; + current.sp = bt->stkptr; + current.fp = bt->frameptr; +@@ -788,8 +791,16 @@ riscv64_back_trace_cmd(struct bt_info *bt) + sizeof(curr_frame), "get stack frame", RETURN_ON_ERROR)) + return; + +- previous.pc = curr_frame.ra; +- previous.fp = curr_frame.fp; ++ /* correct PC and FP of the second frame when the first frame has no callee */ ++ ++ if (regs && (regs->regs[RISCV64_REGS_EPC] == current.pc) && curr_frame.fp & 0x7){ ++ previous.pc = regs->regs[RISCV64_REGS_RA]; ++ previous.fp = curr_frame.ra; ++ } else { ++ previous.pc = curr_frame.ra; ++ previous.fp = curr_frame.fp; ++ } ++ + previous.sp = current.fp; + + riscv64_dump_backtrace_entry(bt, symbol, ¤t, &previous, level++); +-- +2.41.0 + diff --git a/SOURCES/0012-arm64-rewrite-the-arm64_get_vmcoreinfo_ul-to-arm64_g.patch b/SOURCES/0012-arm64-rewrite-the-arm64_get_vmcoreinfo_ul-to-arm64_g.patch new file mode 100644 index 0000000..ca1a749 --- /dev/null +++ b/SOURCES/0012-arm64-rewrite-the-arm64_get_vmcoreinfo_ul-to-arm64_g.patch @@ -0,0 +1,265 @@ +From 19d3c56c9fca9dea49dced0414becc6d1b12e9fc Mon Sep 17 00:00:00 2001 +From: Huang Shijie +Date: Thu, 14 Dec 2023 15:15:20 +0800 +Subject: [PATCH 12/14] arm64: rewrite the arm64_get_vmcoreinfo_ul to + arm64_get_vmcoreinfo + +Rewrite the arm64_get_vmcoreinfo_ul to arm64_get_vmcoreinfo, +add a new parameter "base" for it. + +Also use it to simplify the arm64 code. + +Signed-off-by: Huang Shijie +Signed-off-by: Lianbo Jiang +--- + arm64.c | 99 +++++++++++++++++++++++---------------------------------- + 1 file changed, 39 insertions(+), 60 deletions(-) + +diff --git a/arm64.c b/arm64.c +index 2b6b0e588d4e..57965c6cb3c8 100644 +--- a/arm64.c ++++ b/arm64.c +@@ -92,6 +92,7 @@ static void arm64_get_crash_notes(void); + static void arm64_calc_VA_BITS(void); + static int arm64_is_uvaddr(ulong, struct task_context *); + static void arm64_calc_KERNELPACMASK(void); ++static int arm64_get_vmcoreinfo(unsigned long *vaddr, const char *label, int base); + + struct kernel_range { + unsigned long modules_vaddr, modules_end; +@@ -124,7 +125,6 @@ void + arm64_init(int when) + { + ulong value; +- char *string; + struct machine_specific *ms; + + #if defined(__x86_64__) +@@ -160,11 +160,8 @@ arm64_init(int when) + if (!ms->kimage_voffset && STREQ(pc->live_memsrc, "/dev/crash")) + ioctl(pc->mfd, DEV_CRASH_ARCH_DATA, &ms->kimage_voffset); + +- if (!ms->kimage_voffset && +- (string = pc->read_vmcoreinfo("NUMBER(kimage_voffset)"))) { +- ms->kimage_voffset = htol(string, QUIET, NULL); +- free(string); +- } ++ if (!ms->kimage_voffset) ++ arm64_get_vmcoreinfo(&ms->kimage_voffset, "NUMBER(kimage_voffset)", NUM_HEX); + + if (ms->kimage_voffset || + (ACTIVE() && (symbol_value_from_proc_kallsyms("kimage_voffset") != BADVAL))) { +@@ -185,11 +182,8 @@ arm64_init(int when) + if (kernel_symbol_exists("kimage_voffset")) + machdep->flags |= NEW_VMEMMAP; + +- if (!machdep->pagesize && +- (string = pc->read_vmcoreinfo("PAGESIZE"))) { +- machdep->pagesize = atoi(string); +- free(string); +- } ++ if (!machdep->pagesize && arm64_get_vmcoreinfo(&value, "PAGESIZE", NUM_DEC)) ++ machdep->pagesize = (unsigned int)value; + + if (!machdep->pagesize) { + /* +@@ -443,9 +437,8 @@ arm64_init(int when) + arm64_get_section_size_bits(); + + if (!machdep->max_physmem_bits) { +- if ((string = pc->read_vmcoreinfo("NUMBER(MAX_PHYSMEM_BITS)"))) { +- machdep->max_physmem_bits = atol(string); +- free(string); ++ if (arm64_get_vmcoreinfo(&machdep->max_physmem_bits, "NUMBER(MAX_PHYSMEM_BITS)", NUM_DEC)) { ++ /* nothing */ + } else if (machdep->machspec->VA_BITS == 52) /* guess */ + machdep->max_physmem_bits = _MAX_PHYSMEM_BITS_52; + else if (THIS_KERNEL_VERSION >= LINUX(3,17,0)) +@@ -573,16 +566,28 @@ static int arm64_get_struct_page_max_shift(struct machine_specific *ms) + } + + /* Return TRUE if we succeed, return FALSE on failure. */ +-static int arm64_get_vmcoreinfo_ul(unsigned long *vaddr, const char* label) ++static int ++arm64_get_vmcoreinfo(unsigned long *vaddr, const char *label, int base) + { ++ int err = 0; + char *string = pc->read_vmcoreinfo(label); + + if (!string) + return FALSE; + +- *vaddr = strtoul(string, NULL, 0); ++ switch (base) { ++ case NUM_HEX: ++ *vaddr = strtoul(string, NULL, 16); ++ break; ++ case NUM_DEC: ++ *vaddr = strtoul(string, NULL, 10); ++ break; ++ default: ++ err++; ++ error(INFO, "Unknown type:%#x, (NUM_HEX|NUM_DEC)\n", base); ++ } + free(string); +- return TRUE; ++ return err ? FALSE: TRUE; + } + + /* +@@ -594,21 +599,21 @@ static struct kernel_range *arm64_get_range_v5_18(struct machine_specific *ms) + struct kernel_range *r = &tmp_range; + + /* Get the MODULES_VADDR ~ MODULES_END */ +- if (!arm64_get_vmcoreinfo_ul(&r->modules_vaddr, "NUMBER(MODULES_VADDR)")) ++ if (!arm64_get_vmcoreinfo(&r->modules_vaddr, "NUMBER(MODULES_VADDR)", NUM_HEX)) + return NULL; +- if (!arm64_get_vmcoreinfo_ul(&r->modules_end, "NUMBER(MODULES_END)")) ++ if (!arm64_get_vmcoreinfo(&r->modules_end, "NUMBER(MODULES_END)", NUM_HEX)) + return NULL; + + /* Get the VMEMMAP_START ~ VMEMMAP_END */ +- if (!arm64_get_vmcoreinfo_ul(&r->vmemmap_vaddr, "NUMBER(VMEMMAP_START)")) ++ if (!arm64_get_vmcoreinfo(&r->vmemmap_vaddr, "NUMBER(VMEMMAP_START)", NUM_HEX)) + return NULL; +- if (!arm64_get_vmcoreinfo_ul(&r->vmemmap_end, "NUMBER(VMEMMAP_END)")) ++ if (!arm64_get_vmcoreinfo(&r->vmemmap_end, "NUMBER(VMEMMAP_END)", NUM_HEX)) + return NULL; + + /* Get the VMALLOC_START ~ VMALLOC_END */ +- if (!arm64_get_vmcoreinfo_ul(&r->vmalloc_start_addr, "NUMBER(VMALLOC_START)")) ++ if (!arm64_get_vmcoreinfo(&r->vmalloc_start_addr, "NUMBER(VMALLOC_START)", NUM_HEX)) + return NULL; +- if (!arm64_get_vmcoreinfo_ul(&r->vmalloc_end, "NUMBER(VMALLOC_END)")) ++ if (!arm64_get_vmcoreinfo(&r->vmalloc_end, "NUMBER(VMALLOC_END)", NUM_HEX)) + return NULL; + + return r; +@@ -888,12 +893,7 @@ range_failed: + /* Get the size of struct page {} */ + static void arm64_get_struct_page_size(struct machine_specific *ms) + { +- char *string; +- +- string = pc->read_vmcoreinfo("SIZE(page)"); +- if (string) +- ms->struct_page_size = atol(string); +- free(string); ++ arm64_get_vmcoreinfo(&ms->struct_page_size, "SIZE(page)", NUM_DEC); + } + + /* +@@ -1469,16 +1469,12 @@ arm64_calc_phys_offset(void) + physaddr_t paddr; + ulong vaddr; + struct syment *sp; +- char *string; + + if ((machdep->flags & NEW_VMEMMAP) && + ms->kimage_voffset && (sp = kernel_symbol_search("memstart_addr"))) { + if (pc->flags & PROC_KCORE) { +- if ((string = pc->read_vmcoreinfo("NUMBER(PHYS_OFFSET)"))) { +- ms->phys_offset = htol(string, QUIET, NULL); +- free(string); ++ if (arm64_get_vmcoreinfo(&ms->phys_offset, "NUMBER(PHYS_OFFSET)", NUM_HEX)) + return; +- } + vaddr = symbol_value_from_proc_kallsyms("memstart_addr"); + if (vaddr == BADVAL) + vaddr = sp->value; +@@ -1560,9 +1556,8 @@ arm64_get_section_size_bits(void) + } else + machdep->section_size_bits = _SECTION_SIZE_BITS; + +- if ((string = pc->read_vmcoreinfo("NUMBER(SECTION_SIZE_BITS)"))) { +- machdep->section_size_bits = atol(string); +- free(string); ++ if (arm64_get_vmcoreinfo(&machdep->section_size_bits, "NUMBER(SECTION_SIZE_BITS)", NUM_DEC)) { ++ /* nothing */ + } else if (kt->ikconfig_flags & IKCONFIG_AVAIL) { + if ((ret = get_kernel_config("CONFIG_MEMORY_HOTPLUG", NULL)) == IKCONFIG_Y) { + if ((ret = get_kernel_config("CONFIG_HOTPLUG_SIZE_BITS", &string)) == IKCONFIG_STR) +@@ -1581,15 +1576,11 @@ arm64_get_section_size_bits(void) + static int + arm64_kdump_phys_base(ulong *phys_offset) + { +- char *string; + struct syment *sp; + physaddr_t paddr; + +- if ((string = pc->read_vmcoreinfo("NUMBER(PHYS_OFFSET)"))) { +- *phys_offset = htol(string, QUIET, NULL); +- free(string); ++ if (arm64_get_vmcoreinfo(phys_offset, "NUMBER(PHYS_OFFSET)", NUM_HEX)) + return TRUE; +- } + + if ((machdep->flags & NEW_VMEMMAP) && + machdep->machspec->kimage_voffset && +@@ -4592,10 +4583,9 @@ static int + arm64_set_va_bits_by_tcr(void) + { + ulong value; +- char *string; + +- if ((string = pc->read_vmcoreinfo("NUMBER(TCR_EL1_T1SZ)")) || +- (string = pc->read_vmcoreinfo("NUMBER(tcr_el1_t1sz)"))) { ++ if (arm64_get_vmcoreinfo(&value, "NUMBER(TCR_EL1_T1SZ)", NUM_HEX) || ++ arm64_get_vmcoreinfo(&value, "NUMBER(tcr_el1_t1sz)", NUM_HEX)) { + /* See ARMv8 ARM for the description of + * TCR_EL1.T1SZ and how it can be used + * to calculate the vabits_actual +@@ -4604,10 +4594,9 @@ arm64_set_va_bits_by_tcr(void) + * Basically: + * vabits_actual = 64 - T1SZ; + */ +- value = 64 - strtoll(string, NULL, 0); ++ value = 64 - value; + if (CRASHDEBUG(1)) + fprintf(fp, "vmcoreinfo : vabits_actual: %ld\n", value); +- free(string); + machdep->machspec->VA_BITS_ACTUAL = value; + machdep->machspec->VA_BITS = value; + machdep->machspec->VA_START = _VA_START(machdep->machspec->VA_BITS_ACTUAL); +@@ -4623,13 +4612,8 @@ arm64_calc_VA_BITS(void) + int bitval; + struct syment *sp; + ulong vabits_actual, value; +- char *string; + +- if ((string = pc->read_vmcoreinfo("NUMBER(VA_BITS)"))) { +- value = atol(string); +- free(string); +- machdep->machspec->CONFIG_ARM64_VA_BITS = value; +- } ++ arm64_get_vmcoreinfo(&machdep->machspec->CONFIG_ARM64_VA_BITS, "NUMBER(VA_BITS)", NUM_DEC); + + if (kernel_symbol_exists("vabits_actual")) { + if (pc->flags & PROC_KCORE) { +@@ -4754,9 +4738,7 @@ arm64_calc_virtual_memory_ranges(void) + ulong PUD_SIZE = UNINITIALIZED; + + if (!machdep->machspec->CONFIG_ARM64_VA_BITS) { +- if ((string = pc->read_vmcoreinfo("NUMBER(VA_BITS)"))) { +- value = atol(string); +- free(string); ++ if (arm64_get_vmcoreinfo(&value, "NUMBER(VA_BITS)", NUM_DEC)) { + machdep->machspec->CONFIG_ARM64_VA_BITS = value; + } else if (kt->ikconfig_flags & IKCONFIG_AVAIL) { + if ((ret = get_kernel_config("CONFIG_ARM64_VA_BITS", +@@ -4852,11 +4834,8 @@ arm64_swp_offset(ulong pte) + static void arm64_calc_KERNELPACMASK(void) + { + ulong value; +- char *string; + +- if ((string = pc->read_vmcoreinfo("NUMBER(KERNELPACMASK)"))) { +- value = htol(string, QUIET, NULL); +- free(string); ++ if (arm64_get_vmcoreinfo(&value, "NUMBER(KERNELPACMASK)", NUM_HEX)) { + machdep->machspec->CONFIG_ARM64_KERNELPACMASK = value; + if (CRASHDEBUG(1)) + fprintf(fp, "CONFIG_ARM64_KERNELPACMASK: %lx\n", value); +-- +2.41.0 + diff --git a/SOURCES/0013-help.c-Remove-kmem-l-help-messages.patch b/SOURCES/0013-help.c-Remove-kmem-l-help-messages.patch new file mode 100644 index 0000000..553471c --- /dev/null +++ b/SOURCES/0013-help.c-Remove-kmem-l-help-messages.patch @@ -0,0 +1,36 @@ +From 38435c3acec075b076353ca28f557a0dfe1341c3 Mon Sep 17 00:00:00 2001 +From: Li Zhijian +Date: Fri, 15 Dec 2023 10:44:21 +0800 +Subject: [PATCH 13/14] help.c: Remove "kmem -l" help messages + +"kmem -l" option has existed when crash git project initialization, but +its help message was not accurate (extra arguments a|i|ic|id was missing). + +In addition, those symbols required by the -l option were for very old +kernels, at least 2.6 kernels don't contain them. Also, this option has +not been fixed for a long time. + +Instead of document this option, hide it from help messages. + +Signed-off-by: Li Zhijian +Signed-off-by: Lianbo Jiang +--- + help.c | 2 -- + 1 file changed, 2 deletions(-) + +diff --git a/help.c b/help.c +index cc7ab20e343e..d80e843703c1 100644 +--- a/help.c ++++ b/help.c +@@ -6888,8 +6888,6 @@ char *help_kmem[] = { + " members of the associated page struct are displayed.", + " address when used with -c, the address must be a page pointer address;", + " the page_hash_table entry containing the page is displayed.", +-" address when used with -l, the address must be a page pointer address;", +-" the page address is displayed if it is contained with the list.", + " address when used with -v, the address can be a mapped kernel virtual", + " address or physical address; the mapped region containing the", + " address is displayed.\n", +-- +2.41.0 + diff --git a/SOURCES/0014-x86_64-check-bt-bptr-before-calculate-framesize.patch b/SOURCES/0014-x86_64-check-bt-bptr-before-calculate-framesize.patch new file mode 100644 index 0000000..19d9666 --- /dev/null +++ b/SOURCES/0014-x86_64-check-bt-bptr-before-calculate-framesize.patch @@ -0,0 +1,65 @@ +From 53d2577cef98b76b122aade94349637a11e06138 Mon Sep 17 00:00:00 2001 +From: Tao Liu +Date: Tue, 26 Dec 2023 09:19:28 +0800 +Subject: [PATCH 14/14] x86_64: check bt->bptr before calculate framesize + +Previously the value of bt->bptr is not checked, which may led to a +wrong prev_sp and framesize. As a result, bt->stackbuf[] will be +accessed out of range, and segfault. + +Before: + crash> set debug 1 + crash> bt + ...snip... + --- --- + #8 [ffffffff9a603e10] __switch_to_asm at ffffffff99800214 + rsp: ffffffff9a603e10 textaddr: ffffffff99800214 -> spo: 0 bpo: 0 spr: 0 bpr: 0 type: 0 end: 0 + #9 [ffffffff9a603e40] __schedule at ffffffff9960dfb1 + rsp: ffffffff9a603e40 textaddr: ffffffff9960dfb1 -> spo: 16 bpo: -16 spr: 4 bpr: 1 type: 0 end: 0 + rsp: ffffffff9a603e40 rbp: ffffb9ca076e7ca8 prev_sp: ffffb9ca076e7cb8 framesize: 1829650024 + Segmentation fault (core dumped) + + (gdb) p/x bt->stackbase + $1 = 0xffffffff9a600000 + (gdb) p/x bt->stacktop + $2 = 0xffffffff9a604000 + +After: + crash> set debug 1 + crash> bt + ...snip... + --- --- + #8 [ffffffff9a603e10] __switch_to_asm at ffffffff99800214 + rsp: ffffffff9a603e10 textaddr: ffffffff99800214 -> spo: 0 bpo: 0 spr: 0 bpr: 0 type: 0 end: 0 + #9 [ffffffff9a603e40] __schedule at ffffffff9960dfb1 + rsp: ffffffff9a603e40 textaddr: ffffffff9960dfb1 -> spo: 16 bpo: -16 spr: 4 bpr: 1 type: 0 end: 0 + #10 [ffffffff9a603e98] schedule_idle at ffffffff9960e87c + rsp: ffffffff9a603e98 textaddr: ffffffff9960e87c -> spo: 8 bpo: 0 spr: 5 bpr: 0 type: 0 end: 0 + rsp: ffffffff9a603e98 prev_sp: ffffffff9a603ea8 framesize: 0 + ...snip... + +Check bt->bptr value before calculate framesize. Only bt->bptr within +the range of bt->stackbase and bt->stacktop will be regarded as valid. + +Signed-off-by: Tao Liu +Signed-off-by: Lianbo Jiang +--- + x86_64.c | 2 +- + 1 file changed, 1 insertion(+), 1 deletion(-) + +diff --git a/x86_64.c b/x86_64.c +index 42ade4817ad9..f59991f8c4c5 100644 +--- a/x86_64.c ++++ b/x86_64.c +@@ -8649,7 +8649,7 @@ x86_64_get_framesize(struct bt_info *bt, ulong textaddr, ulong rsp, char *stack_ + if (CRASHDEBUG(1)) + fprintf(fp, "rsp: %lx prev_sp: %lx framesize: %d\n", + rsp, prev_sp, framesize); +- } else if ((korc->sp_reg == ORC_REG_BP) && bt->bptr) { ++ } else if ((korc->sp_reg == ORC_REG_BP) && bt->bptr && INSTACK(bt->bptr, bt)) { + prev_sp = bt->bptr + korc->sp_offset; + framesize = (prev_sp - (rsp + 8) - 8); + if (CRASHDEBUG(1)) +-- +2.41.0 + diff --git a/SOURCES/crash-8.0.3_build.patch b/SOURCES/crash-8.0.4_build.patch similarity index 91% rename from SOURCES/crash-8.0.3_build.patch rename to SOURCES/crash-8.0.4_build.patch index bb09e2b..3db6f47 100644 --- a/SOURCES/crash-8.0.3_build.patch +++ b/SOURCES/crash-8.0.4_build.patch @@ -1,5 +1,5 @@ ---- crash-8.0.3/Makefile.orig -+++ crash-8.0.3/Makefile +--- crash-8.0.4/Makefile.orig ++++ crash-8.0.4/Makefile @@ -204,7 +204,7 @@ GDB_FLAGS= # TARGET_CFLAGS will be configured automatically by configure TARGET_CFLAGS= @@ -18,8 +18,8 @@ @echo "../../${PROGRAM} ../../${PROGRAM}lib.a" > ${GDB}/gdb/mergeobj @rm -f ${PROGRAM} @if [ ! -f ${GDB}/config.status ]; then \ ---- crash-8.0.3/configure.c.orig -+++ crash-8.0.3/configure.c +--- crash-8.0.4/configure.c.orig ++++ crash-8.0.4/configure.c @@ -810,7 +810,8 @@ build_configure(struct supported_gdb_version *sp) fprintf(fp2, "%s\n", sp->GDB); sprintf(target_data.gdb_version, "%s", &sp->GDB[4]); diff --git a/SOURCES/lzo_snappy_zstd.patch b/SOURCES/lzo_snappy_zstd.patch index 8d50246..67e4ad3 100644 --- a/SOURCES/lzo_snappy_zstd.patch +++ b/SOURCES/lzo_snappy_zstd.patch @@ -1,5 +1,5 @@ ---- crash-8.0.3/Makefile.orig -+++ crash-8.0.3/Makefile +--- crash-8.0.4/Makefile.orig ++++ crash-8.0.4/Makefile @@ -256,7 +256,7 @@ all: make_configure gdb_merge: force @if [ ! -f ${GDB}/README ]; then \ @@ -9,8 +9,8 @@ @echo "../../${PROGRAM} ../../${PROGRAM}lib.a" > ${GDB}/gdb/mergeobj @rm -f ${PROGRAM} @if [ ! -f ${GDB}/config.status ]; then \ ---- crash-8.0.3/diskdump.c.orig -+++ crash-8.0.3/diskdump.c +--- crash-8.0.4/diskdump.c.orig ++++ crash-8.0.4/diskdump.c @@ -23,6 +23,9 @@ * GNU General Public License for more details. */ diff --git a/SPECS/crash.spec b/SPECS/crash.spec index 2217bc9..a41a4dd 100644 --- a/SPECS/crash.spec +++ b/SPECS/crash.spec @@ -3,8 +3,8 @@ # Summary: Kernel analysis utility for live systems, netdump, diskdump, kdump, LKCD or mcore dumpfiles Name: crash -Version: 8.0.3 -Release: 1%{?dist} +Version: 8.0.4 +Release: 3%{?dist} License: GPLv3 Source0: https://github.com/crash-utility/crash/archive/crash-%{version}.tar.gz Source1: http://ftp.gnu.org/gnu/gdb/gdb-10.2.tar.gz @@ -18,7 +18,22 @@ Requires: binutils Provides: bundled(libiberty) Provides: bundled(gdb) = 10.2 Patch0: lzo_snappy_zstd.patch -Patch1: crash-8.0.3_build.patch +Patch1: crash-8.0.4_build.patch +Patch2: 0001-Fix-rd-command-for-zram-data-display-in-Linux-6.2-an.patch +Patch3: 0002-Fix-typos-in-offset_table-and-missing-help-o-items.patch +Patch4: 0003-zram-Fixes-for-lookup_swap_cache.patch +Patch5: 0004-symbols-expand-all-kernel-module-symtable-if-not-all.patch +Patch6: 0005-symbols-skip-load-.init.-sections-if-module-was-succ.patch +Patch7: 0006-use-NR_SWAPCACHE-when-nr_swapper_spaces-isn-t-availa.patch +Patch8: 0007-Fix-identity_map_base-value-dump-on-S390.patch +Patch9: 0008-s390x-fix-virtual-vs-physical-address-confusion.patch +Patch10: 0009-s390x-uncouple-physical-and-virtual-memory-spaces.patch +Patch11: 0010-RISCV64-Dump-NT_PRSTATUS-in-help-n.patch +Patch12: 0011-RISCV64-Fix-bt-output-when-no-ra-on-the-stack-top.patch +Patch13: 0012-arm64-rewrite-the-arm64_get_vmcoreinfo_ul-to-arm64_g.patch +Patch14: 0013-help.c-Remove-kmem-l-help-messages.patch +Patch15: 0014-x86_64-check-bt-bptr-before-calculate-framesize.patch +Patch16: 0001-symbols-skip-the-module-if-the-given-address-is-not-.patch %description The core analysis suite is a self-contained tool that can be used to @@ -38,8 +53,23 @@ offered by Mission Critical Linux, or the LKCD kernel patch. %prep %setup -n %{name}-%{version} -q -%patch0 -p1 -b lzo_snappy_zstd.patch -%patch1 -p1 -b crash-8.0.3_build.patch +%patch -P 0 -p1 -b lzo_snappy_zstd.patch +%patch -P 1 -p1 -b crash-8.0.4_build.patch +%patch -P 2 -p1 +%patch -P 3 -p1 +%patch -P 4 -p1 +%patch -P 5 -p1 +%patch -P 6 -p1 +%patch -P 7 -p1 +%patch -P 8 -p1 +%patch -P 9 -p1 +%patch -P 10 -p1 +%patch -P 11 -p1 +%patch -P 12 -p1 +%patch -P 13 -p1 +%patch -P 14 -p1 +%patch -P 15 -p1 +%patch -P 16 -p1 %build @@ -66,6 +96,17 @@ cp -p defs.h %{buildroot}%{_includedir}/crash %{_includedir}/* %changelog +* Mon Feb 05 2024 Tao Liu - 8.0.4-3 +- Fix bt takes many minutes on some pids in some vmcore + +* Tue Jan 02 2024 Lianbo Jiang - 8.0.4-2 +- Fix the "dis -lr" not displaying the source file names + and line numbers +- Fix incorrect symbol translation by the 'struct blk_mq_ops' + +* Thu Nov 16 2023 Lianbo Jiang - 8.0.4-1 +- Rebase to upstream crash 8.0.4 + * Thu May 04 2023 Lianbo Jiang - 8.0.3-1 - Rebase to upstream crash 8.0.3