From a5f7a5bb3b7f28ba7e4fa246775b29a0e5537255 Mon Sep 17 00:00:00 2001 From: Jinke Han Date: Tue, 8 Sep 2026 15:37:42 +0800 Subject: x86/kprobes: Fix crash when probing CS CALL instructions When using eBPF to probe CS CALL instructions within a function, a crash can be triggered. The eBPF tool probes offset 257 of the __hrtimer_run_queues() function: <__hrtimer_run_queues+249>: nopl 0x0(%rax,%rax,1) <__hrtimer_run_queues+254>: mov %r14,%rdi <__hrtimer_run_queues+257>: cs call <__x86_indirect_thunk_r12> <__hrtimer_run_queues+263>: mov %eax,%r12d <__hrtimer_run_queues+266>: xchg %ax,%ax <__hrtimer_run_queues+268>: mov %r13,%rdi Which triggers this crash: BUG: unable to handle page fault for address: 00000000000f41c9 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 0 P4D 0 Oops: 0002 [#1] SMP NOPTI CPU: 1 PID: 0 Comm: swapper/1 Kdump: loaded Tainted: P RIP: 0010:__hrtimer_run_queues+0x106/0x230 Note that __hrtimer_run_queues+0x106 is __hrtimer_run_queues+262, which is at the 6th byte of the above CS CALL instruction. Since the CS CALL instruction occupies 6 bytes, the exception occurred in the middle of that call instruction. The root cause is that when using eBPF tools to probe in the middle of a function, a kprobe with INT3 is used as the underlying implementation. During single-step emulation of the original CALL instruction, int3_emulate_call() assumes that the probed CALL instruction is 5 bytes long. However, the actual CS-prefixed CALL instruction occupies 6 bytes, so it constructs an incorrect exception return address. When the CPU returns from the kprobe handler, the next instruction to be executed is at the address of the last byte of that CS CALL instruction. Coincidentally, starting from that address, the CPU fetches and decodes a completely different instruction, which ultimately triggers a kernel crash. Fix the issue by using the actual instruction length obtained from the instruction decoder when constructing the exception return address, rather than relying on the hardcoded CALL_INSN_SIZE macro. [ mingo: Refined the changelog ] Fixes: 6256e668b7af ("x86/kprobes: Use int3 instead of debug trap for single-step") Suggested-by: Masami Hiramatsu (Google) Signed-off-by: Jinke Han Signed-off-by: Ingo Molnar Reviewed-by: Masami Hiramatsu (Google) Acked-by: Yafang Shao Acked-by: Borislav Petkov Cc: Peter Zijlstra Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000 --- fs/jffs2/read.c | 228 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 228 insertions(+) create mode 100644 fs/jffs2/read.c (limited to 'fs/jffs2/read.c') diff --git a/fs/jffs2/read.c b/fs/jffs2/read.c new file mode 100644 index 000000000..0b042b1fc --- /dev/null +++ b/fs/jffs2/read.c @@ -0,0 +1,228 @@ +/* + * JFFS2 -- Journalling Flash File System, Version 2. + * + * Copyright © 2001-2007 Red Hat, Inc. + * + * Created by David Woodhouse + * + * For licensing information, see the file 'LICENCE' in this directory. + * + */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#include +#include +#include +#include +#include +#include +#include "nodelist.h" +#include "compr.h" + +int jffs2_read_dnode(struct jffs2_sb_info *c, struct jffs2_inode_info *f, + struct jffs2_full_dnode *fd, unsigned char *buf, + int ofs, int len) +{ + struct jffs2_raw_inode *ri; + size_t readlen; + uint32_t crc; + unsigned char *decomprbuf = NULL; + unsigned char *readbuf = NULL; + int ret = 0; + + ri = jffs2_alloc_raw_inode(); + if (!ri) + return -ENOMEM; + + ret = jffs2_flash_read(c, ref_offset(fd->raw), sizeof(*ri), &readlen, (char *)ri); + if (ret) { + jffs2_free_raw_inode(ri); + pr_warn("Error reading node from 0x%08x: %d\n", + ref_offset(fd->raw), ret); + return ret; + } + if (readlen != sizeof(*ri)) { + jffs2_free_raw_inode(ri); + pr_warn("Short read from 0x%08x: wanted 0x%zx bytes, got 0x%zx\n", + ref_offset(fd->raw), sizeof(*ri), readlen); + return -EIO; + } + crc = crc32(0, ri, sizeof(*ri)-8); + + jffs2_dbg(1, "Node read from %08x: node_crc %08x, calculated CRC %08x. dsize %x, csize %x, offset %x, buf %p\n", + ref_offset(fd->raw), je32_to_cpu(ri->node_crc), + crc, je32_to_cpu(ri->dsize), je32_to_cpu(ri->csize), + je32_to_cpu(ri->offset), buf); + if (crc != je32_to_cpu(ri->node_crc)) { + pr_warn("Node CRC %08x != calculated CRC %08x for node at %08x\n", + je32_to_cpu(ri->node_crc), crc, ref_offset(fd->raw)); + ret = -EIO; + goto out_ri; + } + /* There was a bug where we wrote hole nodes out with csize/dsize + swapped. Deal with it */ + if (ri->compr == JFFS2_COMPR_ZERO && !je32_to_cpu(ri->dsize) && + je32_to_cpu(ri->csize)) { + ri->dsize = ri->csize; + ri->csize = cpu_to_je32(0); + } + + D1(if(ofs + len > je32_to_cpu(ri->dsize)) { + pr_warn("jffs2_read_dnode() asked for %d bytes at %d from %d-byte node\n", + len, ofs, je32_to_cpu(ri->dsize)); + ret = -EINVAL; + goto out_ri; + }); + + + if (ri->compr == JFFS2_COMPR_ZERO) { + memset(buf, 0, len); + goto out_ri; + } + + /* Cases: + Reading whole node and it's uncompressed - read directly to buffer provided, check CRC. + Reading whole node and it's compressed - read into comprbuf, check CRC and decompress to buffer provided + Reading partial node and it's uncompressed - read into readbuf, check CRC, and copy + Reading partial node and it's compressed - read into readbuf, check checksum, decompress to decomprbuf and copy + */ + if (ri->compr == JFFS2_COMPR_NONE && len == je32_to_cpu(ri->dsize)) { + readbuf = buf; + } else { + readbuf = kmalloc(je32_to_cpu(ri->csize), GFP_KERNEL); + if (!readbuf) { + ret = -ENOMEM; + goto out_ri; + } + } + if (ri->compr != JFFS2_COMPR_NONE) { + if (len < je32_to_cpu(ri->dsize)) { + decomprbuf = kmalloc(je32_to_cpu(ri->dsize), GFP_KERNEL); + if (!decomprbuf) { + ret = -ENOMEM; + goto out_readbuf; + } + } else { + decomprbuf = buf; + } + } else { + decomprbuf = readbuf; + } + + jffs2_dbg(2, "Read %d bytes to %p\n", je32_to_cpu(ri->csize), + readbuf); + ret = jffs2_flash_read(c, (ref_offset(fd->raw)) + sizeof(*ri), + je32_to_cpu(ri->csize), &readlen, readbuf); + + if (!ret && readlen != je32_to_cpu(ri->csize)) + ret = -EIO; + if (ret) + goto out_decomprbuf; + + crc = crc32(0, readbuf, je32_to_cpu(ri->csize)); + if (crc != je32_to_cpu(ri->data_crc)) { + pr_warn("Data CRC %08x != calculated CRC %08x for node at %08x\n", + je32_to_cpu(ri->data_crc), crc, ref_offset(fd->raw)); + ret = -EIO; + goto out_decomprbuf; + } + jffs2_dbg(2, "Data CRC matches calculated CRC %08x\n", crc); + if (ri->compr != JFFS2_COMPR_NONE) { + jffs2_dbg(2, "Decompress %d bytes from %p to %d bytes at %p\n", + je32_to_cpu(ri->csize), readbuf, + je32_to_cpu(ri->dsize), decomprbuf); + ret = jffs2_decompress(c, f, ri->compr | (ri->usercompr << 8), readbuf, decomprbuf, je32_to_cpu(ri->csize), je32_to_cpu(ri->dsize)); + if (ret) { + pr_warn("Error: jffs2_decompress returned %d\n", ret); + goto out_decomprbuf; + } + } + + if (len < je32_to_cpu(ri->dsize)) { + memcpy(buf, decomprbuf+ofs, len); + } + out_decomprbuf: + if(decomprbuf != buf && decomprbuf != readbuf) + kfree(decomprbuf); + out_readbuf: + if(readbuf != buf) + kfree(readbuf); + out_ri: + jffs2_free_raw_inode(ri); + + return ret; +} + +int jffs2_read_inode_range(struct jffs2_sb_info *c, struct jffs2_inode_info *f, + unsigned char *buf, uint32_t offset, uint32_t len) +{ + uint32_t end = offset + len; + struct jffs2_node_frag *frag; + int ret; + + jffs2_dbg(1, "%s(): ino #%u, range 0x%08x-0x%08x\n", + __func__, f->inocache->ino, offset, offset + len); + + frag = jffs2_lookup_node_frag(&f->fragtree, offset); + + /* XXX FIXME: Where a single physical node actually shows up in two + frags, we read it twice. Don't do that. */ + /* Now we're pointing at the first frag which overlaps our page + * (or perhaps is before it, if we've been asked to read off the + * end of the file). */ + while(offset < end) { + jffs2_dbg(2, "%s(): offset %d, end %d\n", + __func__, offset, end); + if (unlikely(!frag || frag->ofs > offset || + frag->ofs + frag->size <= offset)) { + uint32_t holesize = end - offset; + if (frag && frag->ofs > offset) { + jffs2_dbg(1, "Eep. Hole in ino #%u fraglist. frag->ofs = 0x%08x, offset = 0x%08x\n", + f->inocache->ino, frag->ofs, offset); + holesize = min(holesize, frag->ofs - offset); + } + jffs2_dbg(1, "Filling non-frag hole from %d-%d\n", + offset, offset + holesize); + memset(buf, 0, holesize); + buf += holesize; + offset += holesize; + continue; + } else if (unlikely(!frag->node)) { + uint32_t holeend = min(end, frag->ofs + frag->size); + jffs2_dbg(1, "Filling frag hole from %d-%d (frag 0x%x 0x%x)\n", + offset, holeend, frag->ofs, + frag->ofs + frag->size); + memset(buf, 0, holeend - offset); + buf += holeend - offset; + offset = holeend; + frag = frag_next(frag); + continue; + } else { + uint32_t readlen; + uint32_t fragofs; /* offset within the frag to start reading */ + + fragofs = offset - frag->ofs; + readlen = min(frag->size - fragofs, end - offset); + jffs2_dbg(1, "Reading %d-%d from node at 0x%08x (%d)\n", + frag->ofs+fragofs, + frag->ofs + fragofs+readlen, + ref_offset(frag->node->raw), + ref_flags(frag->node->raw)); + ret = jffs2_read_dnode(c, f, frag->node, buf, fragofs + frag->ofs - frag->node->ofs, readlen); + jffs2_dbg(2, "node read done\n"); + if (ret) { + jffs2_dbg(1, "%s(): error %d\n", + __func__, ret); + memset(buf, 0, readlen); + return ret; + } + buf += readlen; + offset += readlen; + frag = frag_next(frag); + jffs2_dbg(2, "node read was OK. Looping\n"); + } + } + return 0; +} + -- cgit v1.3.1