Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1297506 > unrolled thread

[PATCH 0/2] [GIT PULL] ftrace/recordmcount: Fix hardlink issue with ccache

Back to article view | Back to linux.kernel

Contents

  [PATCH 0/2] [GIT PULL] ftrace/recordmcount: Fix hardlink issue with ccache Steven Rostedt <rostedt@goodmis.org> - 2015-12-23 16:30 +0100
    [PATCH 2/2] ftrace/scripts: Have recordmcount copy the object file Steven Rostedt <rostedt@goodmis.org> - 2015-12-23 16:30 +0100
    [PATCH 1/2] scripts: recordmcount: break hardlinks Steven Rostedt <rostedt@goodmis.org> - 2015-12-23 16:30 +0100

#1297506 — [PATCH 0/2] [GIT PULL] ftrace/recordmcount: Fix hardlink issue with ccache

Linus,

[ Well, as you know, my test box was not suffering from a harddrive
  failure, but a bug in the block layer. After wasting several days
  to figure that out, I got my test box up and running again, and was
  able to complete my testing. ]

Russell King was reporting lots of warnings when he compiled his kernel
with ftrace enabled. With some investigation it was discovered that it
was his compile setup. He was using ccache with hard links, which allowed
recordmcount to process the same .o twice. When this happens, recordmcount
will detect that it was already done and give a warning about it.

Russell fixed this by having recordmcount detect that the object file
has more than one hard link, and if it does, it unlinks the object file
after it maps it and processes then. This appears to fix the issue.

As you did not like the fact that recordmcount modified the file in place
and thought that it should do the modifications in memory and then write
it out to disk and move it over the old file to prevent other more subtle
issues like the one above, a second patch is added on top of Russell's to
do just that. Luckily the original code had write and lseek wrappers that
I was able to modify to not do inplace writes, but simply keep track
of the changes made in memory. When a write is made, a "update" flag is
set, and at the end of processing, if the update is set, then it writes
the file with changes out to a new file, and then renames it over the
original one.

The file descriptor is still passed to the write and lseek wrappers because
removing that would cause the change to be more intrusive. That can be
removed in a follow up cleanup patch that can wait till the next merge
window.

Please pull the latest trace-v4.4-rc4-2 tree, which can be found at:


  git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git
trace-v4.4-rc4-2

Tag SHA1: c8eaa389dd82351cf142d7b5419964dede5e797b
Head SHA1: a50bd43935586420fb75f4558369eb08566fac5e


Russell King (1):
      scripts: recordmcount: break hardlinks

Steven Rostedt (Red Hat) (1):
      ftrace/scripts: Have recordmcount copy the object file

----
 scripts/recordmcount.c | 137 ++++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 113 insertions(+), 24 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [next] | [standalone]

#1297507 — [PATCH 2/2] ftrace/scripts: Have recordmcount copy the object file

From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>

Russell King found that he had weird side effects when compiling the kernel
with hard linked ccache. The reason was that recordmcount modified the
kernel in place via mmap, and when a file gets modified twice by
recordmcount, it will complain about it. To fix this issue, Russell wrote a
patch that checked if the file was hard linked more than once and would
unlink it if it was.

Linus Torvalds was not happy with the fact that recordmcount does this in
place modification. Instead of doing the unlink only if the file has two or
more hard links, it does the unlink all the time. In otherwords, it always
does a copy if it changed something. That is, it does the write out if a
change was made.

Cc: stable@vger.kernel.org # 2.6.37+
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 scripts/recordmcount.c | 145 +++++++++++++++++++++++++++++++++++++------------
 1 file changed, 110 insertions(+), 35 deletions(-)

diff --git a/scripts/recordmcount.c b/scripts/recordmcount.c
index 91705ef30402..301d70b0174f 100644
--- a/scripts/recordmcount.c
+++ b/scripts/recordmcount.c
@@ -48,12 +48,17 @@
 
 static int fd_map;	/* File descriptor for file being modified. */
 static int mmap_failed; /* Boolean flag. */
-static void *ehdr_curr; /* current ElfXX_Ehdr *  for resource cleanup */
 static char gpfx;	/* prefix for global symbol name (sometimes '_') */
 static struct stat sb;	/* Remember .st_size, etc. */
 static jmp_buf jmpenv;	/* setjmp/longjmp per-file error escape */
 static const char *altmcount;	/* alternate mcount symbol name */
 static int warn_on_notrace_sect; /* warn when section has mcount not being recorded */
+static void *file_map;	/* pointer of the mapped file */
+static void *file_end;	/* pointer to the end of the mapped file */
+static int file_updated; /* flag to state file was changed */
+static void *file_ptr;	/* current file pointer location */
+static void *file_append; /* added to the end of the file */
+static size_t file_append_size; /* how much is added to end of file */
 
 /* setjmp() return values */
 enum {
@@ -67,10 +72,14 @@ static void
 cleanup(void)
 {
 	if (!mmap_failed)
-		munmap(ehdr_curr, sb.st_size);
+		munmap(file_map, sb.st_size);
 	else
-		free(ehdr_curr);
-	close(fd_map);
+		free(file_map);
+	file_map = NULL;
+	free(file_append);
+	file_append = NULL;
+	file_append_size = 0;
+	file_updated = 0;
 }
 
 static void __attribute__((noreturn))
@@ -92,12 +101,22 @@ succeed_file(void)
 static off_t
 ulseek(int const fd, off_t const offset, int const whence)
 {
-	off_t const w = lseek(fd, offset, whence);
-	if (w == (off_t)-1) {
-		perror("lseek");
+	switch (whence) {
+	case SEEK_SET:
+		file_ptr = file_map + offset;
+		break;
+	case SEEK_CUR:
+		file_ptr += offset;
+		break;
+	case SEEK_END:
+		file_ptr = file_map + (sb.st_size - offset);
+		break;
+	}
+	if (file_ptr < file_map) {
+		fprintf(stderr, "lseek: seek before file\n");
 		fail_file();
 	}
-	return w;
+	return file_ptr - file_map;
 }
 
 static size_t
@@ -114,12 +133,38 @@ uread(int const fd, void *const buf, size_t const count)
 static size_t
 uwrite(int const fd, void const *const buf, size_t const count)
 {
-	size_t const n = write(fd, buf, count);
-	if (n != count) {
-		perror("write");
-		fail_file();
+	size_t cnt = count;
+	off_t idx = 0;
+
+	file_updated = 1;
+
+	if (file_ptr + count >= file_end) {
+		off_t aoffset = (file_ptr + count) - file_end;
+
+		if (aoffset > file_append_size) {
+			file_append = realloc(file_append, aoffset);
+			file_append_size = aoffset;
+		}
+		if (!file_append) {
+			perror("write");
+			fail_file();
+		}
+		if (file_ptr < file_end) {
+			cnt = file_end - file_ptr;
+		} else {
+			cnt = 0;
+			idx = aoffset - count;
+		}
 	}
-	return n;
+
+	if (cnt)
+		memcpy(file_ptr, buf, cnt);
+
+	if (cnt < count)
+		memcpy(file_append + idx, buf + cnt, count - cnt);
+
+	file_ptr += count;
+	return count;
 }
 
 static void *
@@ -192,9 +237,7 @@ static int make_nop_arm64(void *map, size_t const offset)
  */
 static void *mmap_file(char const *fname)
 {
-	void *addr;
-
-	fd_map = open(fname, O_RDWR);
+	fd_map = open(fname, O_RDONLY);
 	if (fd_map < 0 || fstat(fd_map, &sb) < 0) {
 		perror(fname);
 		fail_file();
@@ -203,29 +246,58 @@ static void *mmap_file(char const *fname)
 		fprintf(stderr, "not a regular file: %s\n", fname);
 		fail_file();
 	}
-	addr = mmap(0, sb.st_size, PROT_READ|PROT_WRITE, MAP_PRIVATE,
-		    fd_map, 0);
+	file_map = mmap(0, sb.st_size, PROT_READ|PROT_WRITE, MAP_PRIVATE,
+			fd_map, 0);
 	mmap_failed = 0;
-	if (addr == MAP_FAILED) {
+	if (file_map == MAP_FAILED) {
 		mmap_failed = 1;
-		addr = umalloc(sb.st_size);
-		uread(fd_map, addr, sb.st_size);
+		file_map = umalloc(sb.st_size);
+		uread(fd_map, file_map, sb.st_size);
 	}
-	if (sb.st_nlink != 1) {
-		/* file is hard-linked, break the hard link */
-		close(fd_map);
-		if (unlink(fname) < 0) {
-			perror(fname);
-			fail_file();
-		}
-		fd_map = open(fname, O_RDWR | O_CREAT, sb.st_mode);
-		if (fd_map < 0) {
-			perror(fname);
+	close(fd_map);
+
+	file_end = file_map + sb.st_size;
+
+	return file_map;
+}
+
+static void write_file(const char *fname)
+{
+	char tmp_file[strlen(fname) + 4];
+	size_t n;
+
+	if (!file_updated)
+		return;
+
+	sprintf(tmp_file, "%s.rc", fname);
+
+	/*
+	 * After reading the entire file into memory, delete it
+	 * and write it back, to prevent weird side effects of modifying
+	 * an object file in place.
+	 */
+	fd_map = open(tmp_file, O_WRONLY | O_TRUNC | O_CREAT, sb.st_mode);
+	if (fd_map < 0) {
+		perror(fname);
+		fail_file();
+	}
+	n = write(fd_map, file_map, sb.st_size);
+	if (n != sb.st_size) {
+		perror("write");
+		fail_file();
+	}
+	if (file_append_size) {
+		n = write(fd_map, file_append, file_append_size);
+		if (n != file_append_size) {
+			perror("write");
 			fail_file();
 		}
-		uwrite(fd_map, addr, sb.st_size);
 	}
-	return addr;
+	close(fd_map);
+	if (rename(tmp_file, fname) < 0) {
+		perror(fname);
+		fail_file();
+	}
 }
 
 /* w8rev, w8nat, ...: Handle endianness. */
@@ -332,7 +404,6 @@ do_file(char const *const fname)
 	Elf32_Ehdr *const ehdr = mmap_file(fname);
 	unsigned int reltype = 0;
 
-	ehdr_curr = ehdr;
 	w = w4nat;
 	w2 = w2nat;
 	w8 = w8nat;
@@ -455,6 +526,7 @@ do_file(char const *const fname)
 	}
 	}  /* end switch */
 
+	write_file(fname);
 	cleanup();
 }
 
@@ -507,11 +579,14 @@ main(int argc, char *argv[])
 		case SJ_SETJMP:    /* normal sequence */
 			/* Avoid problems if early cleanup() */
 			fd_map = -1;
-			ehdr_curr = NULL;
 			mmap_failed = 1;
+			file_map = NULL;
+			file_ptr = NULL;
+			file_updated = 0;
 			do_file(file);
 			break;
 		case SJ_FAIL:    /* error in do_file or below */
+			sprintf("%s: failed\n", file);
 			++n_error;
 			break;
 		case SJ_SUCCEED:    /* premature success */
-- 
2.6.2


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1297508 — [PATCH 1/2] scripts: recordmcount: break hardlinks

From: Russell King <rmk+kernel@arm.linux.org.uk>

recordmcount edits the file in-place, which can cause problems when
using ccache in hardlink mode.  Arrange for recordmcount to break a
hardlinked object.

Link: http://lkml.kernel.org/r/E1a7MVT-0000et-62@rmk-PC.arm.linux.org.uk

Cc: stable@vger.kernel.org # 2.6.37+
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 scripts/recordmcount.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/scripts/recordmcount.c b/scripts/recordmcount.c
index 698768bdc581..91705ef30402 100644
--- a/scripts/recordmcount.c
+++ b/scripts/recordmcount.c
@@ -211,6 +211,20 @@ static void *mmap_file(char const *fname)
 		addr = umalloc(sb.st_size);
 		uread(fd_map, addr, sb.st_size);
 	}
+	if (sb.st_nlink != 1) {
+		/* file is hard-linked, break the hard link */
+		close(fd_map);
+		if (unlink(fname) < 0) {
+			perror(fname);
+			fail_file();
+		}
+		fd_map = open(fname, O_RDWR | O_CREAT, sb.st_mode);
+		if (fd_map < 0) {
+			perror(fname);
+			fail_file();
+		}
+		uwrite(fd_map, addr, sb.st_size);
+	}
 	return addr;
 }
 
-- 
2.6.2


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web