libpmcstat: Fix a few ARM-specific issues with function symbols.

- Refine the checks for ARM mapping symbols and apply them on arm64 as
  well as 32-bit arm.  In particular, mapping symbols can have
  additional characters and are not strictly limited to just "$a" but
  can append additional characters (e.g. "$a.1").  Add "$x" to the
  list of mapping symbol prefixes.

- Clear the LSB of function symbol addresses.  Thumb function
  addresses set the LSB to enable Thumb mode.  However, the actual
  function starts at the aligned address with LSB clear.  Not clearing
  the LSB can cause pmcannotate to pass misaligned addresses to
  objdump when extracting disassembly.

Reviewed by:	andrew
Obtained from:	CheriBSD
Sponsored by:	University of Cambridge, Google, Inc.
Differential Revision:	https://reviews.freebsd.org/D34416

1 files changed, 14 insertions, 4 deletions

diff --git a/lib/libpmcstat/libpmcstat_image.c b/lib/libpmcstat/libpmcstat_image.c
index 97109f203806..c63f00e03253 100644
--- a/lib/libpmcstat/libpmcstat_image.c
+++ b/lib/libpmcstat/libpmcstat_image.c
@@ -119,13 +119,23 @@ pmcstat_image_add_symbols(struct pmcstat_image *image, Elf *e,
 		if ((fnname = elf_strptr(e, sh->sh_link, sym.st_name))
 		    == NULL)
 			continue;
-#ifdef __arm__
-		/* Remove spurious ARM function name. */
+
+#if defined(__aarch64__) || defined(__arm__)
+		/* Ignore ARM mapping symbols. */
 		if (fnname[0] == '$' &&
 		    (fnname[1] == 'a' || fnname[1] == 't' ||
-		    fnname[1] == 'd') &&
-		    fnname[2] == '\0')
+		    fnname[1] == 'd' || fnname[1] == 'x'))
 			continue;
+
+		/*
+		 * Clear LSB from starting addresses for functions
+		 * which execute in Thumb mode.  We should perhaps
+		 * only do this for functions in a $t mapping symbol
+		 * range, but parsing mapping symbols would be a lot
+		 * of work and function addresses shouldn't have the
+		 * LSB set otherwise.
+		 */
+		sym.st_value &= ~1;
 #endif
 
 		symptr->ps_name  = pmcstat_string_intern(fnname);