From: Steve McIntyre <steve.mcintyre@linaro.org>
To: libc-ports@sourceware.org
Cc: Carlos O'Donell <carlos@systemhalted.org>,
"Joseph S. Myers" <joseph@codesourcery.com>,
Marcus Shawcroft <marcus.shawcroft@linaro.org>
Subject: [PATCH] ARM: Identify the hard-float ABI in older ARM binaries
Date: Mon, 03 Dec 2012 15:42:00 -0000 [thread overview]
Message-ID: <20121203154154.GF22509@einval.com> (raw)
In-Reply-To: <20121203153747.GE22509@einval.com>
Hi,
I expect this one to be more controversial and I'll understand if
people don't want to accept it! It's a variation on the code that's in
use today in Debian/Ubuntu to be able to distinguish between soft- and
hard-float ABI libraries. In time, the new ELF flags will make this
obsolete. Until that time, this adds a fallback for older binaries
created before those flags went into binutils.
==============================
Identify the hard-float ABI in older ARM binaries
* ports/sysdeps/unix/sysv/linux/arm/readelflib.c (read_uleb128):
New function.
(is_library_hf): Ditto.
(process_elf_file): Call is_library_hf() to detect HF ABI if
we don't have clear evidence from the ELF flags.
Signed-off-by: Steve McIntyre <steve.mcintyre@linaro.org>
---
ports/sysdeps/unix/sysv/linux/arm/readelflib.c | 129 ++++++++++++++++++++++++
1 file changed, 129 insertions(+)
diff --git a/ports/sysdeps/unix/sysv/linux/arm/readelflib.c b/ports/sysdeps/unix/sysv/linux/arm/readelflib.c
index e4ba2b0..0287644 100644
--- a/ports/sysdeps/unix/sysv/linux/arm/readelflib.c
+++ b/ports/sysdeps/unix/sysv/linux/arm/readelflib.c
@@ -25,6 +25,133 @@ int process_elf64_file (const char *file_name, const char *lib, int *flag,
unsigned int *osversion, char **soname,
void *file_contents, size_t file_length);
+#undef __ELF_NATIVE_CLASS
+#define __ELF_NATIVE_CLASS 32
+
+/* Read an unsigned leb128 value from P, store the value in VAL, return
+ P incremented past the value. We assume that a word is large enough to
+ hold any value so encoded; if it is smaller than a pointer on some target,
+ pointers should not be leb128 encoded on that target. */
+static const unsigned char *
+read_uleb128 (const unsigned char *p, unsigned long *val)
+{
+ unsigned int shift = 0;
+ unsigned char byte;
+ unsigned long result;
+
+ result = 0;
+ do
+ {
+ byte = *p++;
+ result |= (byte & 0x7f) << shift;
+ shift += 7;
+ }
+ while (byte & 0x80);
+
+ *val = result;
+ return p;
+}
+
+#define ATTR_TAG_FILE 1
+#define ABI_VFP_args 28
+#define VFP_ARGS_IN_VFP_REGS 1
+
+/* Check the ABI in the ARM attributes. Search through the section
+ headers looking for the ARM attributes section, then check the
+ VFP_ARGS attribute. */
+static int is_library_hf(const char *file_name, void *file_contents, size_t file_length)
+{
+ unsigned int i;
+ ElfW(Ehdr) *ehdr = (ElfW(Ehdr) *) file_contents;
+ ElfW(Shdr) *shdrs;
+
+ shdrs = file_contents + ehdr->e_shoff;
+ for (i = 0; i < ehdr->e_shnum; i++)
+ {
+ if (SHT_ARM_ATTRIBUTES == shdrs[i].sh_type)
+ {
+ /* We've found a likely section. Load the contents and
+ * check the tags */
+ unsigned char *p = (unsigned char *)file_contents + shdrs[i].sh_offset;
+ unsigned char * end;
+
+ /* Sanity-check the attribute section details. Make sure
+ * that it's the "aeabi" section, that's all we care
+ * about. */
+ if (*p == 'A')
+ {
+ unsigned long len = shdrs[i].sh_size - 1;
+ unsigned long namelen;
+ p++;
+
+ while (len > 0)
+ {
+ unsigned long section_len = p[0] | p[1] << 8 | p[2] << 16 | p[3] << 24;
+ if (section_len > len)
+ section_len = len;
+
+ p += 4;
+ len -= section_len;
+ section_len -= 4;
+
+ if (0 != strcmp((char *)p, "aeabi"))
+ {
+ p += section_len;
+ continue;
+ }
+ namelen = strlen((char *)p) + 1;
+ p += namelen;
+ section_len -= namelen;
+
+ /* We're in a valid section. Walk through this
+ * section looking for the tag we care about
+ * (ABI_VFP_args) */
+ while (section_len > 0)
+ {
+ unsigned long tag, val = 0;
+ unsigned long size;
+
+ end = p;
+ tag = (*p++);
+ size = p[0] | p[1] << 8 | p[2] << 16 | p[3] << 24;
+ if (size > section_len)
+ size = section_len;
+ p += 4;
+
+ section_len -= size;
+ end += size;
+ if (ATTR_TAG_FILE != tag)
+ {
+ /* ignore, we don't care */
+ p = end;
+ continue;
+ }
+ while (p < end)
+ {
+ p = read_uleb128 (p, &tag);
+ /* Handle the different types of tag. */
+ if ( (tag == 4) || (tag == 5) || (tag == 67) )
+ {
+ /* Special cases for string values */
+ namelen = strlen((char *)p) + 1;
+ p += namelen;
+ }
+ else
+ {
+ p = read_uleb128 (p, &val);
+ }
+ if ( (tag == ABI_VFP_args) && (val == VFP_ARGS_IN_VFP_REGS) )
+ return 1;
+ }
+ }
+ }
+ }
+ }
+ }
+ return 0;
+}
+
+
/* Returns 0 if everything is ok, != 0 in case of error. */
int
process_elf_file (const char *file_name, const char *lib, int *flag,
@@ -47,6 +174,8 @@ process_elf_file (const char *file_name, const char *lib, int *flag,
*flag = FLAG_ARM_LIBHF|FLAG_ELF_LIBC6;
else if (elf32_header->e_flags & EF_ARM_ABI_FLOAT_SOFT)
*flag = FLAG_ELF_LIBC6;
+ else if (is_library_hf(file_name, file_contents, file_length))
+ *flag = FLAG_ARM_LIBHF|FLAG_ELF_LIBC6;
}
}
else
--
1.7.10.4
Cheers,
--
Steve McIntyre steve.mcintyre@linaro.org
<http://www.linaro.org/> Linaro.org | Open source software for ARM SoCs
next prev parent reply other threads:[~2012-12-03 15:42 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-12-03 15:35 [PATCH] ARM: Tag ARM and AArch64 binaries in the ldconfig cache Steve McIntyre
2012-12-03 15:38 ` [PATCH] ARM: Check for the FLAG_ARM_LIBHF flag " Steve McIntyre
2012-12-03 15:42 ` Steve McIntyre [this message]
2012-12-03 16:00 ` [PATCH] ARM: Identify the hard-float ABI in older ARM binaries Joseph S. Myers
2012-12-03 16:44 ` Steve McIntyre
2012-12-03 16:49 ` Joseph S. Myers
2012-12-03 16:54 ` Steve McIntyre
2012-12-03 16:59 ` Joseph S. Myers
2012-12-03 17:11 ` Steve McIntyre
2012-12-03 15:59 ` [PATCH] ARM: Check for the FLAG_ARM_LIBHF flag in the ldconfig cache Joseph S. Myers
2012-12-03 16:32 ` Steve McIntyre
2012-12-03 15:44 ` [PATCH] ARM: Tag ARM libc6-dependent binaries with FLAG_ELF_LIBC6 Steve McIntyre
2012-12-03 16:00 ` Joseph S. Myers
2012-12-03 15:58 ` [PATCH] ARM: Tag ARM and AArch64 binaries in the ldconfig cache Joseph S. Myers
2012-12-03 16:31 ` Steve McIntyre
2013-01-09 2:14 ` Joseph S. Myers
2013-01-09 13:50 ` Steve McIntyre
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20121203154154.GF22509@einval.com \
--to=steve.mcintyre@linaro.org \
--cc=carlos@systemhalted.org \
--cc=joseph@codesourcery.com \
--cc=libc-ports@sourceware.org \
--cc=marcus.shawcroft@linaro.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).