From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 12354 invoked by alias); 11 Oct 2011 09:30:20 -0000 Received: (qmail 12340 invoked by uid 22791); 11 Oct 2011 09:30:18 -0000 X-SWARE-Spam-Status: No, hits=0.5 required=5.0 tests=AWL,BAYES_99,RP_MATCHES_RCVD X-Spam-Check-By: sourceware.org Received: from sibelius.xs4all.nl (HELO glazunov.sibelius.xs4all.nl) (83.163.83.176) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 11 Oct 2011 09:30:03 +0000 Received: from glazunov.sibelius.xs4all.nl (kettenis@localhost [127.0.0.1]) by glazunov.sibelius.xs4all.nl (8.14.5/8.14.3) with ESMTP id p9B9TtCh021486; Tue, 11 Oct 2011 11:29:55 +0200 (CEST) Received: (from kettenis@localhost) by glazunov.sibelius.xs4all.nl (8.14.5/8.14.3/Submit) id p9B9Tsx2019112; Tue, 11 Oct 2011 11:29:54 +0200 (CEST) Date: Tue, 11 Oct 2011 09:30:00 -0000 Message-Id: <201110110929.p9B9Tsx2019112@glazunov.sibelius.xs4all.nl> From: Mark Kettenis To: davem@davemloft.net CC: gdb-patches@sourceware.org In-reply-to: <20111010.225847.435074482632472630.davem@davemloft.net> (message from David Miller on Mon, 10 Oct 2011 22:58:47 -0400 (EDT)) Subject: Re: [PATCH] Fix jit.exp on most 32-bit targets. References: <20111010.225847.435074482632472630.davem@davemloft.net> Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2011-10/txt/msg00299.txt.bz2 > Date: Mon, 10 Oct 2011 22:58:47 -0400 (EDT) > From: David Miller > > Outside of i386, most 32-bit targets align long long types on a 64-bit > boundary. But the JIT support code reading in the descriptor from the > target just assumes that the size field will come right after the > first three pointer sized fields. > > This makes jit.exp fail. > > Ok to commit? Somehow this feels wrong to me. But I guess it is too late to make the JIT people change the design of their interface :(. > 2011-10-10 David S. Miller > > * gdbarch.sh: New field 'long_long_align_bit'. > * gdbarch.c, gdbarch.h: Regenerate. > * i386-tdep.c (i386_gdbarch_init): Set long_long_align_bit to 32. > * jit.c (jit_read_code_entry): Use it to determine correct size offset. > > diff --git a/gdb/gdbarch.c b/gdb/gdbarch.c > index 2b892b6..ea48fa7 100644 > --- a/gdb/gdbarch.c > +++ b/gdb/gdbarch.c > @@ -145,6 +145,7 @@ struct gdbarch > int int_bit; > int long_bit; > int long_long_bit; > + int long_long_align_bit; > int half_bit; > const struct floatformat ** half_format; > int float_bit; > @@ -299,6 +300,7 @@ struct gdbarch startup_gdbarch = > 8 * sizeof (int), /* int_bit */ > 8 * sizeof (long), /* long_bit */ > 8 * sizeof (LONGEST), /* long_long_bit */ > + 8 * sizeof (LONGEST), /* long_long_align_bit */ > 16, /* half_bit */ > 0, /* half_format */ > 8 * sizeof (float), /* float_bit */ > @@ -463,6 +465,7 @@ gdbarch_alloc (const struct gdbarch_info *info, > gdbarch->int_bit = 4*TARGET_CHAR_BIT; > gdbarch->long_bit = 4*TARGET_CHAR_BIT; > gdbarch->long_long_bit = 2*gdbarch->long_bit; > + gdbarch->long_long_align_bit = 2*gdbarch->long_bit; > gdbarch->half_bit = 2*TARGET_CHAR_BIT; > gdbarch->float_bit = 4*TARGET_CHAR_BIT; > gdbarch->double_bit = 8*TARGET_CHAR_BIT; > @@ -576,6 +579,7 @@ verify_gdbarch (struct gdbarch *gdbarch) > /* Skip verify of int_bit, invalid_p == 0 */ > /* Skip verify of long_bit, invalid_p == 0 */ > /* Skip verify of long_long_bit, invalid_p == 0 */ > + /* Skip verify of long_long_align_bit, invalid_p == 0 */ > /* Skip verify of half_bit, invalid_p == 0 */ > if (gdbarch->half_format == 0) > gdbarch->half_format = floatformats_ieee_half; > @@ -1020,6 +1024,9 @@ gdbarch_dump (struct gdbarch *gdbarch, struct ui_file *file) > "gdbarch_dump: long_double_format = %s\n", > pformat (gdbarch->long_double_format)); > fprintf_unfiltered (file, > + "gdbarch_dump: long_long_align_bit = %s\n", > + plongest (gdbarch->long_long_align_bit)); > + fprintf_unfiltered (file, > "gdbarch_dump: long_long_bit = %s\n", > plongest (gdbarch->long_long_bit)); > fprintf_unfiltered (file, > @@ -1424,6 +1431,23 @@ set_gdbarch_long_long_bit (struct gdbarch *gdbarch, > } > > int > +gdbarch_long_long_align_bit (struct gdbarch *gdbarch) > +{ > + gdb_assert (gdbarch != NULL); > + /* Skip verify of long_long_align_bit, invalid_p == 0 */ > + if (gdbarch_debug >= 2) > + fprintf_unfiltered (gdb_stdlog, "gdbarch_long_long_align_bit called\n"); > + return gdbarch->long_long_align_bit; > +} > + > +void > +set_gdbarch_long_long_align_bit (struct gdbarch *gdbarch, > + int long_long_align_bit) > +{ > + gdbarch->long_long_align_bit = long_long_align_bit; > +} > + > +int > gdbarch_half_bit (struct gdbarch *gdbarch) > { > gdb_assert (gdbarch != NULL); > diff --git a/gdb/gdbarch.h b/gdb/gdbarch.h > index 0117322..d8420cc 100644 > --- a/gdb/gdbarch.h > +++ b/gdb/gdbarch.h > @@ -120,6 +120,12 @@ extern void set_gdbarch_long_bit (struct gdbarch *gdbarch, int long_bit); > extern int gdbarch_long_long_bit (struct gdbarch *gdbarch); > extern void set_gdbarch_long_long_bit (struct gdbarch *gdbarch, int long_long_bit); > > +/* Alignment of a long long or unsigned long long for the target > + machine. */ > + > +extern int gdbarch_long_long_align_bit (struct gdbarch *gdbarch); > +extern void set_gdbarch_long_long_align_bit (struct gdbarch *gdbarch, int long_long_align_bit); > + > /* The ABI default bit-size and format for "half", "float", "double", and > "long double". These bit/format pairs should eventually be combined > into a single object. For the moment, just initialize them as a pair. > @@ -586,7 +592,7 @@ extern void set_gdbarch_smash_text_address (struct gdbarch *gdbarch, gdbarch_sma > FIXME/cagney/2001-01-18: The logic is backwards. It should be asking if the > target can single step. If not, then implement single step using breakpoints. > > - A return value of 1 means that the software_single_step breakpoints > + A return value of 1 means that the software_single_step breakpoints > were inserted; 0 means they were not. */ > > extern int gdbarch_software_single_step_p (struct gdbarch *gdbarch); > diff --git a/gdb/gdbarch.sh b/gdb/gdbarch.sh > index dcf0343..bf78173 100755 > --- a/gdb/gdbarch.sh > +++ b/gdb/gdbarch.sh > @@ -362,6 +362,9 @@ v:int:long_bit:::8 * sizeof (long):4*TARGET_CHAR_BIT::0 > # Number of bits in a long long or unsigned long long for the target > # machine. > v:int:long_long_bit:::8 * sizeof (LONGEST):2*gdbarch->long_bit::0 > +# Alignment of a long long or unsigned long long for the target > +# machine. > +v:int:long_long_align_bit:::8 * sizeof (LONGEST):2*gdbarch->long_bit::0 > > # The ABI default bit-size and format for "half", "float", "double", and > # "long double". These bit/format pairs should eventually be combined > diff --git a/gdb/i386-tdep.c b/gdb/i386-tdep.c > index 179bc45..2f0b6f5 100644 > --- a/gdb/i386-tdep.c > +++ b/gdb/i386-tdep.c > @@ -7281,6 +7281,8 @@ i386_gdbarch_init (struct gdbarch_info info, struct gdbarch_list *arches) > > tdep->record_regmap = i386_record_regmap; > > + set_gdbarch_long_long_align_bit (gdbarch, 32); > + > /* The format used for `long double' on almost all i386 targets is > the i387 extended floating-point format. In fact, of all targets > in the GCC 2.95 tree, only OSF/1 does it different, and insists > diff --git a/gdb/jit.c b/gdb/jit.c > index eb1bcc7..283ccdb 100644 > --- a/gdb/jit.c > +++ b/gdb/jit.c > @@ -205,10 +205,11 @@ static void > jit_read_code_entry (struct gdbarch *gdbarch, > CORE_ADDR code_addr, struct jit_code_entry *code_entry) > { > - int err; > + int err, off; > struct type *ptr_type; > int ptr_size; > int entry_size; > + int align_bytes; > gdb_byte *entry_buf; > enum bfd_endian byte_order = gdbarch_byte_order (gdbarch); > > @@ -230,8 +231,13 @@ jit_read_code_entry (struct gdbarch *gdbarch, > extract_typed_address (&entry_buf[ptr_size], ptr_type); > code_entry->symfile_addr = > extract_typed_address (&entry_buf[2 * ptr_size], ptr_type); > + > + align_bytes = gdbarch_long_long_align_bit (gdbarch) / 8; > + off = 3 * ptr_size; > + off = (off + (align_bytes - 1)) & ~(align_bytes - 1); > + > code_entry->symfile_size = > - extract_unsigned_integer (&entry_buf[3 * ptr_size], 8, byte_order); > + extract_unsigned_integer (&entry_buf[off], 8, byte_order); > } > > /* This function registers code associated with a JIT code entry. It uses the >