From: Dave Brolley <brolley@redhat.com>
To: Kevin Buettner <kevinb@redhat.com>
Cc: gdb-patches@sources.redhat.com
Subject: Re: [RFA] FR-V SIM: Check availability of GR and FR regs
Date: Fri, 07 Nov 2003 17:30:00 -0000 [thread overview]
Message-ID: <3FABD6AC.5000806@redhat.com> (raw)
In-Reply-To: <1031107063317.ZM15719@localhost.localdomain>
I'm not sure if it's ok to include gdb/sim-frv.h from within the
simulator code. I thought the interface went the other way around (i.e.
gdb includes sim/frv-sim.h).
Other than that, it looks ok to me.
Kevin Buettner wrote:
>The patch below adds tests for checking the availability of the GR and
>FR registers in frvbf_fetch_register() and frvbf_store_register().
>
>This patch fixes a gdb segfault that I was seeing when running the
>simulator against a program built with -mcpu=fr400. The fr400 series
>only has 32 GRs and 32 FRs. This was causing a segfault during
>simulator initialization that occurred as a result of the "target sim"
>command.
>
>The reason for this segfault was much the same as that for the SPR
>related segfault that you committed a patch for last week - basically,
>if the existence of the register is not checked for ahead of time, the
>simulator attempts to generate an exception for those registers which
>are unavailable. We don't want to generate an "unavailable register"
>exception (or whatever it's really called) when attempting to
>fetch/store registers from GDB.
>
>I also took the opportunity to revise frvbf_{fetch,store}_register() to
>use the register constants defined in include/gdb/sim-frv.h.
>
>Okay?
>
> * frv-sim.h (GR_REGNUM_MAX, FR_REGNUM_MAX, PC_REGNUM, SPR_REGNUM_MIN)
> (SPR_REGNUM_MAX): Delete.
> * frv.c (gdb/sim-frv.h): Include.
> (frvbf_fetch_register, frvbf_store_register): Use register number
> constants from gdb/sim-frv.h. Check availability of general
> purpose and float registers.
>
>Index: frv/frv-sim.h
>===================================================================
>RCS file: /cvs/src/src/sim/frv/frv-sim.h,v
>retrieving revision 1.4
>diff -u -p -r1.4 frv-sim.h
>--- frv/frv-sim.h 31 Oct 2003 18:23:47 -0000 1.4
>+++ frv/frv-sim.h 7 Nov 2003 06:08:39 -0000
>@@ -29,13 +29,6 @@ with this program; if not, write to the
> #define H_SPR_ACCG4 1476
> #define H_SPR_ACCG63 1535
>
>-/* gdb register numbers. */
>-#define GR_REGNUM_MAX 63
>-#define FR_REGNUM_MAX 127
>-#define PC_REGNUM 128
>-#define SPR_REGNUM_MIN 129
>-#define SPR_REGNUM_MAX (SPR_REGNUM_MIN + 4096 - 1)
>-
> /* Initialization of the frv cpu. */
> void frv_initialize (SIM_CPU *, SIM_DESC);
> void frv_term (SIM_DESC);
>Index: frv/frv.c
>===================================================================
>RCS file: /cvs/src/src/sim/frv/frv.c,v
>retrieving revision 1.4
>diff -u -p -r1.4 frv.c
>--- frv/frv.c 31 Oct 2003 18:23:47 -0000 1.4
>+++ frv/frv.c 7 Nov 2003 06:08:39 -0000
>@@ -27,6 +27,7 @@ with this program; if not, write to the
> #include "cgen-engine.h"
> #include "cgen-par.h"
> #include "bfd.h"
>+#include "gdb/sim-frv.h"
> #include <math.h>
>
> /* Maintain a flag in order to know when to write the address of the next
>@@ -38,17 +39,37 @@ int frvbf_write_next_vliw_addr_to_LR;
> int
> frvbf_fetch_register (SIM_CPU *current_cpu, int rn, unsigned char *buf, int len)
> {
>- if (rn <= GR_REGNUM_MAX)
>- SETTSI (buf, GET_H_GR (rn));
>- else if (rn <= FR_REGNUM_MAX)
>- SETTSI (buf, GET_H_FR (rn - GR_REGNUM_MAX - 1));
>- else if (rn == PC_REGNUM)
>+ if (SIM_FRV_GR0_REGNUM <= rn && rn <= SIM_FRV_GR63_REGNUM)
>+ {
>+ int hi_available, lo_available;
>+ int grn = rn - SIM_FRV_GR0_REGNUM;
>+
>+ frv_gr_registers_available (current_cpu, &hi_available, &lo_available);
>+
>+ if ((grn < 32 && !lo_available) || (grn >= 32 && !hi_available))
>+ return 0;
>+ else
>+ SETTSI (buf, GET_H_GR (grn));
>+ }
>+ else if (SIM_FRV_FR0_REGNUM <= rn && rn <= SIM_FRV_FR63_REGNUM)
>+ {
>+ int hi_available, lo_available;
>+ int frn = rn - SIM_FRV_FR0_REGNUM;
>+
>+ frv_fr_registers_available (current_cpu, &hi_available, &lo_available);
>+
>+ if ((frn < 32 && !lo_available) || (frn >= 32 && !hi_available))
>+ return 0;
>+ else
>+ SETTSI (buf, GET_H_FR (frn));
>+ }
>+ else if (rn == SIM_FRV_PC_REGNUM)
> SETTSI (buf, GET_H_PC ());
>- else if (rn >= SPR_REGNUM_MIN && rn <= SPR_REGNUM_MAX)
>+ else if (SIM_FRV_SPR0_REGNUM <= rn && rn <= SIM_FRV_SPR4095_REGNUM)
> {
> /* Make sure the register is implemented. */
> FRV_REGISTER_CONTROL *control = CPU_REGISTER_CONTROL (current_cpu);
>- int spr = rn - SPR_REGNUM_MIN;
>+ int spr = rn - SIM_FRV_SPR0_REGNUM;
> if (! control->spr[spr].implemented)
> return 0;
> SETTSI (buf, GET_H_SPR (spr));
>@@ -67,17 +88,37 @@ frvbf_fetch_register (SIM_CPU *current_c
> int
> frvbf_store_register (SIM_CPU *current_cpu, int rn, unsigned char *buf, int len)
> {
>- if (rn <= GR_REGNUM_MAX)
>- SET_H_GR (rn, GETTSI (buf));
>- else if (rn <= FR_REGNUM_MAX)
>- SET_H_FR (rn - GR_REGNUM_MAX - 1, GETTSI (buf));
>- else if (rn == PC_REGNUM)
>+ if (SIM_FRV_GR0_REGNUM <= rn && rn <= SIM_FRV_GR63_REGNUM)
>+ {
>+ int hi_available, lo_available;
>+ int grn = rn - SIM_FRV_GR0_REGNUM;
>+
>+ frv_gr_registers_available (current_cpu, &hi_available, &lo_available);
>+
>+ if ((grn < 32 && !lo_available) || (grn >= 32 && !hi_available))
>+ return 0;
>+ else
>+ SET_H_GR (grn, GETTSI (buf));
>+ }
>+ else if (SIM_FRV_FR0_REGNUM <= rn && rn <= SIM_FRV_FR63_REGNUM)
>+ {
>+ int hi_available, lo_available;
>+ int frn = rn - SIM_FRV_FR0_REGNUM;
>+
>+ frv_fr_registers_available (current_cpu, &hi_available, &lo_available);
>+
>+ if ((frn < 32 && !lo_available) || (frn >= 32 && !hi_available))
>+ return 0;
>+ else
>+ SET_H_FR (frn, GETTSI (buf));
>+ }
>+ else if (rn == SIM_FRV_PC_REGNUM)
> SET_H_PC (GETTSI (buf));
>- else if (rn >= SPR_REGNUM_MIN && rn <= SPR_REGNUM_MAX)
>+ else if (SIM_FRV_SPR0_REGNUM <= rn && rn <= SIM_FRV_SPR4095_REGNUM)
> {
> /* Make sure the register is implemented. */
> FRV_REGISTER_CONTROL *control = CPU_REGISTER_CONTROL (current_cpu);
>- int spr = rn - SPR_REGNUM_MIN;
>+ int spr = rn - SIM_FRV_SPR0_REGNUM;
> if (! control->spr[spr].implemented)
> return 0;
> SET_H_SPR (spr, GETTSI (buf));
>
>
next prev parent reply other threads:[~2003-11-07 17:30 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2003-11-07 6:33 Kevin Buettner
2003-11-07 17:30 ` Dave Brolley [this message]
2003-11-07 17:49 ` Kevin Buettner
2003-11-07 19:12 ` Michael Snyder
2003-11-24 16:47 ` Kevin Buettner
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=3FABD6AC.5000806@redhat.com \
--to=brolley@redhat.com \
--cc=gdb-patches@sources.redhat.com \
--cc=kevinb@redhat.com \
/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