Skip to contentSkip to conversation

Kconfig: conf segfault (with invalid kconfig contents)

DISCUSSION Olof Johansson 3 participants 4 messages Oct 24, 2007 – Dec 27, 2008 Kconfig

lore
Share & tools
lore
#0 lore raw

Hi,

Sorry, no patch, just a bugreport. :)

I did a mistake in some Kconfig changes (locally) and got a segfault from conf:

#0  0x10005fe0 in sym_check_deps ()
#1  0x10005f40 in sym_check_expr_deps ()
#2  0x1000605c in sym_check_deps ()
#3  0x1000f3f4 in conf_parse ()
#4  0x10003824 in main ()

The way to reproduce is simple. I just added the equivalent of this to arch/powerpc/Kconfig:

---
config TEST1

        bool
        depends on TEST2 && PCI
        select TEST2

config TEST2
bool
---

Neither TEST1 nor TEST2 exists in the existing .config when I run make oldconfig to reproduce.

It's clearly invalid in that it depends on what it selects, but it should just abort instead.

-Olof

#1 lore raw

Hi Olof.

Quoted 27 lines · Olof Johansson

Hi,

Sorry, no patch, just a bugreport. :)

I did a mistake in some Kconfig changes (locally) and got a segfault from conf:

#0  0x10005fe0 in sym_check_deps ()
#1  0x10005f40 in sym_check_expr_deps ()
#2  0x1000605c in sym_check_deps ()
#3  0x1000f3f4 in conf_parse ()
#4  0x10003824 in main ()

The way to reproduce is simple. I just added the equivalent of this to arch/powerpc/Kconfig:

---
config TEST1

        bool
        depends on TEST2 && PCI
        select TEST2

config TEST2
bool
---

Quoted 5 lines

Neither TEST1 nor TEST2 exists in the existing .config when I run make oldconfig to reproduce.

It's clearly invalid in that it depends on what it selects, but it should just abort instead.

Thanks for the report and especially for the testcase!
I will try to look at it a bit later if noone bites me (I'm afraid not).

Sam

#2 lore raw

Hi,

Quoted 5 lines · Sam Ravnborg
Quoted 2 lines

It's clearly invalid in that it depends on what it selects, but it should just abort instead.

Thanks for the report and especially for the testcase!
I will try to look at it a bit later if noone bites me (I'm afraid not).

Well, you're also responsible for it. :)
http://lkml.org/lkml/2007/5/6/14

It actually finds the recursive dependency, but it crashes because sym->prop is NULL.

bye, Roman

#3 lore raw
In reply to Olof Johansson
Quoted 27 lines · Olof Johansson

Hi,

Sorry, no patch, just a bugreport. :)

I did a mistake in some Kconfig changes (locally) and got a segfault from conf:

#0  0x10005fe0 in sym_check_deps ()
#1  0x10005f40 in sym_check_expr_deps ()
#2  0x1000605c in sym_check_deps ()
#3  0x1000f3f4 in conf_parse ()
#4  0x10003824 in main ()

The way to reproduce is simple. I just added the equivalent of this to arch/powerpc/Kconfig:

---
config TEST1

        bool
        depends on TEST2 && PCI
        select TEST2

config TEST2
bool
---

Quoted 5 lines

Neither TEST1 nor TEST2 exists in the existing .config when I run make oldconfig to reproduce.

It's clearly invalid in that it depends on what it selects, but it should just abort instead.

Hi Olof.

BAsed on your nice small testcase I have finally looked at it and fixed the bug. I forgot to credit you in the patch I sent to lkml (and linux-kbuild) but I have done so in my local repository.

See updated commit below.

Sam

[PATCH 1/3] kconfig: improve readout when we hit recursive dependencies

Following sample:

    config TEST1
            bool
            depends on TEST2 && PCI
            select TEST2

    config TEST2
            bool

Will result in the following output:

    Kconfig:1:error: found recursive dependency:
                TEST1
             -> TEST2 (arch/x86/Kconfig:6)
             -> TEST1 (arch/x86/Kconfig:1)

The sample above gave a segmentation fault
which is also fixed by this patch.

To save the filename/lineno info add a new property P_SYMBOL which is added each time we see a symbol definition.

Thanks to Olof for reporting and providing the sample Kconfig.

---
 scripts/kconfig/expr.h   |    1 +
 scripts/kconfig/menu.c   |    2 ++
 scripts/kconfig/symbol.c |   11 ++++++++---
 3 files changed, 11 insertions(+), 3 deletions(-)
scripts/kconfig/expr.h +1 -0
index 6408fef..d587895 100644
@@ -132,6 +132,7 @@ enum prop_type {
 	P_SELECT,   /* select BAR */
 	P_RANGE,    /* range 7..100 (for a symbol) */
 	P_ENV,      /* value from environment variable */
+	P_SYMBOL,   /* where a symbol is defined */
 };
 
 struct property {
scripts/kconfig/menu.c +2 -0
index 07ff8d1..aa0e2f2 100644
@@ -55,6 +55,8 @@ void menu_add_entry(struct symbol *sym)
 	*last_entry_ptr = menu;
 	last_entry_ptr = &menu->next;
 	current_entry = menu;
+	if (sym)
+		 menu_add_prop(P_SYMBOL, NULL, NULL, NULL);
 }
 
 void menu_end_entry(void)
scripts/kconfig/symbol.c +8 -3
index 18f3e5c..474dae2 100644
@@ -856,8 +856,9 @@ struct symbol *sym_check_deps(struct symbol *sym)
 	struct property *prop;
 
 	if (sym->flags & SYMBOL_CHECK) {
-		fprintf(stderr, "%s:%d:error: found recursive dependency: %s",
-		        sym->prop->file->name, sym->prop->lineno,
+		fprintf(stderr, "%s:%d:error: found recursive dependency:\n",
+		        sym->prop->file->name, sym->prop->lineno);
+		fprintf(stderr, "\t    %s\n",
 			sym->name ? sym->name : "<choice>");
 		return sym;
 	}
@@ -877,7 +878,9 @@ struct symbol *sym_check_deps(struct symbol *sym)
 	}
 
 	if (sym2) {
-		fprintf(stderr, " -> %s", sym->name ? sym->name : "<choice>");
+		fprintf(stderr, "\t -> %s (%s:%d)\n",
+		        sym->name ? sym->name : "<choice>",
+			sym->prop->file->name, sym->prop->lineno);
 		if (sym2 == sym) {
 			fprintf(stderr, "\n");
 			zconfnerrs++;
@@ -937,6 +940,8 @@ const char *prop_get_type_name(enum prop_type type)
 		return "select";
 	case P_RANGE:
 		return "range";
+	case P_SYMBOL:
+		return "symbol";
 	case P_UNKNOWN:
 		break;
 	}
Signature · 1 line
1.6.0.2.GIT