[PATCH] mem-leak fix for acpi_thermal_write_trip_points()

From: Jesper Juhl
Date: Sat May 13 2006 - 20:27:58 EST


There's a memory leak in
drivers/acpi/thermal.c::acpi_thermal_write_trip_points()
found by the Coverity checker as bug #1243

The problem is simply that if the allocation of 'active' fails, then we'll
return from the function without freeing the memory previously allocated
to 'limit_string'.

One fix would have been the insertion of kfree(limit_string); just before
the return_VALUE(-ENOMEM); call, but I chose a different fix.
Except for this one return the function only has a single exit point at
the end: label, and there it does proper cleanup of everything. So even
though setting the return value (the 'count' variable) and jumping there
results in a pointless kfree(NULL) call (since 'active' will be NULL), I
thought this was a small price to pay for the bennefit of having just a
single exit path for the function.


Signed-off-by: Jesper Juhl <jesper.juhl@xxxxxxxxx>
---

drivers/acpi/thermal.c | 6 ++++--
1 files changed, 4 insertions(+), 2 deletions(-)


--- linux-2.6.17-rc4-git2-orig/drivers/acpi/thermal.c 2006-03-20 06:53:29.000000000 +0100
+++ linux-2.6.17-rc4-git2/drivers/acpi/thermal.c 2006-05-14 02:18:30.000000000 +0200
@@ -942,8 +942,10 @@ acpi_thermal_write_trip_points(struct fi
memset(limit_string, 0, ACPI_THERMAL_MAX_LIMIT_STR_LEN);

active = kmalloc(ACPI_THERMAL_MAX_ACTIVE * sizeof(int), GFP_KERNEL);
- if (!active)
- return_VALUE(-ENOMEM);
+ if (!active) {
+ count = -ENOMEM;
+ goto end;
+ }

if (!tz || (count > ACPI_THERMAL_MAX_LIMIT_STR_LEN - 1)) {
ACPI_DEBUG_PRINT((ACPI_DB_ERROR, "Invalid argument\n"));



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/