Currently kernel command line 'stack_guard_gap=', which does not provide an explicit pages gap value, still assigns 0 to 'stack_guard_gap', which would not have been expected. In such cases it should just retain the default gap value (DEFAULT_STACK_GUARD_GAP). Instead let's assert a positive value for input gap pages before proceeding any further. While here, this tightens up cmdline_parse_stack_guard_gap() for other scenarios, where the command line parameter 'stack_guard_gap' is not successfully handled as expected. Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> Cc: linux-kernel@xxxxxxxxxxxxxxx Cc: linux-mm@xxxxxxxxx Signed-off-by: Anshuman Khandual <anshuman.khandual@xxxxxxx> --- Depends on the following patch. https://lore.kernel.org/all/20230828035248.678960-1-anshuman.khandual@xxxxxxx/ mm/mmap.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) diff --git a/mm/mmap.c b/mm/mmap.c index 8679750333bb..adaa81d95518 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -2122,16 +2122,19 @@ int expand_downwards(struct vm_area_struct *vma, unsigned long address) /* enforced gap between the expanding stack and other mappings. */ unsigned long stack_guard_gap = DEFAULT_STACK_GUARD_GAP; -static int __init cmdline_parse_stack_guard_gap(char *p) +static int __init cmdline_parse_stack_guard_gap(char *str) { unsigned long val; - char *endptr; - val = simple_strtoul(p, &endptr, 10); - if (!*endptr) - stack_guard_gap = val << PAGE_SHIFT; + if (!str) + return 0; - return 1; + val = simple_strtoul(str, &str, 10); + if (!*str && val) { + stack_guard_gap = val << PAGE_SHIFT; + return 1; + } + return 0; } __setup("stack_guard_gap=", cmdline_parse_stack_guard_gap); -- 2.30.2
