When update kernel with the latest mips-next, we can not login through a
graphical interface, this is because drm radeon GPU driver does not work,
we can not see the boot message "[drm] radeon kernel modesetting enabled."
through the serial console.

drivers/gpu/drm/radeon/radeon_drv.c
static int __init radeon_module_init(void)
{
        [...]
        DRM_INFO("radeon kernel modesetting enabled.\n");
        [...]
}

I use git bisect to find out the commit 04324f44cb69 ("MIPS: Remove
get_fs/set_fs") is the first bad commit.

I analysis and test the changes in the above first bad commit and then
find out the following obvious difference which leads to the login issue.

arch/mips/include/asm/uaccess.h
static inline long strnlen_user(const char __user *s, long n)
{
        [...]
        if (!access_ok(s, n))
                return -0;
        [...]
}

I use dump_stack() to find out the following call trace:
load_module()
        strndup_user()
                strnlen_user()

load_module() failed in the following error path, we can see that the
second argument of strndup_user() is very big.

static int load_module(struct load_info *info, const char __user *uargs,
                       int flags)
{
        [...]
        mod->args = strndup_user(uargs, ~0UL >> 1);
        if (IS_ERR(mod->args)) {
                err = PTR_ERR(mod->args);
                goto free_arch_cleanup;
        }
        [...]
}

As discussed earlier [1], it seems that just modify the exception check
condition in strnlen_user() to fix load_module() failure, like this:

diff --git a/arch/mips/include/asm/uaccess.h b/arch/mips/include/asm/uaccess.h
index 9e9e3aa..f4fb8ca 100644
--- a/arch/mips/include/asm/uaccess.h
+++ b/arch/mips/include/asm/uaccess.h
@@ -612,8 +612,8 @@ static inline long strnlen_user(const char __user *s, long 
n)
 {
        long res;

-       if (!access_ok(s, n))
-               return -0;
+       if (!access_ok(s, 1))
+               return 0;

        might_fault();
        __asm__ __volatile__(

At the other hand, I search strndup_user() in the kernel tree, the second
argument of them are almost a macro or a fixed value which is relatively
small, such as PAGE_SIZE, PATH_MAX. So I think maybe we can use ARG_MAX as
second argument of strndup_user() in load_module().

With this patch, the load_module() failure disappered and we can login
normally through a graphical interface.

[1] https://lore.kernel.org/patchwork/patch/1411214/

Signed-off-by: Tiezhu Yang <yangtie...@loongson.cn>
---
 kernel/module.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/module.c b/kernel/module.c
index 3047935..30d320b 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -3998,7 +3998,7 @@ static int load_module(struct load_info *info, const char 
__user *uargs,
        flush_module_icache(mod);
 
        /* Now copy in args */
-       mod->args = strndup_user(uargs, ~0UL >> 1);
+       mod->args = strndup_user(uargs, ARG_MAX);
        if (IS_ERR(mod->args)) {
                err = PTR_ERR(mod->args);
                goto free_arch_cleanup;
-- 
2.1.0

Reply via email to