Re: [patch V2 1/3] x86/ldt: Plug memory leak in error path

[Andy Lutomirski]

> On Dec 31, 2017, at 2:24 AM, Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
> 
> The error path in write_ldt() tries to free old_ldt instead of the newly
> allocated new_ldt resulting in a memory leak. It also misses to clean up a
> half populated LDT pagetable, which is not a leak as it gets cleaned up
> when the process exits.
> 
> Free both the potentially half populated LDT pagetable and the newly
> allocated LDT struct. This can be done unconditionally because once a LDT
> is mapped subsequent maps will succeed because the PTE page is already
> populated and the two LDTs fit into that single page.
> 
> Fixes: f55f0501cbf6 ("x86/pti: Put the LDT in its own PGD if PTI is on")
> Reported-by: Mathieu Desnoyers <mathieu.desnoyers@xxxxxxxxxxxx>
> Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> ---
> arch/x86/kernel/ldt.c |    8 +++++++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
> 
> --- a/arch/x86/kernel/ldt.c
> +++ b/arch/x86/kernel/ldt.c
> @@ -421,7 +421,13 @@ static int write_ldt(void __user *ptr, u
>     */
>    error = map_ldt_struct(mm, new_ldt, old_ldt ? !old_ldt->slot : 0);
>    if (error) {
> -        free_ldt_struct(old_ldt);
> +        /*
> +         * This only can fail for the first LDT setup. If a LDT is
> +         * already installed then the PTE page is already
> +         * populated. Mop up a half populated page table.
> +         */

I liked it better with the conditional.  If this ever fails due to fault injection, some silly accounting issue, a paravirt glitch, or whatever, then we'll oops.

> +        free_ldt_pgtables(mm);
> +        free_ldt_struct(new_ldt);
>        goto out_unlock;
>    }
>