[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMzpN2hf-CFpO6x58aDK_FX_6C2MBKh1g7PdV4Y=ypaeUNVfRw@mail.gmail.com>
Date: Sun, 29 Dec 2024 18:05:32 -0500
From: Brian Gerst <brgerst@...il.com>
To: Kaixiong Yu <yukaixiong@...wei.com>
Cc: akpm@...ux-foundation.org, mcgrof@...nel.org, ysato@...rs.sourceforge.jp,
dalias@...c.org, glaubitz@...sik.fu-berlin.de, luto@...nel.org,
tglx@...utronix.de, mingo@...hat.com, bp@...en8.de,
dave.hansen@...ux.intel.com, hpa@...or.com, viro@...iv.linux.org.uk,
brauner@...nel.org, jack@...e.cz, kees@...nel.org, j.granados@...sung.com,
willy@...radead.org, Liam.Howlett@...cle.com, vbabka@...e.cz,
lorenzo.stoakes@...cle.com, trondmy@...nel.org, anna@...nel.org,
chuck.lever@...cle.com, jlayton@...nel.org, neilb@...e.de,
okorniev@...hat.com, Dai.Ngo@...cle.com, tom@...pey.com, davem@...emloft.net,
edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com, paul@...l-moore.com,
jmorris@...ei.org, linux-sh@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-fsdevel@...r.kernel.org, linux-mm@...ck.org, linux-nfs@...r.kernel.org,
netdev@...r.kernel.org, linux-security-module@...r.kernel.org,
dhowells@...hat.com, haifeng.xu@...pee.com, baolin.wang@...ux.alibaba.com,
shikemeng@...weicloud.com, dchinner@...hat.com, bfoster@...hat.com,
souravpanda@...gle.com, hannes@...xchg.org, rientjes@...gle.com,
pasha.tatashin@...een.com, david@...hat.com, ryan.roberts@....com,
ying.huang@...el.com, yang@...amperecomputing.com, zev@...ilderbeest.net,
serge@...lyn.com, vegard.nossum@...cle.com, wangkefeng.wang@...wei.com
Subject: Re: [PATCH v4 -next 13/15] x86: vdso: move the sysctl to arch/x86/entry/vdso/vdso32-setup.c
On Sat, Dec 28, 2024 at 10:17 AM Kaixiong Yu <yukaixiong@...wei.com> wrote:
>
> When CONFIG_X86_32 is defined and CONFIG_UML is not defined,
> vdso_enabled belongs to arch/x86/entry/vdso/vdso32-setup.c.
> So, move it into its own file.
>
> Before this patch, vdso_enabled was allowed to be set to
> a value exceeding 1 on x86_32 architecture. After this patch is
> applied, vdso_enabled is not permitted to set the value more than 1.
> It does not matter, because according to the function load_vdso32(),
> only vdso_enabled is set to 1, VDSO would be enabled. Other values
> all mean "disabled". The same limitation could be seen in the
> function vdso32_setup().
>
> Signed-off-by: Kaixiong Yu <yukaixiong@...wei.com>
> Reviewed-by: Kees Cook <kees@...nel.org>
> ---
> v4:
> - const qualify struct ctl_table vdso_table
> ---
> ---
> arch/x86/entry/vdso/vdso32-setup.c | 16 +++++++++++-----
> kernel/sysctl.c | 8 +-------
> 2 files changed, 12 insertions(+), 12 deletions(-)
>
> diff --git a/arch/x86/entry/vdso/vdso32-setup.c b/arch/x86/entry/vdso/vdso32-setup.c
> index 76e4e74f35b5..f71625f99bf9 100644
> --- a/arch/x86/entry/vdso/vdso32-setup.c
> +++ b/arch/x86/entry/vdso/vdso32-setup.c
> @@ -51,15 +51,17 @@ __setup("vdso32=", vdso32_setup);
> __setup_param("vdso=", vdso_setup, vdso32_setup, 0);
> #endif
>
> -#ifdef CONFIG_X86_64
>
> #ifdef CONFIG_SYSCTL
> -/* Register vsyscall32 into the ABI table */
> #include <linux/sysctl.h>
>
> -static struct ctl_table abi_table2[] = {
> +static const struct ctl_table vdso_table[] = {
> {
> +#ifdef CONFIG_X86_64
> .procname = "vsyscall32",
> +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML))
vdso32-setup,.c is not used when building UML, so this can be reduced
to "#else".
> + .procname = "vdso_enabled",
> +#endif
> .data = &vdso32_enabled,
> .maxlen = sizeof(int),
> .mode = 0644,
> @@ -71,10 +73,14 @@ static struct ctl_table abi_table2[] = {
>
> static __init int ia32_binfmt_init(void)
> {
> - register_sysctl("abi", abi_table2);
> +#ifdef CONFIG_X86_64
> + /* Register vsyscall32 into the ABI table */
> + register_sysctl("abi", vdso_table);
> +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML))
Same as above.
> + register_sysctl_init("vm", vdso_table);
> +#endif
> return 0;
> }
> __initcall(ia32_binfmt_init);
> #endif /* CONFIG_SYSCTL */
>
> -#endif /* CONFIG_X86_64 */
Brian Gerst
Powered by blists - more mailing lists