[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZsMf1jIfbwwmqe1m@bogus>
Date: Mon, 19 Aug 2024 11:35:02 +0100
From: Sudeep Holla <sudeep.holla@....com>
To: Jinjie Ruan <ruanjinjie@...wei.com>
Cc: <linux-arm-kernel@...ts.infradead.org>, <linux-kernel@...r.kernel.org>,
Sudeep Holla <sudeep.holla@....com>
Subject: Re: [PATCH -next] firmware: arm_ffa: Fix beyond size of field warning
On Fri, Aug 16, 2024 at 06:02:58PM +0800, Jinjie Ruan wrote:
> An allmodconfig build of arm64 resulted in following warning:
>
> In function ‘fortify_memcpy_chk’,
> inlined from ‘export_uuid’ at ./include/linux/uuid.h:88:2,
> inlined from ‘ffa_msg_send_direct_req2’ at ./drivers/firmware/arm_ffa/driver.c:488:2:
> ./include/linux/fortify-string.h:571:25: error: call to ‘__write_overflow_field’ declared with attribute warning: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror=attribute-warning]
> 571 | __write_overflow_field(p_size_field, size);
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> In function ‘fortify_memcpy_chk’,
> inlined from ‘ffa_msg_send_direct_req2’ at ./drivers/firmware/arm_ffa/driver.c:489:2:
> ./linux-next/include/linux/fortify-string.h:571:25: error: call to ‘__write_overflow_field’ declared with attribute warning: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror=attribute-warning]
> 571 | __write_overflow_field(p_size_field, size);
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> Because ffa_msg_send_direct_req2() memcpy uuid_t and struct
> ffa_send_direct_data2 data to unsigned long dst, the copy size is 2 or
> or 14 unsigned long which beyond size of dst size, fix it by using a temp
> array for memcpy.
>
> Signed-off-by: Jinjie Ruan <ruanjinjie@...wei.com>
> ---
> drivers/firmware/arm_ffa/driver.c | 14 ++++++++++++--
> 1 file changed, 12 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/firmware/arm_ffa/driver.c b/drivers/firmware/arm_ffa/driver.c
> index 1e3764852118..674fbe008ea6 100644
> --- a/drivers/firmware/arm_ffa/driver.c
> +++ b/drivers/firmware/arm_ffa/driver.c
> @@ -480,13 +480,23 @@ static int ffa_msg_send2(u16 src_id, u16 dst_id, void *buf, size_t sz)
> static int ffa_msg_send_direct_req2(u16 src_id, u16 dst_id, const uuid_t *uuid,
> struct ffa_send_direct_data2 *data)
> {
> + unsigned long args_data[14];
> + unsigned long args_uuid[2];
> + unsigned long *data_ptr;
> +
> u32 src_dst_ids = PACK_TARGET_INFO(src_id, dst_id);
> ffa_value_t ret, args = {
> .a0 = FFA_MSG_SEND_DIRECT_REQ2, .a1 = src_dst_ids,
> };
>
> - export_uuid((u8 *)&args.a2, uuid);
> - memcpy(&args.a4, data, sizeof(*data));
> + memcpy(args_uuid, uuid, sizeof(uuid_t));
> + args.a2 = args_uuid[0];
> + args.a3 = args_uuid[1];
> +
> + memcpy(args_data, data, sizeof(*data));
> + data_ptr = &args.a4;
> + for (int i = 0; i < 14; i++)
> + *data_ptr++ = args_data[i];
>
So we end up with double copy for both uuid and ffa_send_direct_data2 ?
This is not correct and not needed.
Which toolchain are you using ? I got error only for memcpy which I forgot
to push to -next, now fixed. It must appear in -next soon.
--
Regards,
Sudeep
Powered by blists - more mailing lists