From: Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>
To: libc-alpha@sourceware.org
Subject: Re: [PATCH v6] ifaddrs: Get rid of alloca
Date: Thu, 22 Jun 2023 10:47:58 -0300 [thread overview]
Message-ID: <ccd8858c-985a-f7ce-6de9-310526a6803b@linaro.org> (raw)
In-Reply-To: <20230621200053.3021604-1-josimmon@redhat.com>
On 21/06/23 17:00, Joe Simmons-Talbott via Libc-alpha wrote:
> Use scratch_buffer and malloc rather than alloca to avoid potential stack
> overflows.
LGTM, thanks.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
> ---
> Changes to v5:
> * Don't bypass the __libc_scratch_buffer_set_array_size check by
> passing both the item size and count to scratch_buffer_set_array_size.
>
> sysdeps/unix/sysv/linux/ifaddrs.c | 46 ++++++++++++++-----------------
> 1 file changed, 20 insertions(+), 26 deletions(-)
>
> diff --git a/sysdeps/unix/sysv/linux/ifaddrs.c b/sysdeps/unix/sysv/linux/ifaddrs.c
> index 184ee224cb..0db9bb7847 100644
> --- a/sysdeps/unix/sysv/linux/ifaddrs.c
> +++ b/sysdeps/unix/sysv/linux/ifaddrs.c
> @@ -16,13 +16,13 @@
> License along with the GNU C Library; if not, see
> <https://www.gnu.org/licenses/>. */
>
> -#include <alloca.h>
> #include <assert.h>
> #include <errno.h>
> #include <ifaddrs.h>
> #include <net/if.h>
> #include <netinet/in.h>
> #include <netpacket/packet.h>
> +#include <scratch_buffer.h>
> #include <stdbool.h>
> #include <stdint.h>
> #include <stdlib.h>
> @@ -131,26 +131,14 @@ __netlink_request (struct netlink_handle *h, int type)
> ssize_t read_len;
> bool done = false;
>
> -#ifdef PAGE_SIZE
> - /* Help the compiler optimize out the malloc call if PAGE_SIZE
> - is constant and smaller or equal to PTHREAD_STACK_MIN/4. */
> - const size_t buf_size = PAGE_SIZE;
> -#else
> - const size_t buf_size = __getpagesize ();
> -#endif
> - bool use_malloc = false;
> - char *buf;
> -
> - if (__libc_use_alloca (buf_size))
> - buf = alloca (buf_size);
> - else
> - {
> - buf = malloc (buf_size);
> - if (buf != NULL)
> - use_malloc = true;
> - else
> - goto out_fail;
> - }
> + /* Netlink requires that user buffer needs to be either 8kb or page size
> + (whichever is bigger), however this has been changed over time and now
> + 8Kb is sufficient (check NLMSG_DEFAULT_SIZE on Linux
> + linux/include/linux/netlink.h). */
> + const size_t buf_size = 8192;
> + char *buf = malloc (buf_size);
> + if (buf == NULL)
> + goto out_fail;
>
> struct iovec iov = { buf, buf_size };
>
> @@ -229,13 +217,11 @@ __netlink_request (struct netlink_handle *h, int type)
> h->end_ptr = nlm_next;
> }
>
> - if (use_malloc)
> - free (buf);
> + free(buf);
> return 0;
>
> out_fail:
> - if (use_malloc)
> - free (buf);
> + free(buf);
> return -1;
> }
>
> @@ -324,6 +310,8 @@ getifaddrs_internal (struct ifaddrs **ifap)
> char *ifa_data_ptr; /* Pointer to the unused part of memory for
> ifa_data. */
> int result = 0;
> + struct scratch_buffer buf;
> + scratch_buffer_init (&buf);
>
> *ifap = NULL;
>
> @@ -425,7 +413,12 @@ getifaddrs_internal (struct ifaddrs **ifap)
> }
>
> /* Table for mapping kernel index to entry in our list. */
> - map_newlink_data = alloca (newlink * sizeof (int));
> + if (!scratch_buffer_set_array_size (&buf, newlink, sizeof (int)))
> + {
> + result = -1;
> + goto exit_free;
> + }
> + map_newlink_data = buf.data;
> memset (map_newlink_data, '\xff', newlink * sizeof (int));
>
> ifa_data_ptr = (char *) &ifas[newlink + newaddr];
> @@ -820,6 +813,7 @@ getifaddrs_internal (struct ifaddrs **ifap)
> exit_free:
> __netlink_free_handle (&nh);
> __netlink_close (&nh);
> + scratch_buffer_free (&buf);
>
> return result;
> }
next prev parent reply other threads:[~2023-06-22 13:48 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-06-21 20:00 Joe Simmons-Talbott
2023-06-22 13:47 ` Adhemerval Zanella Netto [this message]
2023-06-26 13:10 ` Siddhesh Poyarekar
2023-06-26 19:49 ` Adhemerval Zanella Netto
2023-06-26 21:04 ` Joe Simmons-Talbott
2023-06-27 11:50 ` Siddhesh Poyarekar
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ccd8858c-985a-f7ce-6de9-310526a6803b@linaro.org \
--to=adhemerval.zanella@linaro.org \
--cc=libc-alpha@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).