lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CA+_ehUwbUYc6GBGx1VFfqknWrFTk=Xm7NmMUhbyxMSBE4KdnjQ@mail.gmail.com>
Date:   Tue, 16 Nov 2021 14:30:59 +0100
From:   Ansuel Smith <ansuelsmth@...il.com>
To:     Dan Carpenter <dan.carpenter@...cle.com>
Cc:     kbuild@...ts.01.org, kernel test robot <lkp@...el.com>,
        kbuild-all@...ts.01.org, open list <linux-kernel@...r.kernel.org>
Subject: Re: drivers/net/dsa/qca8k.c:944 qca8k_parse_port_config() error:
 testing array offset 'cpu_port_index' after use.

> On Mon, Nov 15, 2021 at 07:08:30PM +0100, Ansuel Smith wrote:
> > > tree:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git  master
> > > head:   debe436e77c72fcee804fb867f275e6d31aa999c
> > > commit: 5654ec78dd7e64b1e04777b24007344329e6a63b net: dsa: qca8k: rework rgmii delay logic and scan for cpu port 6
> > > config: i386-randconfig-m021-20211025 (attached as .config)
> > > compiler: gcc-9 (Debian 9.3.0-22) 9.3.0
> > >
> > > If you fix the issue, kindly add following tag as appropriate
> > > Reported-by: kernel test robot <lkp@...el.com>
> > > Reported-by: Dan Carpenter <dan.carpenter@...cle.com>
> > >
> >
> > This should already be fixed by 06dd34a628ae5b6a839b757e746de165d6789ca8
> > Can you confirm this?
> >
>
> No, it doesn't fix the problem.  The check is either useless and should
> be removed or there is an out of bounds bug.  Checking for an out of
> bounds *after* you've already written to the memory is *never* useful.
>
> regards,
> dan carpenter
>

Again sorry if I insist and I'm 99% missing something.
There was an out of bounds bug.
It was fixed by 06dd34a628ae5b6a839b757e746de165d6789ca8 by
starting the counter to -1.
The extra check is useless, yes.

> > > smatch warnings:
> > > drivers/net/dsa/qca8k.c:944 qca8k_parse_port_config() error: testing array offset 'cpu_port_index' after use.
> > >
> > > vim +/cpu_port_index +944 drivers/net/dsa/qca8k.c
> > >
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   934  static int
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   935  qca8k_parse_port_config(struct qca8k_priv *priv)
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   936  {
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   937    int port, cpu_port_index = 0, ret;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   938    struct device_node *port_dn;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   939    phy_interface_t mode;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   940    struct dsa_port *dp;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   941    u32 delay;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   942
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   943    /* We have 2 CPU port. Check them */
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14  @944    for (port = 0; port < QCA8K_NUM_PORTS && cpu_port_index < QCA8K_NUM_CPU_PORTS; port++) {
> > >                                                                                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > Assume cpu_port_index = QCA8K_NUM_CPU_PORTS - 1;
> > >
> > >
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   945            /* Skip every other port */
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   946            if (port != 0 && port != 6)
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   947                    continue;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   948
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   949            dp = dsa_to_port(priv->ds, port);
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   950            port_dn = dp->dn;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   951            cpu_port_index++;
> > >                                                         ^^^^^^^^^^^^^^^^^
> > > cpu_port_index is now out of bounds.
> > >
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   952
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   953            if (!of_device_is_available(port_dn))
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   954                    continue;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   955
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   956            ret = of_get_phy_mode(port_dn, &mode);
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   957            if (ret)
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   958                    continue;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   959
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   960            switch (mode) {
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   961            case PHY_INTERFACE_MODE_RGMII:
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   962            case PHY_INTERFACE_MODE_RGMII_ID:
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   963            case PHY_INTERFACE_MODE_RGMII_TXID:
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   964            case PHY_INTERFACE_MODE_RGMII_RXID:
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   965                    delay = 0;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   966
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   967                    if (!of_property_read_u32(port_dn, "tx-internal-delay-ps", &delay))
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   968                            /* Switch regs accept value in ns, convert ps to ns */
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   969                            delay = delay / 1000;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   970                    else if (mode == PHY_INTERFACE_MODE_RGMII_ID ||
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   971                             mode == PHY_INTERFACE_MODE_RGMII_TXID)
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   972                            delay = 1;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   973
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   974                    if (delay > QCA8K_MAX_DELAY) {
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   975                            dev_err(priv->dev, "rgmii tx delay is limited to a max value of 3ns, setting to the max value");
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   976                            delay = 3;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   977                    }
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   978
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   979                    priv->rgmii_tx_delay[cpu_port_index] = delay;
> > >                                                                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > Out of bounds
> > >
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   980
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   981                    delay = 0;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   982
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   983                    if (!of_property_read_u32(port_dn, "rx-internal-delay-ps", &delay))
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   984                            /* Switch regs accept value in ns, convert ps to ns */
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   985                            delay = delay / 1000;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   986                    else if (mode == PHY_INTERFACE_MODE_RGMII_ID ||
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   987                             mode == PHY_INTERFACE_MODE_RGMII_RXID)
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   988                            delay = 2;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   989
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   990                    if (delay > QCA8K_MAX_DELAY) {
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   991                            dev_err(priv->dev, "rgmii rx delay is limited to a max value of 3ns, setting to the max value");
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   992                            delay = 3;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   993                    }
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   994
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   995                    priv->rgmii_rx_delay[cpu_port_index] = delay;
> > >                                                                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > Out of bounds
> > >
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   996
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   997                    break;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14   998            case PHY_INTERFACE_MODE_SGMII:
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14   999                    if (of_property_read_bool(port_dn, "qca,sgmii-txclk-falling-edge"))
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14  1000                            priv->sgmii_tx_clk_falling_edge = true;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14  1001
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14  1002                    if (of_property_read_bool(port_dn, "qca,sgmii-rxclk-falling-edge"))
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14  1003                            priv->sgmii_rx_clk_falling_edge = true;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14  1004
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14  1005                    break;
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14  1006            default:
> > > 5654ec78dd7e64 Ansuel Smith 2021-10-14  1007                    continue;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14  1008            }
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14  1009    }
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14  1010
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14  1011    return 0;
> > > 6c43809bf1bee7 Ansuel Smith 2021-10-14  1012  }
> > >
> > > ---
> > > 0-DAY CI Kernel Test Service, Intel Corporation
> > > https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
> > >

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ