| 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
| ||
|
Message-ID: <YMthNX8uHwLo4i/y@chrisdown.name>
Date: Thu, 17 Jun 2021 15:50:29 +0100
From: Chris Down <chris@...isdown.name>
To: Andy Shevchenko <andy.shevchenko@...il.com>
Cc: Petr Mladek <pmladek@...e.com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Jessica Yu <jeyu@...nel.org>,
Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
John Ogness <john.ogness@...utronix.de>,
Steven Rostedt <rostedt@...dmis.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Johannes Weiner <hannes@...xchg.org>,
Kees Cook <keescook@...omium.org>,
Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
Rasmus Villemoes <linux@...musvillemoes.dk>, kernel-team@...com
Subject: Re: [PATCH v7 0/5] printk: Userspace format indexing support
Andy Shevchenko writes:
>Assuming that Chris indeed spent time on checking string_esape_mem()
>users along with %*pE (and all its variations with hardcoded length)
>and haven't found any problems,
>Acked-by: Andy Shevchenko <andy.shevchenko@...il.com>
Thanks! Probably worth documenting my methodology :-)
Forgive the lack of wrapping -- I suspect it will probably make this easier to
read.
% git grep '"[^"]*%[*0-9]*pE[^"]*"'
drivers/gpu/drm/drm_dp_cec.c: seq_printf(file, "ID: %*pE\n",
drivers/gpu/drm/drm_dp_dual_mode_helper.c: drm_dbg_kms(dev, "DP dual mode HDMI ID: %*pE (err %zd)\n",
drivers/gpu/drm/drm_dp_helper.c: "%s: DP %s: OUI %*phD dev-ID %*pE HW-rev %d.%d SW-rev %d.%d quirks 0x%04x\n",
drivers/net/wireless/intel/ipw2x00/ipw2100.c: IPW_DEBUG_INFO("%s: Associated with '%*pE' at %s, channel %d (BSSID=%pM)\n",
drivers/net/wireless/intel/ipw2x00/ipw2100.c: IPW_DEBUG_HC("SSID: '%*pE'\n", ssid_len, essid);
drivers/net/wireless/intel/ipw2x00/ipw2100.c: "disassociated: '%*pE' %pM\n", priv->essid_len, priv->essid,
drivers/net/wireless/intel/ipw2x00/ipw2100.c: IPW_DEBUG_WX("Setting ESSID: '%*pE' (%d)\n", length, essid, length);
drivers/net/wireless/intel/ipw2x00/ipw2100.c: IPW_DEBUG_WX("Getting essid: '%*pE'\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: "associated: '%*pE' %pM\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: "deauthenticated: '%*pE' %pM: (0x%04X) - %s\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: "authenticated: '%*pE' %pM\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: "disassociated: '%*pE' %pM\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: "authenticated: '%*pE' %pM\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: "deauthenticated: '%*pE' %pM\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded due to capability mismatch.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded because of non-network ESSID.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded because of ESSID mismatch: '%*pE'.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE excluded because newer than current network.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE excluded because newer than current network.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded because of age: %ums.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded because of channel mismatch: %d != %d.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded because of privacy mismatch: %s != %s.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded because of the same BSSID match: %pM.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded because of invalid frequency/mode combination.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded because configured rate mask excludes AP mandatory rate.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' excluded because of no compatible rates.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("Network '%*pE (%pM)' is a viable match.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_MERGE("remove network %*pE\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded due to capability mismatch.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of non-network ESSID.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of ESSID mismatch: '%*pE'.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because '%*pE (%pM)' has a stronger signal.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of storming (%ums since last assoc attempt).\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of age: %ums.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of channel mismatch: %d != %d.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of privacy mismatch: %s != %s.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of BSSID mismatch: %pM.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of invalid frequency/mode combination.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of invalid channel in current GEO\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because configured rate mask excludes AP mandatory rate.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' excluded because of no compatible rates.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Network '%*pE (%pM)' is a viable match.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_INFO("ESSID locked to '%*pE'\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("%ssociation attempt: '%*pE', channel %d, 802.11%c [%d], %s[:%s], enc=%s%s%s%c%c\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG(IPW_DL_STATE, "associating: '%*pE' %pM\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_ASSOC("Expired '%*pE' (%pM) from network list.\n",
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_WX("Setting ESSID: '%*pE' (%d)\n", length, extra, length);
drivers/net/wireless/intel/ipw2x00/ipw2200.c: IPW_DEBUG_WX("Getting essid: '%*pE'\n",
drivers/net/wireless/intel/ipw2x00/libipw_rx.c: LIBIPW_DEBUG_MGMT("WLAN_EID_SSID: '%*pE' len=%d.\n",
drivers/net/wireless/intel/ipw2x00/libipw_rx.c: LIBIPW_DEBUG_SCAN("Filtered out '%*pE (%pM)' network.\n",
drivers/net/wireless/intel/ipw2x00/libipw_rx.c: LIBIPW_DEBUG_SCAN("'%*pE' (%pM): %c%c%c%c %c%c%c%c-%c%c%c%c %c%c%c%c\n",
drivers/net/wireless/intel/ipw2x00/libipw_rx.c: LIBIPW_DEBUG_SCAN("Dropped '%*pE' (%pM) via %s.\n",
drivers/net/wireless/intel/ipw2x00/libipw_rx.c: LIBIPW_DEBUG_SCAN("Expired '%*pE' (%pM) from network list.\n",
drivers/net/wireless/intel/ipw2x00/libipw_rx.c: LIBIPW_DEBUG_SCAN("Adding '%*pE' (%pM) via %s.\n",
drivers/net/wireless/intel/ipw2x00/libipw_rx.c: LIBIPW_DEBUG_SCAN("Updating '%*pE' (%pM) via %s.\n",
drivers/net/wireless/intel/ipw2x00/libipw_wx.c: LIBIPW_DEBUG_SCAN("Not showing network '%*pE (%pM)' due to age (%ums).\n",
drivers/net/wireless/intel/ipw2x00/libipw_wx.c: LIBIPW_DEBUG_WX("Setting key %d to '%*pE' (%d:%d bytes)\n",
drivers/net/wireless/intersil/hostap/hostap_proc.c: seq_printf(m, "%*pE", (int)bss->ssid_len, bss->ssid);
drivers/net/wireless/marvell/libertas/cfg.c: lbs_deb_scan("scan: %pM, capa %04x, chan %2d, %*pE, %d dBm\n",
drivers/net/wireless/marvell/libertas/mesh.c: lbs_deb_cmd("mesh config action %d type %x channel %d SSID %*pE\n",
drivers/platform/olpc/olpc-xo175-ec.c: dev_dbg(&priv->spi->dev, "got debug string [%*pE]\n",
drivers/platform/surface/surface3_power.c: snprintf(bix->serial, ARRAY_SIZE(bix->serial), "%3pE%6pE", buf + 7, buf);
drivers/platform/surface/surface3_power.c: snprintf(bix->OEM, ARRAY_SIZE(bix->OEM), "%3pE", buf);
drivers/platform/x86/wmi.c: pr_info("\tobject_id: %2pE\n", g->object_id);
drivers/scsi/scsi_lib.c: id_size = snprintf(id, id_len, "t10.%*pE",
drivers/soc/qcom/cmd-db.c: seq_printf(seq, "0x%05x: %*pEp", le32_to_cpu(ent->addr),
drivers/staging/rtl8192e/rtllib.h: snprintf(escaped, sizeof(escaped), "%*pE", essid_len, essid);
drivers/staging/rtl8192u/ieee80211/ieee80211.h: snprintf(escaped, sizeof(escaped), "%*pE", essid_len, essid);
drivers/staging/wlan-ng/prism2sta.c: netdev_info(wlandev->netdev, "Prism2 card SN: %*pE\n",
drivers/thunderbolt/xdomain.c: return sprintf(buf, "%*pE\n", (int)strlen(svc->key), svc->key);
drivers/tty/mips_ejtag_fdc.c: dev_dbg(priv->dev, "%s%u: out %08x: \"%*pE%*pE\"\n",
drivers/tty/mips_ejtag_fdc.c: dev_dbg(priv->dev, "%s%u: in %08x: \"%*pE\"\n",
drivers/tty/serial/serial_core.c: pr_info("SysRq is enabled by magic sequence '%*pE' on serial\n",
fs/overlayfs/overlayfs.h: pr_debug("getxattr(%pd2, \"%s\", \"%*pE\", %zu, 0) = %i\n",
fs/overlayfs/overlayfs.h: pr_debug("setxattr(%pd2, \"%s\", \"%*pE\", %zu, 0) = %i\n",
lib/test_printf.c: test("(null)", "%pE", NULL);
lib/test_printf.c: test("(efault)", "%pE", ERR_PTR(-11));
lib/test_printf.c: test("(efault)", "%pE", PTR_INVALID);
net/ceph/debugfs.c: seq_printf(s, "%*pE/%*pE\t0x%x",
net/ceph/debugfs.c: seq_printf(s, "%*pE\t0x%x", t->target_oid.name_len,
For all ESSID cases, there's no clear standard, but according to Cisco, double
quote is illegal in SSID (although there's no formalised standard). At the very
least it's extremely unusual, and even if it happens, I don't see how it could
cause problems for the cases here.
- drivers/gpu/drm/drm_dp_cec.c: Device ID, chance of quotes is approaching zero.
- drivers/gpu/drm/drm_dp_dual_mode_helper.c: Debugging message only.
- drivers/gpu/drm/drm_dp_helper.c: Debugging message only.
- drivers/net/wireless/intel/ipw2x00/ipw2100.c: ESSID case, debugging message only.
- drivers/net/wireless/intel/ipw2x00/ipw2200.c: ESSID case, debugging message only.
- drivers/net/wireless/intersil/hostap/hostap_proc.c: ESSID case, /proc.
- drivers/net/wireless/marvell/libertas/cfg.c: ESSID case, debugging message only.
- drivers/net/wireless/marvell/libertas/mesh.c: ESSID case, debugging message only.
- drivers/platform/olpc/olpc-xo175-ec.c: Debugging message only.
- drivers/platform/surface/surface3_power.c: It wouldn't make sense for either the OEM or serial to contain quotes, especially considering how low level this is.
- drivers/platform/x86/wmi.c: Debug dump only, there's not gonna be quotes in 2 character escaping anyway.
- drivers/scsi/scsi_lib.c: Reading vendor ID. Passed back as length anyway, so that's fine -- it just needs to be unique.
- drivers/soc/qcom/cmd-db.c: Debug dump, and judging by the code no way it's gonna contain quotes.
- drivers/staging/rtl8192e/rtllib.h: ESSID case, only used for netdev_dbg messages anyway.
- drivers/staging/rtl8192u/ieee80211/ieee80211.h: Ditto rtllib.h.
- drivers/staging/wlan-ng/prism2sta.c: Serial number, it's not gonna contain quotes. For debugging only anyway.
- drivers/thunderbolt/xdomain.c: Used in key_show, which is used as a rare device_attribute. Only used for device debug.
- drivers/tty/mips_ejtag_fdc.c: Debugging messages only.
- drivers/tty/serial/serial_core.c: Debugging message only, and benefits from quoting (but seems highly unlikely it would be there anyway).
- fs/overlayfs/overlayfs.h: Debugging messages only, and would actually even benefit from the new quoting.
- lib/test_printf.c: None of these have quotes.
- net/ceph/debugfs.c: Debugging only, and looks unlikely to be affected regardless.
Powered by blists - more mailing lists