mirror of
https://github.com/kernelkit/infix.git
synced 2026-08-02 05:43:02 +02:00
Joachim Wiberg
442b572478
- Allow parenthesis in tag names - Unescape Linux /dev/kmsg messages for unicode characters Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
143 lines
4.0 KiB
Diff
143 lines
4.0 KiB
Diff
From c1c3f253a8963c6e9e41fafcf9b888f3f5a34906 Mon Sep 17 00:00:00 2001
|
|
From: Joachim Wiberg <troglobit@gmail.com>
|
|
Date: Mon, 3 Nov 2025 13:11:56 +0100
|
|
Subject: [PATCH 4/4] syslogd: unescape Linux /dev/kmsg messages before
|
|
sanitization
|
|
Organization: Wires
|
|
|
|
Linux's /dev/kmsg interface escapes all non-printable characters and
|
|
backslashes using C-style hex encoding (\xHH format). Preventing the
|
|
recent UTF-8 sanitization improvements from working correctly, since
|
|
UTF-8 sequences arrive as escaped text like "\xe2\x80\x94" rather than
|
|
as actual bytes, see [1] for details.
|
|
|
|
This commit implements "unescaping" of the kernel's C-style format
|
|
before applying the UTF-8-aware sanitization. Ensuring that UTF-8
|
|
content also in kernel messages is properly preserved when using the
|
|
-8 flag.
|
|
|
|
[1]: https://www.kernel.org/doc/Documentation/ABI/testing/dev-kmsg
|
|
|
|
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
|
|
---
|
|
src/syslogd.c | 81 ++++++++++++++++++++++++++++++++++++++++++++++++---
|
|
1 file changed, 77 insertions(+), 4 deletions(-)
|
|
|
|
diff --git a/src/syslogd.c b/src/syslogd.c
|
|
index 37e1920..0b6bc30 100644
|
|
--- a/src/syslogd.c
|
|
+++ b/src/syslogd.c
|
|
@@ -155,6 +155,7 @@ static int KernLog = 1; /* Track kernel logs by default */
|
|
static int KeepKernFac; /* Keep remotely logged kernel facility */
|
|
static int KeepKernTime; /* Keep kernel timestamp, evern after initial read */
|
|
static int KeepKernConsole; /* Keep kernel logging to console */
|
|
+static int IsLinuxKmsg; /* Set if reading from Linux /dev/kmsg */
|
|
|
|
static int rotate_opt; /* Set if command line option has been given (wins) */
|
|
static off_t RotateSz = 0; /* Max file size (bytes) before rotating, disabled by default */
|
|
@@ -651,8 +652,10 @@ int main(int argc, char *argv[])
|
|
_PATH_KLOG);
|
|
else
|
|
kern_console_off();
|
|
- } else
|
|
+ } else {
|
|
+ IsLinuxKmsg = 1;
|
|
kern_console_off();
|
|
+ }
|
|
}
|
|
no_klogd:
|
|
consfile.f_type = F_CONSOLE;
|
|
@@ -1062,6 +1065,61 @@ utf8_valid(const unsigned char *in, size_t len)
|
|
return 1;
|
|
}
|
|
|
|
+/*
|
|
+ * Unescapes Linux /dev/kmsg messages that use C-style hex encoding.
|
|
+ * Converts "\xHH" sequences back to bytes and "\\" back to "\".
|
|
+ * Returns the new length of the unescaped string.
|
|
+ */
|
|
+static size_t kmsg_unescape(char *msg)
|
|
+{
|
|
+ char *src, *dst;
|
|
+ int hi, lo;
|
|
+
|
|
+ src = dst = msg;
|
|
+ while (*src) {
|
|
+ if (*src == '\\' && src[1]) {
|
|
+ if (src[1] == 'x' && src[2] && src[3]) {
|
|
+ /* Decode \xHH */
|
|
+ hi = src[2];
|
|
+ lo = src[3];
|
|
+
|
|
+ /* Convert hex digits to values */
|
|
+ if (hi >= '0' && hi <= '9')
|
|
+ hi = hi - '0';
|
|
+ else if (hi >= 'a' && hi <= 'f')
|
|
+ hi = hi - 'a' + 10;
|
|
+ else if (hi >= 'A' && hi <= 'F')
|
|
+ hi = hi - 'A' + 10;
|
|
+ else
|
|
+ goto copy_literal;
|
|
+
|
|
+ if (lo >= '0' && lo <= '9')
|
|
+ lo = lo - '0';
|
|
+ else if (lo >= 'a' && lo <= 'f')
|
|
+ lo = lo - 'a' + 10;
|
|
+ else if (lo >= 'A' && lo <= 'F')
|
|
+ lo = lo - 'A' + 10;
|
|
+ else
|
|
+ goto copy_literal;
|
|
+
|
|
+ *dst++ = (char)((hi << 4) | lo);
|
|
+ src += 4;
|
|
+ continue;
|
|
+ } else if (src[1] == '\\') {
|
|
+ /* Decode \\ to \ */
|
|
+ *dst++ = '\\';
|
|
+ src += 2;
|
|
+ continue;
|
|
+ }
|
|
+ }
|
|
+copy_literal:
|
|
+ *dst++ = *src++;
|
|
+ }
|
|
+ *dst = '\0';
|
|
+
|
|
+ return dst - msg;
|
|
+}
|
|
+
|
|
/*
|
|
* Removes characters from log messages that are unsafe to display.
|
|
* Preserves valid UTF-8 sequences, including BOM, with -8 flag.
|
|
@@ -1773,9 +1831,24 @@ void printsys(char *msg)
|
|
parsemsg_rfc3164_app_name_procid(&p, &buffer.app_name, &buffer.proc_id);
|
|
|
|
q = lp;
|
|
- while (*p != '\0' && (c = *p++) != '\n' && q < &line[MAXLINE])
|
|
- *q++ = c;
|
|
- *q = '\0';
|
|
+ if (IsLinuxKmsg) {
|
|
+ char tmp[MAXLINE + 1];
|
|
+ char *t = tmp;
|
|
+
|
|
+ while (*p != '\0' && *p != '\n' && t < &tmp[MAXLINE])
|
|
+ *t++ = *p++;
|
|
+ *t = '\0';
|
|
+
|
|
+ /* Unescape \xHH sequences and \\ */
|
|
+ kmsg_unescape(tmp);
|
|
+
|
|
+ /* Sanitize the unescaped message with UTF-8 support */
|
|
+ parsemsg_remove_unsafe_characters(tmp, lp, MAXLINE);
|
|
+ } else {
|
|
+ while (*p != '\0' && (c = *p++) != '\n' && q < &line[MAXLINE])
|
|
+ *q++ = c;
|
|
+ *q = '\0';
|
|
+ }
|
|
|
|
logmsg(&buffer);
|
|
}
|
|
--
|
|
2.43.0
|
|
|