Thread (4 messages) 4 messages, 4 authors, 2022-03-22

Re: [PATCH 1/1 module-next] kallsyms: enhance %pS/s/b printing when KALLSYSMS is disabled

From: Petr Mladek <pmladek@suse.com>
Date: 2022-03-17 10:59:24
Also in: lkml

Possibly related (same subject, not in this thread)

On Wed 2022-03-16 10:05:40, Maninder Singh wrote:
quoted hunk ↗ jump to hunk
print module information when KALLSYMS is disabled.

No change for %pB, as it needs to know symbol name to adjust address
value which can't be done without KALLSYMS.

(A) original output with KALLSYMS:
[8.842129] ps function_1 [crash]
[8.842735] pS function_1+0x4/0x2c [crash]
[8.842890] pSb function_1+0x4/0x2c [crash b367e79021b9f3b0172f9a36d4261c1f528ca1b3]
[8.843175] pB function_1+0x4/0x2c [crash]
[8.843362] pBb function_1+0x4/0x2c [crash b367e79021b9f3b0172f9a36d4261c1f528ca1b3]

(B) original output without KALLSYMS:
[12.487424] ps 0xffff800000eb008c
[12.487598] pS 0xffff800000eb008c
[12.487723] pSb 0xffff800000eb008c
[12.487850] pB 0xffff800000eb008c
[12.487967] pBb 0xffff800000eb008c

(C) With patched kernel
with KALLYSMS:
[41.974576] ps function_1 [crash]
[41.975173] pS function_1+0x4/0x2c [crash]
[41.975386] pSb function_1+0x4/0x2c [crash a8b20caaec9635b316cf4812f6b55598fe2b7cee]
[41.975879] pB function_1+0x4/0x2c [crash]
[41.976076] pBb function_1+0x4/0x2c [crash a8b20caaec9635b316cf4812f6b55598fe2b7cee]

without KALLSYMS:
[9.624152] ps 0xffff800001bd008c [crash]	// similar to original, no changes
[9.624548] pS 0x(____ptrval____)+0x8c [crash]   // base address hashed and offset is without hash
[9.624847] pSb 0x(____ptrval____)+0x8c [crash a8b20caaec9635b316cf4812f6b55598fe2b7cee]
[9.625388] pB 0x(____ptrval____)+0x8c [crash]
[9.625594] pBb 0x(____ptrval____)+0x8c [crash a8b20caaec9635b316cf4812f6b55598fe2b7cee]

with disable hashing:
[8.563916] ps 0xffff800000f2008c [crash]
[8.564574] pS 0xffff800000f20000+0x8c [crash]
[8.564749] pSb 0xffff800000f20000+0x8c [crash 3423a8993a7033fb79e5add14bf9d8d6b56330ca]
[8.565008] pB 0xffff800000f20000+0x8c [crash]
[8.565154] pBb 0xffff800000f20000+0x8c [crash 3423a8993a7033fb79e5add14bf9d8d6b56330ca]
--- a/lib/vsprintf.c
+++ b/lib/vsprintf.c
@@ -979,33 +979,92 @@ char *bdev_name(char *buf, char *end, struct block_device *bdev,
 }
 #endif
 
+#if !defined(CONFIG_KALLSYMS) && defined(CONFIG_MODULES)
+static int sprint_module_info(char *buf, unsigned long value,
+			     int modbuildid, int backtrace, int symbol)
+{
+	struct module *mod;
+	unsigned long offset;
+	void *base;
+	char *modname;
+	int len;
+	const unsigned char *buildid = NULL;
+	bool add_offset;
+
+	if (is_ksym_addr(value))
+		return 0;
+
+	if (backtrace || symbol)
+		add_offset = true;
+	else
+		add_offset = false;
+
+	preempt_disable();
+	mod = __module_address(value);
+	if (mod) {
+		modname = mod->name;
+#if IS_ENABLED(CONFIG_STACKTRACE_BUILD_ID)
+		if (modbuildid)
+			buildid = mod->build_id;
+#endif
+		if (add_offset) {
+			base = mod->core_layout.base;
+			offset = value - (unsigned long)base;
+		}
+	}
+	preempt_enable();
+	if (!mod)
+		return 0;
I think that some earlier version of the patch allowed to print
also the address from vmlinux with the offset. My concern was
that it would show non-hashed base pointer. IMHO, it is fine
to show it hashed.
+
+	/* address belongs to module */
+	if (add_offset)
+		len = sprintf(buf, "0x%p+0x%lx", base, offset);
+	else
+		len = sprintf(buf, "0x%lx", value);
+
+	return len + fill_name_build_id(buf, modname, modbuildid, buildid, len);
+}
Otherwise, it looks good to me. I did also some basic testing.
The vmlinux address with offset can be added by a followup patch.
Feel free to use:

Reviewed-by: Petr Mladek <pmladek@suse.com>
Tested-by: Petr Mladek <pmladek@suse.com>

Best Regards,
Petr
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help