From: Rusty Russell <rusty@rustcorp.com.au>
To: Jeff Arnold <jbarnold@mit.edu>
Cc: Andrew Morton <akpm@linux-foundation.org>,
linux-kernel@vger.kernel.org,
Denys Vlasenko <vda.linux@googlemail.com>,
Tim Abbott <tabbott@mit.edu>, Anders Kaseorg <andersk@mit.edu>,
Waseem Daher <wdaher@mit.edu>,
Nikanth Karthikesan <knikanth@suse.de>
Subject: Re: [PATCH 3/7] Ksplice: Make find_symbol return a struct kernel_symbol
Date: Wed, 4 Feb 2009 19:56:13 +1030 [thread overview]
Message-ID: <200902041956.13958.rusty@rustcorp.com.au> (raw)
In-Reply-To: <1228521840-3886-4-git-send-email-jbarnold@mit.edu>
On Saturday 06 December 2008 10:33:56 Jeff Arnold wrote:
> From: Tim Abbott <tabbott@mit.edu>
>
> Ksplice needs access to the kernel_symbol structure in order to support
> modifications to the exported symbol table.
Applied, with some fixes:
> preempt_disable();
> - if (IS_ERR_VALUE(find_symbol(symbol, &owner, NULL, true, false)))
> + if (find_symbol(symbol, &owner, NULL, true, false) == NULL)
> BUG();
I skipped the == NULL part; if (find_symbol... reads quite well. Same with
three or so other places.
> - ret = find_symbol(name, &owner, &crc,
> + sym = find_symbol(name, &owner, &crc,
> !(mod->taints & (1 << TAINT_PROPRIETARY_MODULE)), true);
> - if (!IS_ERR_VALUE(ret)) {
> - /* use_module can fail due to OOM,
> - or module initialization or unloading */
> - if (!check_version(sechdrs, versindex, name, mod, crc) ||
> - !use_module(mod, owner))
> - ret = -EINVAL;
> - }
> - return ret;
> + /* use_module can fail due to OOM,
> + or module initialization or unloading */
> + if (!check_version(sechdrs, versindex, name, mod, crc) ||
> + !use_module(mod, owner))
> + sym = NULL;
> + return sym;
> }
If sym is NULL those other values aren't valid! You can't take out the
branch....
> - return (void *)value;
> + return sym == NULL ? (void *)0 : (void *)sym->value;
> }
One NULL too many, and one too few:
return sym ? (void *)sym->value : NULL;
Here's the nett diff:
diff --git a/kernel/module.c b/kernel/module.c
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -860,7 +860,7 @@ void __symbol_put(const char *symbol)
struct module *owner;
preempt_disable();
- if (find_symbol(symbol, &owner, NULL, true, false) == NULL)
+ if (!find_symbol(symbol, &owner, NULL, true, false))
BUG();
module_put(owner);
preempt_enable();
@@ -1023,7 +1023,7 @@ static inline int check_modstruct_versio
{
const unsigned long *crc;
- if (find_symbol("struct_module", NULL, &crc, true, false) == NULL)
+ if (!find_symbol("struct_module", NULL, &crc, true, false))
BUG();
return check_version(sechdrs, versindex, "struct_module", mod, crc);
}
@@ -1077,9 +1077,11 @@ static const struct kernel_symbol *resol
!(mod->taints & (1 << TAINT_PROPRIETARY_MODULE)), true);
/* use_module can fail due to OOM,
or module initialization or unloading */
- if (!check_version(sechdrs, versindex, name, mod, crc) ||
- !use_module(mod, owner))
- sym = NULL;
+ if (sym) {
+ if (!check_version(sechdrs, versindex, name, mod, crc) ||
+ !use_module(mod, owner))
+ sym = NULL;
+ }
return sym;
}
@@ -1481,11 +1483,11 @@ void *__symbol_get(const char *symbol)
preempt_disable();
sym = find_symbol(symbol, &owner, NULL, true, true);
- if (sym != NULL && strong_try_module_get(owner))
+ if (sym && strong_try_module_get(owner))
sym = NULL;
preempt_enable();
- return sym == NULL ? (void *)0 : (void *)sym->value;
+ return sym ? (void *)sym->value : NULL;
}
EXPORT_SYMBOL_GPL(__symbol_get);
@@ -1513,8 +1515,7 @@ static int verify_export_symbols(struct
for (i = 0; i < ARRAY_SIZE(arr); i++) {
for (s = arr[i].sym; s < arr[i].sym + arr[i].num; s++) {
- if (find_symbol(s->name, &owner, NULL, true, false)
- != NULL) {
+ if (find_symbol(s->name, &owner, NULL, true, false)) {
printk(KERN_ERR
"%s: exports duplicate symbol %s"
" (owned by %s)\n",
@@ -1561,7 +1562,7 @@ static int simplify_symbols(Elf_Shdr *se
ksym = resolve_symbol(sechdrs, versindex,
strtab + sym[i].st_name, mod);
/* Ok if resolved. */
- if (ksym != NULL) {
+ if (ksym) {
sym[i].st_value = ksym->value;
break;
}
next prev parent reply other threads:[~2009-02-04 9:26 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-12-06 0:03 [PATCH 0/7] Ksplice: Rebootless kernel updates Jeff Arnold
2008-12-06 0:03 ` [PATCH 1/7] Make section names compatible with -ffunction-sections -fdata-sections Jeff Arnold
2008-12-06 0:03 ` [PATCH 2/7] x86: Add an option to compile " Jeff Arnold
2008-12-06 0:03 ` [PATCH 3/7] Ksplice: Make find_symbol return a struct kernel_symbol Jeff Arnold
2008-12-06 0:03 ` [PATCH 4/7] Ksplice: Add module_data_address (the analogue of module_text_address) Jeff Arnold
2008-12-06 0:03 ` [PATCH 5/7] Ksplice: Add functions for walking kallsyms symbols Jeff Arnold
2008-12-06 0:03 ` [PATCH 6/7] Ksplice: Export symbols needed for Ksplice Jeff Arnold
2008-12-06 0:04 ` [PATCH 7/7] Ksplice: Support updating x86-32 and x86-64 Jeff Arnold
2008-12-17 5:41 ` Theodore Tso
2008-12-18 2:09 ` Tim Abbott
2009-02-07 2:36 ` Rusty Russell
2009-02-10 1:01 ` Tim Abbott
2009-02-04 11:35 ` [PATCH 6/7] Ksplice: Export symbols needed for Ksplice Rusty Russell
2009-02-13 1:46 ` Tim Abbott
2009-02-16 7:11 ` Rusty Russell
2009-02-04 11:30 ` [PATCH 5/7] Ksplice: Add functions for walking kallsyms symbols Rusty Russell
2009-02-04 21:31 ` Anders Kaseorg
2009-02-04 11:21 ` [PATCH 4/7] Ksplice: Add module_data_address (the analogue of module_text_address) Rusty Russell
2009-02-04 20:48 ` Anders Kaseorg
2009-02-07 12:45 ` Rusty Russell
2009-02-04 9:26 ` Rusty Russell [this message]
2009-02-04 8:38 ` [PATCH 2/7] x86: Add an option to compile with -ffunction-sections -fdata-sections Rusty Russell
2009-02-04 10:26 ` Anders Kaseorg
2009-02-04 10:58 ` Rusty Russell
2009-02-04 20:50 ` Anders Kaseorg
2008-12-06 8:46 ` [PATCH 1/7] Make section names compatible " Andi Kleen
2008-12-31 19:18 ` Tim Abbott
2008-12-31 19:52 ` Andi Kleen
2008-12-31 21:59 ` Tim Abbott
2009-01-01 16:32 ` Andrew Morton
2009-01-04 19:20 ` Tim Abbott
2009-01-12 21:51 ` Andrew Morton
2009-01-12 22:11 ` Andi Kleen
2009-02-04 8:15 ` Rusty Russell
2009-02-05 1:11 ` Anders Kaseorg
2009-02-05 2:00 ` Anders Kaseorg
2008-12-17 2:48 ` [PATCH 0/7] Ksplice: Rebootless kernel updates Tim Abbott
2008-12-17 3:07 ` Andrew Morton
2008-12-17 3:53 ` Dave Jones
2008-12-17 17:19 ` Jeff Arnold
2008-12-17 5:05 ` Tim Abbott
2008-12-17 12:09 ` Ben Collins
2008-12-17 12:06 ` Ben Collins
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=200902041956.13958.rusty@rustcorp.com.au \
--to=rusty@rustcorp.com.au \
--cc=akpm@linux-foundation.org \
--cc=andersk@mit.edu \
--cc=jbarnold@mit.edu \
--cc=knikanth@suse.de \
--cc=linux-kernel@vger.kernel.org \
--cc=tabbott@mit.edu \
--cc=vda.linux@googlemail.com \
--cc=wdaher@mit.edu \
/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