From: Thomas Hurd <thurd@tresys.com>
To: Stephen Smalley <sds@tycho.nsa.gov>,
Steve Lawrence <slawrence@tresys.com>, <selinux@tycho.nsa.gov>
Subject: Re: [PATCH] libsemanage: Change bunzip to use heap instead of stack for buffer.
Date: Thu, 19 Mar 2015 10:59:44 -0400 [thread overview]
Message-ID: <550AE460.1040004@tresys.com> (raw)
In-Reply-To: <5509C4EF.9040701@tycho.nsa.gov>
On 03/18/2015 02:33 PM, Stephen Smalley wrote:
> On 03/18/2015 02:28 PM, Stephen Smalley wrote:
>> On 02/26/2015 11:26 AM, Steve Lawrence wrote:
>>> On 02/26/2015 09:03 AM, Thomas Hurd wrote:
>>>> Fixes segfault on systems with less than 256K stack size.
>>>> After change, I was able to run semodule -l with a 32K stack size.
>>>> Additionally, fix potential memory leak on realloc failure.
>>>>
>>>> Signed-off-by: Thomas Hurd <thurd@tresys.com>
>>> Acked-by: Steve Lawrence <slawrence@tresys.com>
>>>
>>> Thanks!
>> This broke installing .pp files that are not bzip2'd.
>> Previously, one could transparently installed compressed or uncompressed
>> .pp files without difficulty.
> Before:
> /usr/sbin/semodule -i test_policy/test_policy.pp
>
> After:
> /usr/sbin/semodule -i test_policy/test_policy.pp
> libsemanage.bunzip: Failure reading bz2 archive. (No such file or
> directory).
>
Thanks. It turns out it's only the error message being printed is
incorrect. Uncompressed pp files will still be installed.
In the case of an uncompressed pp it will fail magic number checking in
BZ2_bzRead(). I think the fix is to unconditionally
check the magic number before BZ2_bzReadOpen() instead of only when
bzip_blocksize=0, since it falls through there anyway.
That way if the file is not compressed it will return immediately and
map_file can mmap it. If the file is compressed then it will
go through the BZ2 functions and any errors will be reported correctly.
Sending a patch with that fix.
>>>> ---
>>>> libsemanage/src/direct_api.c | 61 ++++++++++++++++++++++++++++++++------------
>>>> 1 file changed, 44 insertions(+), 17 deletions(-)
>>>>
>>>> diff --git a/libsemanage/src/direct_api.c b/libsemanage/src/direct_api.c
>>>> index b0ed338..be6bd3c 100644
>>>> --- a/libsemanage/src/direct_api.c
>>>> +++ b/libsemanage/src/direct_api.c
>>>> @@ -428,49 +428,76 @@ static ssize_t bzip(semanage_handle_t *sh, const char *filename, char *data,
>>>> * in the file. Returns -1 if file could not be decompressed. */
>>>> ssize_t bunzip(semanage_handle_t *sh, FILE *f, char **data)
>>>> {
>>>> - BZFILE* b;
>>>> + BZFILE* b = NULL;
>>>> size_t nBuf;
>>>> - char buf[1<<18];
>>>> - size_t size = sizeof(buf);
>>>> + char* buf = NULL;
>>>> + size_t size = 1<<18;
>>>> + size_t bufsize = size;
>>>> int bzerror;
>>>> size_t total=0;
>>>> + char* uncompress = NULL;
>>>> + char* tmpalloc = NULL;
>>>> + int ret = -1;
>>>> +
>>>> + buf = malloc(bufsize);
>>>> + if (buf == NULL) {
>>>> + ERR(sh, "Failure allocating memory.");
>>>> + goto exit;
>>>> + }
>>>>
>>>> if (!sh->conf->bzip_blocksize) {
>>>> bzerror = fread(buf, 1, BZ2_MAGICLEN, f);
>>>> rewind(f);
>>>> - if ((bzerror != BZ2_MAGICLEN) || memcmp(buf, BZ2_MAGICSTR, BZ2_MAGICLEN))
>>>> - return -1;
>>>> + if ((bzerror != BZ2_MAGICLEN) || memcmp(buf, BZ2_MAGICSTR, BZ2_MAGICLEN)) {
>>>> + ERR(sh, "bz2 magic number not found.");
>>>> + goto exit;
>>>> + }
>>>> /* fall through */
>>>> }
>>>>
>>>> b = BZ2_bzReadOpen ( &bzerror, f, 0, sh->conf->bzip_small, NULL, 0 );
>>>> if ( bzerror != BZ_OK ) {
>>>> - BZ2_bzReadClose ( &bzerror, b );
>>>> - return -1;
>>>> + ERR(sh, "Failure opening bz2 archive.");
>>>> + goto exit;
>>>> }
>>>> -
>>>> - char *uncompress = realloc(NULL, size);
>>>> -
>>>> +
>>>> + uncompress = malloc(size);
>>>> + if (uncompress == NULL) {
>>>> + ERR(sh, "Failure allocating memory.");
>>>> + goto exit;
>>>> + }
>>>> +
>>>> while ( bzerror == BZ_OK) {
>>>> - nBuf = BZ2_bzRead ( &bzerror, b, buf, sizeof(buf));
>>>> + nBuf = BZ2_bzRead ( &bzerror, b, buf, bufsize);
>>>> if (( bzerror == BZ_OK ) || ( bzerror == BZ_STREAM_END )) {
>>>> if (total + nBuf > size) {
>>>> size *= 2;
>>>> - uncompress = realloc(uncompress, size);
>>>> + tmpalloc = realloc(uncompress, size);
>>>> + if (tmpalloc == NULL) {
>>>> + ERR(sh, "Failure allocating memory.");
>>>> + goto exit;
>>>> + }
>>>> + uncompress = tmpalloc;
>>>> }
>>>> memcpy(&uncompress[total], buf, nBuf);
>>>> total += nBuf;
>>>> }
>>>> }
>>>> if ( bzerror != BZ_STREAM_END ) {
>>>> - BZ2_bzReadClose ( &bzerror, b );
>>>> - free(uncompress);
>>>> - return -1;
>>>> + ERR(sh, "Failure reading bz2 archive.");
>>>> + goto exit;
>>>> }
>>>> - BZ2_bzReadClose ( &bzerror, b );
>>>>
>>>> + ret = total;
>>>> *data = uncompress;
>>>> - return total;
>>>> +
>>>> +exit:
>>>> + BZ2_bzReadClose ( &bzerror, b );
>>>> + free(buf);
>>>> + if ( ret < 0 ) {
>>>> + free(uncompress);
>>>> + }
>>>> + return ret;
>>>> }
>>>>
>>>> /* mmap() a file to '*data',
>>>>
>>> _______________________________________________
>>> Selinux mailing list
>>> Selinux@tycho.nsa.gov
>>> To unsubscribe, send email to Selinux-leave@tycho.nsa.gov.
>>> To get help, send an email containing "help" to Selinux-request@tycho.nsa.gov.
>>>
>>>
prev parent reply other threads:[~2015-03-19 14:59 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-02-26 14:03 [PATCH] libsemanage: Change bunzip to use heap instead of stack for buffer Thomas Hurd
2015-02-26 16:26 ` Steve Lawrence
2015-03-18 18:28 ` Stephen Smalley
2015-03-18 18:33 ` Stephen Smalley
2015-03-19 14:59 ` Thomas Hurd [this message]
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=550AE460.1040004@tresys.com \
--to=thurd@tresys.com \
--cc=sds@tycho.nsa.gov \
--cc=selinux@tycho.nsa.gov \
--cc=slawrence@tresys.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.