1
0
mirror of https://github.com/samba-team/samba.git synced 2025-03-11 16:58:40 +03:00

37 Commits

Author SHA1 Message Date
Volker Lendecke
be9aaffdac r17333: Some C++ warnings 2007-10-10 11:38:26 -05:00
Jeremy Allison
dcef65acb5 r15162: Patch for bug #3668. Windows has a bug with LARGE_READX
where if you ask for exactly 64k bytes it returns 0.
Jeremy.
2007-10-10 11:16:27 -05:00
Jeremy Allison
1d710d06a2 r13915: Fixed a very interesting class of realloc() bugs found by Coverity.
realloc can return NULL in one of two cases - (1) the realloc failed,
(2) realloc succeeded but the new size requested was zero, in which
case this is identical to a free() call.

The error paths dealing with these two cases should be different,
but mostly weren't. Secondly the standard idiom for dealing with
realloc when you know the new size is non-zero is the following :

 tmp = realloc(p, size);
 if (!tmp) {
    SAFE_FREE(p);
    return error;
 } else {
    p = tmp;
 }

However, there were *many* *many* places in Samba where we were
using the old (broken) idiom of :

 p = realloc(p, size)
 if (!p) {
    return error;
 }

which will leak the memory pointed to by p on realloc fail.

This commit (hopefully) fixes all these cases by moving to
a standard idiom of :

 p = SMB_REALLOC(p, size)
 if (!p) {
    return error;
 }

Where if the realloc returns null due to the realloc failing
or size == 0 we *guarentee* that the storage pointed to by p
has been freed. This allows me to remove a lot of code that
was dealing with the standard (more verbose) method that required
a tmp pointer. This is almost always what you want. When a
realloc fails you never usually want the old memory, you
want to free it and get into your error processing asap.

For the 11 remaining cases where we really do need to keep the
old pointer I have invented the new macro SMB_REALLOC_KEEP_OLD_ON_ERROR,
which can be used as follows :

 tmp = SMB_REALLOC_KEEP_OLD_ON_ERROR(p, size);
 if (!tmp) {
    SAFE_FREE(p);
    return error;
 } else {
    p = tmp;
 }

SMB_REALLOC_KEEP_OLD_ON_ERROR guarentees never to free the
pointer p, even on size == 0 or realloc fail. All this is
done by a hidden extra argument to Realloc(), BOOL free_old_on_error
which is set appropriately by the SMB_REALLOC and SMB_REALLOC_KEEP_OLD_ON_ERROR
macros (and their array counterparts).

It remains to be seen what this will do to our Coverity bug count :-).

Jeremy.
2007-10-10 11:10:59 -05:00
Jeremy Allison
103cac7dd3 r13119: Fix for #1779 from William Jojo <jojowil@hvcc.edu>
Jeremy.
2007-10-10 11:06:14 -05:00
Gerald Carter
939c3cb5d7 r10656: BIG merge from trunk. Features not copied over
* \PIPE\unixinfo
* winbindd's {group,alias}membership new functions
* winbindd's lookupsids() functionality
* swat (trunk changes to be reverted as per discussion with Deryck)
2007-10-10 11:04:48 -05:00
Tim Potter
985dbb47d9 r8572: Remove crufty #define NO_SYSLOG as it's not used at all anymore. 2007-10-10 11:00:11 -05:00
Jeremy Allison
94779ccb39 r4212: Ensure we only look at the bottom bit of large_readx.
Set the 14 word version of write if size > 0xffff as
well as 64-bit offset.
Jeremy.
2007-10-10 10:53:37 -05:00
Jeremy Allison
9d4e57f06c r4188: Ensure we add in the upper length in the right place !
Jeremy.
2007-10-10 10:53:36 -05:00
Jeremy Allison
831cb21a87 r4186: Fix client & server to allow 127k READX calls.
Jeremy.
2007-10-10 10:53:36 -05:00
Jeremy Allison
620f2e608f r4088: Get medieval on our ass about malloc.... :-). Take control of all our allocation
functions so we can funnel through some well known functions. Should help greatly with
malloc checking.
HEAD patch to follow.
Jeremy.
2007-10-10 10:53:32 -05:00
Richard Sharpe
3eb33fbc64 r2959: If we want to support writes >= 65536 with cli_write, then it had better
return a size_t, not an ssize_t, and we had better left shift the upper
part of the write count, not right shift it.
2007-10-10 10:52:57 -05:00
Jeremy Allison
b1033fc77c r2373: Fix typo.
Jeremy.
2007-10-10 10:52:41 -05:00
Jeremy Allison
330025d1a6 r2371: Fix for talking to OS/2 clients (max_mux ignored) by Guenter Kukkukk <guenter.kukkukk@kukkukk.com>.
Bugid #1590.
Jeremy.
2007-10-10 10:52:41 -05:00
Jeremy Allison
68590b9e22 RPC fix from Ronan Waide <waider@waider.ie>. Tested with rpcecho.
Jeremy.
-
Gerald Carter
0fb724b321 *lots of small merges form HEAD
*sync up configure.in
*don't build torture tools in make all
*make sure to remove torture tools as part of make clean
-
Andrew Bartlett
3a7458f947 Merge from HEAD - make Samba compile with -Wwrite-strings without additional
warnings.  (Adds a lot of const).

Andrew Bartlett
-
Jeremy Allison
5431bae894 Merge Richard's write > 4Gb fix.
Jeremy.
-
Herb Lewis
17f685fdbf merge from 2.2 fix for smbclient large files -
Jeremy Allison
fff7f3cbe2 Test was reversed for ERRmoredata in cli_read.
Jeremy.
-
Gerald Carter
65e7b5273b sync'ing up for 3.0alpha20 release -
Andrew Tridgell
03ac082dcb updated the 3.0 branch from the head branch - ready for alpha18 -
Jeremy Allison
5b04b5f1df Correctly increment offset in cli_smbwrite.
Jeremy.
-
Jeremy Allison
24ef6258a1 Test against W2K that we're doing large read/writes correctly (we are).
At least with 14 word writes.
Jeremy.
-
Tim Potter
6a58c9bd06 Removed version number from file header.
Changed "SMB/Netbios" to "SMB/CIFS" in file header.
-
Jeremy Allison
01ff6ce496 Same fix as went into 2.2 (I'm waiting for jerry to finish some code).
Jeremy.
-
Andrew Tridgell
55d5828e60 use cli_is_error() instead of looking in smb_rcls, otherwise NT status
codes don't work correctly
-
Andrew Tridgell
1b778bc7d2 started converting NTSTATUS to be a structure on systems with gcc in order to make it type incompatible with BOOL so we catch errors sooner. This has already found a number of bugs -
Jeremy Allison
fe85a19b4b Re-added readbraw call to test with smbtorture. This code not yet
tested...
Jeremy.
-
Tim Potter
6dbdb0d813 A rewrite of the error handling in the libsmb client code. I've separated
out the error handling into a bunch of separate functions rather than all
being handled in one big function.

Fetch error codes from the last received packet:

    void cli_dos_error(struct cli_state *cli, uint8 *eclass, uint32 *num);
    uint32 cli_nt_error(struct cli_state *);

Convert errors to UNIX errno values:

    int cli_errno_from_dos(uint8 eclass, uint32 num);
    int cli_errno_from_nt(uint32 status);
    int cli_errno(struct cli_state *cli);

Detect different kinds of errors:

    BOOL cli_is_dos_error(struct cli_state *cli);
    BOOL cli_is_nt_error(struct cli_state *cli);
    BOOL cli_is_error(struct cli_state *cli);

This also means we now support CAP_STATUS32 as we can decode and understand
NT errors instead of just DOS errors.  Yay!

Ported a whole bunch of files in libsmb to use this new API instead of the
just the DOS error.
-
Andrew Tridgell
ba79d2a030 cli_read() was reading too many bytes. -
Jeremy Allison
2999eab5ab Use a logical cli_read(), removed the cli_read_one() hack.
Jeremy.
-
Tim Potter
9e074bc2bf Merged cli_read_one() function for reading DCE/RPC reply fragments. -
Andrew Tridgell
bbfbe03cc6 added some comments to make the cli read code clearer -
Andrew Tridgell
674ee2f1d1 next_token() was supposed to be a reentrant replacement for strtok(),
but the code suffered from bitrot and is not now reentrant. That means
we can get bizarre behaviour
i've fixed this by making next_token() reentrant and creating a
next_token_nr() that is a small non-reentrant wrapper for those lumps
of code (mostly smbclient) that have come to rely on the non-reentrant
behaviour
-
Jeremy Allison
6d65556ae8 Set correct reply word in large writeX (greater than 64k) replies.
Also added smbtorture test for this.
Jeremy.
-
Jeremy Allison
c55bcec817 Tidyup removing many of the 0xC0000000 | NT_STATUS_XXX stuff (only need NT_STATUS_XXX).
Removed IS_BITS_xxx macros as they were just reproducing "C" syntax in a more
obscure way.
Jeremy.
-
Andrew Tridgell
10c5470835 split clientgen.c into several parts
the next step is splitting out the auth code, to make adding lukes
NTLMSSP support easier
-