400711 - SSL_CanBypass leaks memory

Status: RESOLVED FIXED

(NSS :: Libraries, defect, P2)

Description

[glen beasley]

reported by Sun engineer Meena Vyas

libumem shows this :
umem_alloc_160 leak: 2 buffers, 160 bytes each, 320 bytes total
ADDR          BUFADDR        TIMESTAMP   THREAD  CACHE
9bbc98         9b9628   1f9d80d3d4b6f0        1  2f008    
libumem.so.1`umem_cache_alloc+0x210
libumem.so.1`umem_alloc+0x60
libumem.so.1`malloc+0x28
libnspr4.so`PR_Malloc+0x78
libnss3.so`PORT_Alloc+0x48
libssl3.so`SSL_CanBypass+0x4d0
libns-httpd40.so`int SSLSocketConfiguration::check_bypassconst+0x140
libns-httpd40.so`void SSLSocketConfiguration::enableSSLconst+0x35c


http://mxr.mozilla.org/security/source/security/nss/lib/ssl/derive.c#683
683             enc_pms.data = (unsigned char*)PORT_Alloc(enc_pms.len)

the allocation is never freed.

Comment 1

[glen beasley]

Attachment: free data if allocated

Comment 2

[Nelson Bolyard (seldom reads bugmail)]

Comment on attachment 286578 [details] [diff] [review]
free data  if allocated 

PORT_Free tests its argument, so this additional test is
redundant.  But even better is to call SECITEM_FreeItem.
Please do that.

>+    if (enc_pms.data)
>+       PORT_Free(enc_pms.data);
>+

Comment 3

[glen beasley]

Attachment: use SECITEM_FreeItem to free

Comment 4

[Nelson Bolyard (seldom reads bugmail)]

Comment on attachment 286724 [details] [diff] [review]
use SECITEM_FreeItem to free 

r=nelson

Comment 5

[glen beasley]

/cvsroot/mozilla/security/nss/lib/ssl/derive.c,v  <--  derive.c
new revision: 1.8; previous revision: 1.7
done

Comment 6

[glen beasley]

3_11_Branch checkin:

/cvsroot/mozilla/security/nss/lib/ssl/derive.c,v  <--  derive.c
new revision: 1.3.2.3; previous revision: 1.3.2.2
done