| Message ID | 20101004192459.GE5692@bicker |
|---|---|
| State | Accepted, archived |
| Delegated to: | David Miller |
| Headers | show |
On Mon, 4 Oct 2010, Dan Carpenter wrote: > In diva_mnt_add_xdi_adapter() we do this: > strcpy (clients[id].drvName, tmp); > strcpy (clients[id].Dbg.drvName, tmp); > > The "clients[id].drvName" is a 128 character buffer and > "clients[id].Dbg.drvName" was originally a 16 character buffer but I've > changed it to 128 as well. We don't actually use 128 characters but we > do use more than 16. I don't see any reason for that change. The driver names here do not use more than 16 characters and when filled, the length is checked anyway. Please avoid changing the size of that structure. Armin > I've also changed the size of "tmp" to 128 characters instead of 256. > > Signed-off-by: Dan Carpenter <error27@gmail.com> > > diff --git a/drivers/isdn/hardware/eicon/debuglib.h b/drivers/isdn/hardware/eicon/debuglib.h > index 8ea5877..02eed6b 100644 > --- a/drivers/isdn/hardware/eicon/debuglib.h > +++ b/drivers/isdn/hardware/eicon/debuglib.h > @@ -249,7 +249,7 @@ typedef struct _DbgHandle_ > } regTime ; /* timestamp for registration */ > void *pIrp ; /* ptr to pending i/o request */ > unsigned long dbgMask ; /* current debug mask */ > - char drvName[16] ; /* ASCII name of registered driver */ > + char drvName[128] ; /* ASCII name of registered driver */ > char drvTag[64] ; /* revision string */ > DbgEnd dbg_end ; /* function for debug closing */ > DbgLog dbg_prt ; /* function for debug appending */ > diff --git a/drivers/isdn/hardware/eicon/debug.c b/drivers/isdn/hardware/eicon/debug.c > index 33ce89e..3626401 100644 > --- a/drivers/isdn/hardware/eicon/debug.c > +++ b/drivers/isdn/hardware/eicon/debug.c > @@ -862,7 +862,7 @@ void diva_mnt_add_xdi_adapter (const DESCRIPTOR* d) { > diva_os_spin_lock_magic_t old_irql, old_irql1; > dword sec, usec, logical, serial, org_mask; > int id, best_id = 0, free_id = -1; > - char tmp[256]; > + char tmp[128]; > diva_dbg_entry_head_t* pmsg = NULL; > int len; > word size; > -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Oct 06, 2010 at 09:25:44AM +0200, Armin Schindler wrote: > On Mon, 4 Oct 2010, Dan Carpenter wrote: >> In diva_mnt_add_xdi_adapter() we do this: >> strcpy (clients[id].drvName, tmp); >> strcpy (clients[id].Dbg.drvName, tmp); >> >> The "clients[id].drvName" is a 128 character buffer and >> "clients[id].Dbg.drvName" was originally a 16 character buffer but I've >> changed it to 128 as well. We don't actually use 128 characters but we >> do use more than 16. > > I don't see any reason for that change. The driver names here do not use > more than 16 characters and when filled, the length is checked anyway. > Please avoid changing the size of that structure. > drivers/isdn/hardware/eicon/debug.c diva_mnt_add_xdi_adapter() 874 sprintf (tmp, "ADAPTER:%d SN:%u-%d", 12345678 90123 45 67 That's a minimum 17 characters. 875 (int)logical, 876 serial & 0x00ffffff, 877 (byte)(((serial & 0xff000000) >> 24) + 1)); 878 } else { 879 sprintf (tmp, "ADAPTER:%d SN:%u", (int)logical, serial); 880 } regards, dan carpenter > Armin -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Dan Carpenter schrieb: > On Wed, Oct 06, 2010 at 09:25:44AM +0200, Armin Schindler wrote: >> On Mon, 4 Oct 2010, Dan Carpenter wrote: >>> In diva_mnt_add_xdi_adapter() we do this: >>> strcpy (clients[id].drvName, tmp); >>> strcpy (clients[id].Dbg.drvName, tmp); >>> >>> The "clients[id].drvName" is a 128 character buffer and >>> "clients[id].Dbg.drvName" was originally a 16 character buffer but I've >>> changed it to 128 as well. We don't actually use 128 characters but we >>> do use more than 16. >> I don't see any reason for that change. The driver names here do not use >> more than 16 characters and when filled, the length is checked anyway. >> Please avoid changing the size of that structure. >> > > drivers/isdn/hardware/eicon/debug.c diva_mnt_add_xdi_adapter() > 874 sprintf (tmp, "ADAPTER:%d SN:%u-%d", > 12345678 90123 45 67 > > That's a minimum 17 characters. > > 875 (int)logical, > 876 serial & 0x00ffffff, > 877 (byte)(((serial & 0xff000000) >> 24) + 1)); > 878 } else { > 879 sprintf (tmp, "ADAPTER:%d SN:%u", (int)logical, serial); > 880 } > for that reason i am a fan of kasprintf(). Maybe this a solution here also ? samething like: kasprintf (&buf, "ADAPTER:%d SN:%u-%d",12345678,90123,45,67): ... kstrncpy(tmp,buf,sizeof(buf)); .... free(buf); that would keep overflows away until the changes in the structure are ready. re, wh -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, 6 Oct 2010, Dan Carpenter wrote: > On Wed, Oct 06, 2010 at 09:25:44AM +0200, Armin Schindler wrote: >> On Mon, 4 Oct 2010, Dan Carpenter wrote: >>> In diva_mnt_add_xdi_adapter() we do this: >>> strcpy (clients[id].drvName, tmp); >>> strcpy (clients[id].Dbg.drvName, tmp); >>> >>> The "clients[id].drvName" is a 128 character buffer and >>> "clients[id].Dbg.drvName" was originally a 16 character buffer but I've >>> changed it to 128 as well. We don't actually use 128 characters but we >>> do use more than 16. >> >> I don't see any reason for that change. The driver names here do not use >> more than 16 characters and when filled, the length is checked anyway. >> Please avoid changing the size of that structure. >> > > drivers/isdn/hardware/eicon/debug.c diva_mnt_add_xdi_adapter() > 874 sprintf (tmp, "ADAPTER:%d SN:%u-%d", > 12345678 90123 45 67 > > That's a minimum 17 characters. > > 875 (int)logical, > 876 serial & 0x00ffffff, > 877 (byte)(((serial & 0xff000000) >> 24) + 1)); > 878 } else { > 879 sprintf (tmp, "ADAPTER:%d SN:%u", (int)logical, serial); > 880 } this is tmp with a bigger size. It seems you are mixing the sizes of drvName and tmp. Armin -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Oct 06, 2010 at 10:21:02AM +0200, Armin Schindler wrote: > On Wed, 6 Oct 2010, Dan Carpenter wrote: >> On Wed, Oct 06, 2010 at 09:25:44AM +0200, Armin Schindler wrote: >>> On Mon, 4 Oct 2010, Dan Carpenter wrote: >>>> In diva_mnt_add_xdi_adapter() we do this: >>>> strcpy (clients[id].drvName, tmp); >>>> strcpy (clients[id].Dbg.drvName, tmp); >>>> >>>> The "clients[id].drvName" is a 128 character buffer and >>>> "clients[id].Dbg.drvName" was originally a 16 character buffer but I've >>>> changed it to 128 as well. We don't actually use 128 characters but we >>>> do use more than 16. >>> >>> I don't see any reason for that change. The driver names here do not use >>> more than 16 characters and when filled, the length is checked anyway. >>> Please avoid changing the size of that structure. >>> >> >> drivers/isdn/hardware/eicon/debug.c diva_mnt_add_xdi_adapter() >> 874 sprintf (tmp, "ADAPTER:%d SN:%u-%d", >> 12345678 90123 45 67 >> >> That's a minimum 17 characters. >> >> 875 (int)logical, >> 876 serial & 0x00ffffff, >> 877 (byte)(((serial & 0xff000000) >> 24) + 1)); >> 878 } else { >> 879 sprintf (tmp, "ADAPTER:%d SN:%u", (int)logical, serial); >> 880 } > > this is tmp with a bigger size. It seems you are mixing the sizes of > drvName and tmp. > What I mean is that later on we use strcpy() to copy "tmp" into "clients[id].Dbg.drvName" 927 strcpy (clients[id].drvName, tmp); 928 strcpy (clients[id].Dbg.drvName, tmp); ^ this buffer is only 16 chars regards, dan carpenter > Armin -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, 6 Oct 2010, Dan Carpenter wrote: > On Wed, Oct 06, 2010 at 10:21:02AM +0200, Armin Schindler wrote: >> On Wed, 6 Oct 2010, Dan Carpenter wrote: >>> On Wed, Oct 06, 2010 at 09:25:44AM +0200, Armin Schindler wrote: >>>> On Mon, 4 Oct 2010, Dan Carpenter wrote: >>>>> In diva_mnt_add_xdi_adapter() we do this: >>>>> strcpy (clients[id].drvName, tmp); >>>>> strcpy (clients[id].Dbg.drvName, tmp); >>>>> >>>>> The "clients[id].drvName" is a 128 character buffer and >>>>> "clients[id].Dbg.drvName" was originally a 16 character buffer but I've >>>>> changed it to 128 as well. We don't actually use 128 characters but we >>>>> do use more than 16. >>>> >>>> I don't see any reason for that change. The driver names here do not use >>>> more than 16 characters and when filled, the length is checked anyway. >>>> Please avoid changing the size of that structure. >>>> >>> >>> drivers/isdn/hardware/eicon/debug.c diva_mnt_add_xdi_adapter() >>> 874 sprintf (tmp, "ADAPTER:%d SN:%u-%d", >>> 12345678 90123 45 67 >>> >>> That's a minimum 17 characters. >>> >>> 875 (int)logical, >>> 876 serial & 0x00ffffff, >>> 877 (byte)(((serial & 0xff000000) >> 24) + 1)); >>> 878 } else { >>> 879 sprintf (tmp, "ADAPTER:%d SN:%u", (int)logical, serial); >>> 880 } >> >> this is tmp with a bigger size. It seems you are mixing the sizes of >> drvName and tmp. >> > > What I mean is that later on we use strcpy() to copy "tmp" into > "clients[id].Dbg.drvName" > > 927 strcpy (clients[id].drvName, tmp); > 928 strcpy (clients[id].Dbg.drvName, tmp); > ^ > this buffer is only 16 chars Now I understand. You are right. So the fix would be to change these strcpy() to e.g. strncpy() or similar. Armin -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Oct 06, 2010 at 10:44:47AM +0200, Armin Schindler wrote: >> 927 strcpy (clients[id].drvName, tmp); >> 928 strcpy (clients[id].Dbg.drvName, tmp); >> ^ >> this buffer is only 16 chars > > Now I understand. You are right. So the fix would be to change these > strcpy() to e.g. strncpy() or similar. > We need more than 16 characters to store the information. What is the problem with just making the buffer larger? regards, dan carpenter > Armin -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, 6 Oct 2010, Dan Carpenter wrote: > On Wed, Oct 06, 2010 at 10:44:47AM +0200, Armin Schindler wrote: >>> 927 strcpy (clients[id].drvName, tmp); >>> 928 strcpy (clients[id].Dbg.drvName, tmp); >>> ^ >>> this buffer is only 16 chars >> >> Now I understand. You are right. So the fix would be to change these >> strcpy() to e.g. strncpy() or similar. >> > > We need more than 16 characters to store the information. What is the > problem with just making the buffer larger? The larger buffer is not really needed. Later versions of the driver will use the strcpy/memcpy with checked size anyway. I just want to avoid a change in the debug structure which is used in more than one module/tool. But it shouldn't be a real problem. Armin -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
From: Dan Carpenter <error27@gmail.com> Date: Mon, 4 Oct 2010 21:24:59 +0200 > In diva_mnt_add_xdi_adapter() we do this: > strcpy (clients[id].drvName, tmp); > strcpy (clients[id].Dbg.drvName, tmp); > > The "clients[id].drvName" is a 128 character buffer and > "clients[id].Dbg.drvName" was originally a 16 character buffer but I've > changed it to 128 as well. We don't actually use 128 characters but we > do use more than 16. > > I've also changed the size of "tmp" to 128 characters instead of 256. > > Signed-off-by: Dan Carpenter <error27@gmail.com> Applied. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/isdn/hardware/eicon/debuglib.h b/drivers/isdn/hardware/eicon/debuglib.h index 8ea5877..02eed6b 100644 --- a/drivers/isdn/hardware/eicon/debuglib.h +++ b/drivers/isdn/hardware/eicon/debuglib.h @@ -249,7 +249,7 @@ typedef struct _DbgHandle_ } regTime ; /* timestamp for registration */ void *pIrp ; /* ptr to pending i/o request */ unsigned long dbgMask ; /* current debug mask */ - char drvName[16] ; /* ASCII name of registered driver */ + char drvName[128] ; /* ASCII name of registered driver */ char drvTag[64] ; /* revision string */ DbgEnd dbg_end ; /* function for debug closing */ DbgLog dbg_prt ; /* function for debug appending */ diff --git a/drivers/isdn/hardware/eicon/debug.c b/drivers/isdn/hardware/eicon/debug.c index 33ce89e..3626401 100644 --- a/drivers/isdn/hardware/eicon/debug.c +++ b/drivers/isdn/hardware/eicon/debug.c @@ -862,7 +862,7 @@ void diva_mnt_add_xdi_adapter (const DESCRIPTOR* d) { diva_os_spin_lock_magic_t old_irql, old_irql1; dword sec, usec, logical, serial, org_mask; int id, best_id = 0, free_id = -1; - char tmp[256]; + char tmp[128]; diva_dbg_entry_head_t* pmsg = NULL; int len; word size;
In diva_mnt_add_xdi_adapter() we do this: strcpy (clients[id].drvName, tmp); strcpy (clients[id].Dbg.drvName, tmp); The "clients[id].drvName" is a 128 character buffer and "clients[id].Dbg.drvName" was originally a 16 character buffer but I've changed it to 128 as well. We don't actually use 128 characters but we do use more than 16. I've also changed the size of "tmp" to 128 characters instead of 256. Signed-off-by: Dan Carpenter <error27@gmail.com> -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html