[Freedos-kernel] Unable to use drive number 32 with some functions

[Damien Guibouret]

Hello,

I found a small problem in ioctl code that leads to drive number 32 to not be
usable with device related functions (int 0x21, ah=0x44..) and leading to
returning information about another drive without signaling any error. As I
suppose this is not widely used (drive 27 to 32 are most often not known by
users), I think this is quite minor. I wrote a patch, but did not test nor even
compile it. In this patch I keep some code that from interface point of view is
wrong but that is the way I understood the "NDN" related comment.

Regards,

Damien

--- ioctl.c.orig 2015-04-26 21:01:18.000000000 +0200
+++ ioctl.c 2015-04-26 21:26:48.000000000 +0200
@@ -164,6 +164,7 @@
{
struct dpb FAR *dpbp;
unsigned attr;
+ BYTE used_drive;
/*
This line previously returned the deviceheader at r->bl. But,
DOS numbers its drives starting at 1, not 0. A=1, B=2, and so
@@ -173,9 +174,18 @@
*/
/* JPP - changed to use default drive if drive=0 */
/* JT Fixed it */
-
- /* NDN feeds the actual ASCII drive letter to this function */
- dpbp = get_dpb((r->BL & 0x1f) == 0 ? default_drive : (r->BL & 0x1f) - 1);
+ if (r->BL == 0)
+ used_drive = default_drive;
+ else if ((r->BL >= 1) && (r->BL <= 32))
+ used_drive = r->BL - 1;
+ else if (((r->BL >= 'A') && (r->BL <= 'Z')) ||
+ ((r->BL >= 'a') && (r->BL <= 'z')))
+ /* NDN feeds the actual ASCII drive letter to this function */
+ used_drive = (r->BL & 0x1f) - 1;
+ else
+ return DE_INVLDDRV;
+
+ dpbp = get_dpb(used_drive);
if (dpbp)
{
CharReqHdr.r_unit = dpbp->dpb_subunit;
@@ -196,7 +206,7 @@
{
/* note from get_dpb() */
/* that if cdsp == NULL then dev must be NULL too */
- struct cds FAR *cdsp = get_cds1(r->BL & 0x1f);
+ struct cds FAR *cdsp = get_cds(used_drive);
if (cdsp == NULL)
return DE_INVLDDRV;
if (cdsp->cdsFlags & CDSSUBST)

[p***@gmail.com]

Thank you. I will try to get a fix implemented this week. Your patch
corrects supporting the 32nd drive if provided as value 1 thru 32 (0 is
default) but still doesn't handle if passed as ascii value ` (0x60).

Jeremy

On Apr 26, 2015 3:51 PM, "Damien Guibouret" <

Post by Damien Guibouret
Hello,
I found a small problem in ioctl code that leads to drive number 32 to not be
usable with device related functions (int 0x21, ah=0x44..) and leading to
returning information about another drive without signaling any error. As I
suppose this is not widely used (drive 27 to 32 are most often not known by
users), I think this is quite minor. I wrote a patch, but did not test nor even
compile it. In this patch I keep some code that from interface point of view is
wrong but that is the way I understood the "NDN" related comment.
Regards,
Damien
--- ioctl.c.orig 2015-04-26 21:01:18.000000000 +0200
+++ ioctl.c 2015-04-26 21:26:48.000000000 +0200
@@ -164,6 +164,7 @@
{
struct dpb FAR *dpbp;
unsigned attr;
+ BYTE used_drive;
/*
This line previously returned the deviceheader at r->bl. But,
DOS numbers its drives starting at 1, not 0. A=1, B=2, and so
@@ -173,9 +174,18 @@
*/
/* JPP - changed to use default drive if drive=0 */
/* JT Fixed it */
-
- /* NDN feeds the actual ASCII drive letter to this function */
- dpbp = get_dpb((r->BL & 0x1f) == 0 ? default_drive : (r->BL & 0x1f) - 1);
+ if (r->BL == 0)
+ used_drive = default_drive;
+ else if ((r->BL >= 1) && (r->BL <= 32))
+ used_drive = r->BL - 1;
+ else if (((r->BL >= 'A') && (r->BL <= 'Z')) ||
+ ((r->BL >= 'a') && (r->BL <= 'z')))
+ /* NDN feeds the actual ASCII drive letter to this function */
+ used_drive = (r->BL & 0x1f) - 1;
+ else
+ return DE_INVLDDRV;
+
+ dpbp = get_dpb(used_drive);
if (dpbp)
{
CharReqHdr.r_unit = dpbp->dpb_subunit;
@@ -196,7 +206,7 @@
{
/* note from get_dpb() */
/* that if cdsp == NULL then dev must be NULL too */
- struct cds FAR *cdsp = get_cds1(r->BL & 0x1f);
+ struct cds FAR *cdsp = get_cds(used_drive);
if (cdsp == NULL)
return DE_INVLDDRV;
if (cdsp->cdsFlags & CDSSUBST)

------------------------------------------------------------------------------

Post by Damien Guibouret
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Freedos-kernel mailing list
https://lists.sourceforge.net/lists/listinfo/freedos-kernel

[Damien Guibouret]

Hello,

That's right, I was not expecting NDN will use it as it is already out of the spec.
You can change the

else if (((r->BL >= 'A') && (r->BL <= 'Z')) ||
((r->BL >= 'a') && (r->BL <= 'z')))
/* NDN feeds the actual ASCII drive letter to this function */
used_drive = (r->BL & 0x1f) - 1;

into

else if (((r->BL >= 'A') && (r->BL <= '`')) ||
((r->BL >= 'a') && (r->BL <= 'z')))
/* NDN feeds the actual ASCII drive letter to this function */
used_drive = (r->BL - 1) & 0x1f;

I think the lowercase case can remain as it is (there is no correspondance
between lower and upper case for letters above 'z' and more over it goes outside
of 7 bits ASCII range).

Regards,

Damien

Post by p***@gmail.com
Thank you. I will try to get a fix implemented this week. Your patch
corrects supporting the 32nd drive if provided as value 1 thru 32 (0 is
default) but still doesn't handle if passed as ascii value ` (0x60).
Jeremy
On Apr 26, 2015 3:51 PM, "Damien Guibouret" <

Post by Damien Guibouret
Hello,
I found a small problem in ioctl code that leads to drive number 32 to

not be

Post by Damien Guibouret
usable with device related functions (int 0x21, ah=0x44..) and leading to
returning information about another drive without signaling any error. As

I

Post by Damien Guibouret
suppose this is not widely used (drive 27 to 32 are most often not known

by

Post by Damien Guibouret
users), I think this is quite minor. I wrote a patch, but did not test

nor even

Post by Damien Guibouret
compile it. In this patch I keep some code that from interface point of

view is

Post by Damien Guibouret
wrong but that is the way I understood the "NDN" related comment.
Regards,
Damien
--- ioctl.c.orig 2015-04-26 21:01:18.000000000 +0200
+++ ioctl.c 2015-04-26 21:26:48.000000000 +0200
@@ -164,6 +164,7 @@
{
struct dpb FAR *dpbp;
unsigned attr;
+ BYTE used_drive;
/*
This line previously returned the deviceheader at r->bl. But,
DOS numbers its drives starting at 1, not 0. A=1, B=2, and so
@@ -173,9 +174,18 @@
*/
/* JPP - changed to use default drive if drive=0 */
/* JT Fixed it */
-
- /* NDN feeds the actual ASCII drive letter to this function */
- dpbp = get_dpb((r->BL & 0x1f) == 0 ? default_drive : (r->BL &

0x1f) - 1);

Post by Damien Guibouret
+ if (r->BL == 0)
+ used_drive = default_drive;
+ else if ((r->BL >= 1) && (r->BL <= 32))
+ used_drive = r->BL - 1;
+ else if (((r->BL >= 'A') && (r->BL <= 'Z')) ||
+ ((r->BL >= 'a') && (r->BL <= 'z')))
+ /* NDN feeds the actual ASCII drive letter to this function */
+ used_drive = (r->BL & 0x1f) - 1;
+ else
+ return DE_INVLDDRV;
+
+ dpbp = get_dpb(used_drive);
if (dpbp)
{
CharReqHdr.r_unit = dpbp->dpb_subunit;
@@ -196,7 +206,7 @@
{
/* note from get_dpb() */
/* that if cdsp == NULL then dev must be NULL too */
- struct cds FAR *cdsp = get_cds1(r->BL & 0x1f);
+ struct cds FAR *cdsp = get_cds(used_drive);
if (cdsp == NULL)
return DE_INVLDDRV;
if (cdsp->cdsFlags & CDSSUBST)