diff mbox series

[3/6] fms: Add TOC/PMA/ATIP DVD-ROM capabilities

Message ID 20210626211820.107310-4-i.kononenko@yadro.com
State New
Headers show
Series USB-gadget: mass-storage: Support DVD-like images. | expand

Commit Message

i.kononenko June 26, 2021, 9:18 p.m. UTC
The DVD-ROM required the SCSI 6.25 READ TOC/PMA/ATIP Command formats:
 * Response Format 0000b: Formatted TOC
 * Response Format 0001b: Multi-session Information
(MMC-6 Specification).

This patch adds an implementation of that described above formats.

Signed-off-by: Igor Kononenko <i.kononenko@yadro.com>
---
 drivers/usb/gadget/function/f_mass_storage.c |  99 +++++++++--
 drivers/usb/gadget/function/storage_common.c |  10 +-
 include/uapi/linux/cdrom.h                   | 177 +++++++++++++++++++
 3 files changed, 261 insertions(+), 25 deletions(-)

Comments

Alan Stern June 27, 2021, 2:29 p.m. UTC | #1
On Sun, Jun 27, 2021 at 12:18:16AM +0300, Igor Kononenko wrote:
> The DVD-ROM required the SCSI 6.25 READ TOC/PMA/ATIP Command formats:
>  * Response Format 0000b: Formatted TOC
>  * Response Format 0001b: Multi-session Information
> (MMC-6 Specification).
> 
> This patch adds an implementation of that described above formats.

I will sum up the last four patches of this series by saying that they
add features for emulating DVD-ROM and BD devices.  Doing so increases
the size of the f_mass_storage driver by a considerable amount and
also adds a large amount of new material to Jens Axboe's
include/uapi/linux/cdrom.h.

Is any of this really needed?  What usage scenarios require
f_mass_storage to emulate a DVD-ROM that couldn't use f_tcm instead?

Alan Stern
i.kononenko June 27, 2021, 6:45 p.m. UTC | #2
On 27.06.2021 17:29, Alan Stern wrote:
> On Sun, Jun 27, 2021 at 12:18:16AM +0300, Igor Kononenko wrote:
>> The DVD-ROM required the SCSI 6.25 READ TOC/PMA/ATIP Command formats:
>>  * Response Format 0000b: Formatted TOC
>>  * Response Format 0001b: Multi-session Information
>> (MMC-6 Specification).
>>
>> This patch adds an implementation of that described above formats.
> 
> I will sum up the last four patches of this series by saying that they
> add features for emulating DVD-ROM and BD devices.  Doing so increases
> the size of the f_mass_storage driver by a considerable amount and
> also adds a large amount of new material to Jens Axboe's
> include/uapi/linux/cdrom.h.

The `include/uapi/linux/cdrom.h` already includes the definition of the
MMC-(2/3) SCSI data structures, and I believe this uses in many cdrom-api
consumers. The current patchset extends this declaration with additional
structures set and clarifies each. Besides, the described above 
SCSI structures are used to implement a mass_storage SCSI-command handler 
to make implementation clearer and avoid the use of magic constants.

> 
> Is any of this really needed?  What usage scenarios require
> f_mass_storage to emulate a DVD-ROM that couldn't use f_tcm instead?

I can't see any impediments to supplement the already existing 
implementation of MMC-(2/3) specification of multimedia devices to 
represent the DVD/BD features. If the kernel presents the CD-ROM SCSI 
commands, why the mass_storage:usb-gadget-function still doesn't include
that for DVD/BD?

Many modern embedded systems (e.g., BMC, OpenBMC) implements their 
required features, e.g., Virtual Media Device, which is based on the 
usb:gadget:mass-storage. 
The purpose of that features is extensive, and their use the mass-storage
not only as a cdrom-device.

The required features of such systems might expect image back-end files
that size is significant than 2.1Gb, but such medium is not the CD-ROM 
device. USB-gadget consumers can incorrectly interpret such device by 
loading the wrong driver. I believe that should be the DVD-medium device,
at least. 

Additionally, please note the current patch also fixes the incorrect 
implementation of retrieving TOC/PMA/ATIP data, which is required for the 
CD-ROM. One system might correct works with retrieving first with the 
last session together, but for some systems, e.g., OS ESXi, OS Windows, 
should retrieving first and last border sessions in separate SCSI-request. 

> 
> Alan Stern
>
Alan Stern June 28, 2021, 2:31 p.m. UTC | #3
On Sun, Jun 27, 2021 at 09:45:07PM +0300, i.kononenko wrote:
> 
> 
> On 27.06.2021 17:29, Alan Stern wrote:
> > Is any of this really needed?  What usage scenarios require
> > f_mass_storage to emulate a DVD-ROM that couldn't use f_tcm instead?
> 
> I can't see any impediments to supplement the already existing 
> implementation of MMC-(2/3) specification of multimedia devices to 
> represent the DVD/BD features. If the kernel presents the CD-ROM SCSI 
> commands, why the mass_storage:usb-gadget-function still doesn't include
> that for DVD/BD?
> 
> Many modern embedded systems (e.g., BMC, OpenBMC) implements their 
> required features, e.g., Virtual Media Device, which is based on the 
> usb:gadget:mass-storage. 
> The purpose of that features is extensive, and their use the mass-storage
> not only as a cdrom-device.
> 
> The required features of such systems might expect image back-end files
> that size is significant than 2.1Gb, but such medium is not the CD-ROM 
> device. USB-gadget consumers can incorrectly interpret such device by 
> loading the wrong driver. I believe that should be the DVD-medium device,
> at least. 

You should include this information in the patch description, so that 
people will understand why you wrote the patch.

> Additionally, please note the current patch also fixes the incorrect 
> implementation of retrieving TOC/PMA/ATIP data, which is required for the 
> CD-ROM. One system might correct works with retrieving first with the 
> last session together, but for some systems, e.g., OS ESXi, OS Windows, 
> should retrieving first and last border sessions in separate SCSI-request. 

What's wrong with the existing implementation?  Are you talking about 
the do_read_toc function?  The driver only supports one session in any 
case.

In general, fixes to existing code and additions of new code should go 
in separate patches.

Alan Stern
diff mbox series

Patch

diff --git a/drivers/usb/gadget/function/f_mass_storage.c b/drivers/usb/gadget/function/f_mass_storage.c
index c3fddee21b53..4865937799aa 100644
--- a/drivers/usb/gadget/function/f_mass_storage.c
+++ b/drivers/usb/gadget/function/f_mass_storage.c
@@ -223,6 +223,7 @@ 
 #include <linux/usb/composite.h>
 
 #include <linux/nospec.h>
+#include <linux/cdrom.h>
 
 #include "configfs.h"
 
@@ -1319,29 +1320,93 @@  static int do_read_header(struct fsg_common *common, struct fsg_buffhd *bh)
 
 static int do_read_toc(struct fsg_common *common, struct fsg_buffhd *bh)
 {
-	struct fsg_lun	*curlun = common->curlun;
-	int		msf = common->cmnd[1] & 0x02;
-	int		start_track = common->cmnd[6];
-	u8		*buf = (u8 *)bh->buf;
+	struct fsg_lun *curlun = common->curlun;
+	struct cdb_read_toc_pma_atip *cdb =
+		(struct cdb_read_toc_pma_atip *)common->cmnd;
+	struct read_tpa_header *header = (struct read_tpa_header *)bh->buf;
+	struct read_tpa_toc_formatted *data =
+		(struct read_tpa_toc_formatted *)((u8 *)bh->buf +
+						  sizeof(*header));
+	size_t data_size = sizeof(*header);
 
-	if ((common->cmnd[1] & ~0x02) != 0 ||	/* Mask away MSF */
-			start_track > 1) {
+	if (cdb->format == 0) {
+		if (cdb->control == READ_TPA_CTRL_MAGIC_SESS) {
+			LDBG(curlun,
+			    "The MMC-3 specifies format a control byte. Using Multi-Session info\n");
+			cdb->format = CDB_TPA_MULTI_SESS_INFO;
+		}
+		if (cdb->control == READ_TPA_CTRL_MAGIC_RAW) {
+			LDBG(curlun,
+			    "The MMC-3 specifies format a control byte. Using RAW TOC\n");
+			cdb->format = CDB_TPA_RAW_TOC;
+		}
+	}
+
+	/* Currently support response format 0000b: Formatted TOC only */
+	if (cdb->format > CDB_TPA_MULTI_SESS_INFO) {
+		LDBG(curlun, "Unsupported TOC/PMA/ATIP format: %02Xh\n",
+		    cdb->format);
 		curlun->sense_data = SS_INVALID_FIELD_IN_CDB;
 		return -EINVAL;
 	}
 
-	memset(buf, 0, 20);
-	buf[1] = (20-2);		/* TOC data length */
-	buf[2] = 1;			/* First track number */
-	buf[3] = 1;			/* Last track number */
-	buf[5] = 0x16;			/* Data track, copying allowed */
-	buf[6] = 0x01;			/* Only track is number 1 */
-	store_cdrom_address(&buf[8], msf, 0);
+	/*
+	 * We only support one track per disk.
+	 * We also needs to indicate the number of the last track
+	 */
+	if (cdb->number > 1 && cdb->number != READ_TPA_LEADOUT_TRACK) {
+		curlun->sense_data = SS_INVALID_FIELD_IN_CDB;
+		return -EINVAL;
+	}
 
-	buf[13] = 0x16;			/* Lead-out track is data */
-	buf[14] = 0xAA;			/* Lead-out track number */
-	store_cdrom_address(&buf[16], msf, curlun->num_sectors);
-	common->data_size_to_handle = 20;
+	/*
+	 * MULTI-SESSIOIN information must be reported only for first track.
+	 */
+	if (cdb->format == CDB_TPA_MULTI_SESS_INFO && cdb->number > 1) {
+		curlun->sense_data = SS_INVALID_FIELD_IN_CDB;
+		return -EINVAL;
+	}
+
+	memset(header, 0, sizeof(*header));
+	header->n_first_stf = 1;
+	header->n_last_stf = 1;
+
+	memset(data, 0, sizeof(*data));
+	data->addr = 1;
+	data->control = TPA_SECTOR_MODE2_MIXED;
+	data->track_number = cdb->number;
+	data_size += sizeof(*data);
+
+	/*
+	 * We have too case:
+	 *	1)	The request Track Number == 1.
+	 *		We shall set 2 descriptors: First Track, Lead-Out Track
+	 *	2)	The requested Track Number == 0xAA
+	 *		Only Lead-Out descriptor shall be set
+	 */
+	if (cdb->number == 1) {
+		DBG(common, "Fill first track addr\n");
+		store_cdrom_address((u8 *)&data->start_addr_track, cdb->msf, 0);
+
+		data += 1; /* Add one more descriptor */
+		data_size += sizeof(*data);
+		memset(data, 0, sizeof(*data));
+		/* setting the lead-out track info. First part of data*/
+		data->addr = 1;
+		data->control = TPA_SECTOR_MODE2_MIXED;
+		data->track_number = READ_TPA_LEADOUT_TRACK;
+	}
+
+	/*
+	 * Lead-out track must be set anyway.
+	 * If 0xAA Track is requested - the first part of data is already set.
+	 */
+	DBG(common, "Fill last track addr\n");
+	store_cdrom_address((u8 *)&data->start_addr_track,
+				cdb->msf, curlun->num_sectors);
+
+	header->length = cpu_to_be16(data_size - sizeof(header->length));
+	common->data_size_to_handle = data_size;
 	return 0;
 }
 
diff --git a/drivers/usb/gadget/function/storage_common.c b/drivers/usb/gadget/function/storage_common.c
index b859a158a414..b883b8b7b05b 100644
--- a/drivers/usb/gadget/function/storage_common.c
+++ b/drivers/usb/gadget/function/storage_common.c
@@ -24,6 +24,7 @@ 
 #include <linux/file.h>
 #include <linux/fs.h>
 #include <linux/usb/composite.h>
+#include <linux/cdrom.h>
 
 #include "storage_common.h"
 
@@ -295,14 +296,7 @@  void store_cdrom_address(u8 *dest, int msf, u32 addr)
 {
 	if (msf) {
 		/* Convert to Minutes-Seconds-Frames */
-		addr >>= 2;		/* Convert to 2048-byte frames */
-		addr += 2*75;		/* Lead-in occupies 2 seconds */
-		dest[3] = addr % 75;	/* Frames */
-		addr /= 75;
-		dest[2] = addr % 60;	/* Seconds */
-		addr /= 60;
-		dest[1] = addr;		/* Minutes */
-		dest[0] = 0;		/* Reserved */
+		lba_to_msf(addr, &dest[1], &dest[2], &dest[3]);
 	} else {
 		/* Absolute sector */
 		put_unaligned_be32(addr, dest);
diff --git a/include/uapi/linux/cdrom.h b/include/uapi/linux/cdrom.h
index 6c34f6e2f1f7..1d7b4333c3aa 100644
--- a/include/uapi/linux/cdrom.h
+++ b/include/uapi/linux/cdrom.h
@@ -947,4 +947,181 @@  struct rm_feature_desc {
 	__u8 reserved4;
 };
 
+/**
+ * The READ TOC/PMA/ATIP format field values
+ */
+enum cdb_read_tpa_format {
+	/**
+	 * The Track/Session Number field specifies starting track number
+	 * for which the data is returned. For multi-session discs, TOC
+	 * data is returned for all sessions. Track number AAh is reported
+	 * only for the Lead-out area of the last complete session.
+	 */
+	CDB_TPA_FORMATTED_TOC,
+	/**
+	 * This format returns the first complete session number, last
+	 * complete session number and last complete session starting address.
+	 * In this format, the Track/Session Number field is reserved and
+	 * should be set to 00h.
+	 * NOTE: This format provides the Host access to the last closed
+	 * session starting address quickly.
+	 */
+	CDB_TPA_MULTI_SESS_INFO,
+	/**
+	 * This format returns all Q sub-code data in the Lead-In (TOC) areas
+	 * starting from a session number as specified in the Number
+	 * Track/Session Number field.
+	 * In this mode, the Drive shall support Q Sub-channel POINT field
+	 * value of A0h, A1h, A2h, Track numbers, B0h, B1h, B2h, B3h, B4h, C0h,
+	 * and C1h.
+	 * There is no defined LBA addressing and MSF bit shall be set to one.
+	 */
+	CDB_TPA_RAW_TOC,
+	/**
+	 * This format returns Q sub-channel data in the PMA area. In this
+	 * format, the Track/Session Number field is reserved and shall be
+	 * set to 00h. There is no defined LBA addressing and MSF bit
+	 * shall be set to one.
+	 */
+	CDB_TPA_PMA,
+	/**
+	 * This format returns ATIP data. In this format, the Track/Session
+	 * Number field is reserved and shall be set to 00h. There is no
+	 * defined LBA addressing and MSF bit shall be set to one.
+	 */
+	CDB_TPA_ATIP,
+	/**
+	 * This format returns CD-TEXT information that is recorded in the
+	 * Lead-in area as R-W Sub-channel Data.
+	 */
+	CDB_TPA_CD_TEXT,
+};
+
+#define TPA_SECTOR_MODE0		(0x00)
+#define TPA_SECTOR_AUDIO		(0x01)
+#define TPA_SECTOR_MODE1		(0x02)
+#define TPA_SECTOR_MODE2		(0x03)
+#define TPA_SECTOR_MODE2_FORM1		(0x04)
+#define TPA_SECTOR_MODE2_FORM2		(0x05)
+#define TPA_SECTOR_MODE2_MIXED		(TPA_SECTOR_MODE1 | TPA_SECTOR_MODE2_FORM1)
+#define TPA_SECTOR_RAW				(0x07)
+#define TPA_SECTOR_RAW_SCRAMBLED	(0x08)
+
+/**
+ * @brief The READ TOC/PMA/ATIP CDB (43h)
+ * The READ TOC/PMA/ATIP command requests that the Drive read data from a
+ * Table of Contents, the Program Memory Area (PMA), or the Absolute Time
+ * in Pre-Grove (ATIP) from CD media, format according to CDB parameters
+ * and transfer the result to the Host.
+ */
+struct cdb_read_toc_pma_atip {
+	__u8 code;
+
+#if defined(__BIG_ENDIAN_BITFIELD)
+	__u8 reserved1 : 6;
+	/**
+	 * When MSF is set to zero, the address fields in some returned data
+	 * formats shall be in LBA form. When MSF is set to one, the address
+	 * fields in some returned data formats shall be in MSF form
+	 */
+	__u8 msf : 1;
+	__u8 reserved2 : 1;
+#elif defined(__LITTLE_ENDIAN_BITFIELD)
+	__u8 reserved2 : 1;
+	__u8 msf : 1;
+	__u8 reserved1 : 6;
+#endif
+
+#if defined(__BIG_ENDIAN_BITFIELD)
+	__u8 reserved3 : 4;
+	/**
+	 * The Format field is used to select specific returned data format
+	 * according to @enum cdb_read_tpa_format
+	 */
+	__u8 format : 4;
+#elif defined(__LITTLE_ENDIAN_BITFIELD)
+	__u8 format : 4;
+	__u8 reserved3 : 4;
+#endif
+
+	__u8 reserved4[3];
+	/**
+	 * The Track/Session Number field provides a method to restrict the
+	 * returned of some data formats to a specific session or a track range
+	 */
+	__u8 number;
+
+	/**
+	 * The Allocation Length field specifies the maximum number of bytes that
+	 * may be returned by the Drive.
+	 * An Allocation Length field of zero shall not be considered an error
+	 */
+	__be16 length;
+
+	__u8 control;
+} __packed;
+
+#define READ_TPA_LEADOUT_TRACK	(0xAAU)
+/*
+ * Control magic byte
+ * Some legacy media recorder implementations set the control byte,
+ * helping determine the relevant TOC/PMA/ATIP formats.
+ * We should support this as well.
+ */
+#define READ_TPA_CTRL_MAGIC_SESS	(0x40U)
+#define READ_TPA_CTRL_MAGIC_RAW		(0x80U)
+
+/**
+ * @brief READ TOC/PMA/ATIP Data list header
+ * The response data list shows the general description of the response data
+ * to the Read TOC/PMA/ATIP command.
+ */
+struct read_tpa_header {
+	__be16 length;
+	/* First Track/Session/Reserved Field */
+	__u8 n_first_stf;
+	/* Last Track/Session/Reserved Field */
+	__u8 n_last_stf;
+} __packed;
+
+/**
+ * @brief Response Format 0000b: Formatted TOC
+ * The response data consist of four header bytes and zero or more track
+ * descriptors.
+ */
+struct read_tpa_toc_formatted {
+	__u8 reserved1;
+#if defined(__BIG_ENDIAN_BITFIELD)
+	/**
+	 * The ADR field gives the type of information encoded in the Q Sub-channel
+	 * of the block where this TOC entry was found.
+	 */
+	__u8 addr : 4;
+	/**
+	 * The CONTROL Field indicates the attributes of the track.
+	 */
+	__u8 control : 4;
+#elif defined(__LITTLE_ENDIAN_BITFIELD)
+	__u8 control : 4;
+	__u8 addr : 4;
+#endif
+	/**
+	 * The Track Start Address contains the address of the first block with user
+	 * information for that track number as read from the Table of Contents.
+	 * A MSF bit of zero indicates that the Track Start Address field shall contain
+	 * a logical block address.
+	 * A MSF bit of one indicates the Logical Block Address field shall contain a
+	 * MSF address
+	 */
+	__u8 track_number;
+	__u8 reserved2;
+	/**
+	 * The Track Number field indicates the track number for that the data in the
+	 * TOC track descriptor is valid. A track number of READ_TPA_LEADOUT_TRACK
+	 * indicates that the track descriptor is for the start of the Lead-out area.
+	 */
+	__be32 start_addr_track;
+} __packed;
+
+
 #endif /* _UAPI_LINUX_CDROM_H */