Skip to contentSkip to conversation

Commit 9f10d9ee breaks CD mounting/burning

DISCUSSION walt 4 participants 8 messages Mar 6, 2008 – Mar 7, 2008 libata

lore
Share & tools
lore

walt

author
#0 lore raw

Hi Bartolmiej,

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

When I revert this commit everything works properly again, including CD burning.

Here is the drive info:

hdc: TSSTcorpCD/DVDW SH-S182M, ATAPI CD/DVD-ROM drive
hdc: host max PIO5 wanted PIO255(auto-tune) selected PIO4
hdc: UDMA/33 mode selected
hdc: ATAPI 48X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache Uniform CD-ROM driver Revision: 3.20
hdc: UDMA/33 mode selected

I'm happy to try patches or supply more info, just let me know what you need.

Thanks.

#1 lore raw
Quoted 34 lines · walt

Hi Bartolmiej,

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

When I revert this commit everything works properly again, including CD burning.

Here is the drive info:

hdc: TSSTcorpCD/DVDW SH-S182M, ATAPI CD/DVD-ROM drive
hdc: host max PIO5 wanted PIO255(auto-tune) selected PIO4
hdc: UDMA/33 mode selected
hdc: ATAPI 48X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache Uniform CD-ROM driver Revision: 3.20
hdc: UDMA/33 mode selected

Quoted 2 lines

I'm happy to try patches or supply more info, just let me know what you need.

Does the following patch help?

[ Borislav, please take a look and double check that it is OK. ]

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH] ide-cd: mark REQ_TYPE_ATA_PC write requests with REQ_RW flag

Quoted 20 lines · walt

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

Quoted 2 lines

When I revert this commit everything works properly again, including CD burning.

It turned out that REQ_TYPE_ATA_PC write requests were not marked as such (the previous commit assumed them to be).

---
 drivers/ide/ide-cd_ioctl.c |    4 ++++
 1 file changed, 4 insertions(+)
drivers/ide/ide-cd_ioctl.c +4 -0
Index: b/drivers/ide/ide-cd_ioctl.c
===================================================================
@@ -457,6 +457,10 @@ int ide_cdrom_packet(struct cdrom_device
 	   layer. the packet must be complete, as we do not
 	   touch it at all. */
 	ide_cd_init_rq(drive, &req);
+
+	if (cgc->data_direction == CGC_DATA_WRITE)
+		req.cmd_flags |= REQ_RW;
+
 	memcpy(req.cmd, cgc->cmd, CDROM_PACKET_SIZE);
 	if (cgc->sense)
 		memset(cgc->sense, 0, sizeof(struct request_sense));
author
#2 lore raw
Quoted 35 lines · Bartlomiej Zolnierkiewicz
Quoted 9 lines · walt

Hi Bartolmiej,

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz<[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

Does the following patch help?

[ Borislav, please take a look and double check that it is OK. ]

From: Bartlomiej Zolnierkiewicz<[email protected]>
Subject: [PATCH] ide-cd: mark REQ_TYPE_ATA_PC write requests with REQ_RW flag

It turned out that REQ_TYPE_ATA_PC write requests were not marked as such (the previous commit assumed them to be).

---
  drivers/ide/ide-cd_ioctl.c |    4 ++++
  1 file changed, 4 insertions(+)

Index: b/drivers/ide/ide-cd_ioctl.c
===================================================================

drivers/ide/ide-cd_ioctl.c int ide_cdrom_packet(struct cdrom_device · line 466

Quoted 10 lines
drivers/ide/ide-cd_ioctl.c +4 -0
@@ -457,6 +457,10 @@ int ide_cdrom_packet(struct cdrom_device
  	   layer. the packet must be complete, as we do not
  	   touch it at all. */
  	ide_cd_init_rq(drive,&req);
+
+	if (cgc->data_direction == CGC_DATA_WRITE)
+		req.cmd_flags |= REQ_RW;
+
  	memcpy(req.cmd, cgc->cmd, CDROM_PACKET_SIZE);
  	if (cgc->sense)
  		memset(cgc->sense, 0, sizeof(struct request_sense));

It's perfect, thanks!

#3 lore raw
In reply to Bartlomiej Zolnierkiewicz
Quoted 86 lines · Bartlomiej Zolnierkiewicz
Quoted 34 lines · walt

Hi Bartolmiej,

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

When I revert this commit everything works properly again, including CD burning.

Here is the drive info:

hdc: TSSTcorpCD/DVDW SH-S182M, ATAPI CD/DVD-ROM drive
hdc: host max PIO5 wanted PIO255(auto-tune) selected PIO4
hdc: UDMA/33 mode selected
hdc: ATAPI 48X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache Uniform CD-ROM driver Revision: 3.20
hdc: UDMA/33 mode selected

Quoted 2 lines

I'm happy to try patches or supply more info, just let me know what you need.

Does the following patch help?

[ Borislav, please take a look and double check that it is OK. ]

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH] ide-cd: mark REQ_TYPE_ATA_PC write requests with REQ_RW flag

Quoted 20 lines · walt

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

Quoted 2 lines

When I revert this commit everything works properly again, including CD burning.

It turned out that REQ_TYPE_ATA_PC write requests were not marked as such (the previous commit assumed them to be).

---
 drivers/ide/ide-cd_ioctl.c |    4 ++++
 1 file changed, 4 insertions(+)

Index: b/drivers/ide/ide-cd_ioctl.c
===================================================================

drivers/ide/ide-cd_ioctl.c int ide_cdrom_packet(struct cdrom_device · line 466

Quoted 10 lines
drivers/ide/ide-cd_ioctl.c +4 -0
@@ -457,6 +457,10 @@ int ide_cdrom_packet(struct cdrom_device
 	   layer. the packet must be complete, as we do not
 	   touch it at all. */
 	ide_cd_init_rq(drive, &req);
+
+	if (cgc->data_direction == CGC_DATA_WRITE)
+		req.cmd_flags |= REQ_RW;
+
 	memcpy(req.cmd, cgc->cmd, CDROM_PACKET_SIZE);
 	if (cgc->sense)
 		memset(cgc->sense, 0, sizeof(struct request_sense));

Here's how i understand it (and correct me if i'm wrong):

req.cmd_flags & 0x1 (i.e. the least sig. bit) was remaining unset for write requests and that's why the ireason check in the interrupt handler was failing loudly. This sets it correctly so that rq_data_dir(rq) evaluates to the correct data direction of the request, no?

Signature · 2 lines
Regards/Gruß,
    Boris.
#4 lore raw
Quoted 100 lines · Borislav Petkov
Quoted 86 lines · Bartlomiej Zolnierkiewicz
Quoted 34 lines · walt

Hi Bartolmiej,

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

When I revert this commit everything works properly again, including CD burning.

Here is the drive info:

hdc: TSSTcorpCD/DVDW SH-S182M, ATAPI CD/DVD-ROM drive
hdc: host max PIO5 wanted PIO255(auto-tune) selected PIO4
hdc: UDMA/33 mode selected
hdc: ATAPI 48X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache Uniform CD-ROM driver Revision: 3.20
hdc: UDMA/33 mode selected

Quoted 2 lines

I'm happy to try patches or supply more info, just let me know what you need.

Does the following patch help?

[ Borislav, please take a look and double check that it is OK. ]

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH] ide-cd: mark REQ_TYPE_ATA_PC write requests with REQ_RW flag

Quoted 20 lines · walt

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

Quoted 2 lines

When I revert this commit everything works properly again, including CD burning.

It turned out that REQ_TYPE_ATA_PC write requests were not marked as such (the previous commit assumed them to be).

---
 drivers/ide/ide-cd_ioctl.c |    4 ++++
 1 file changed, 4 insertions(+)

Index: b/drivers/ide/ide-cd_ioctl.c
===================================================================

drivers/ide/ide-cd_ioctl.c int ide_cdrom_packet(struct cdrom_device · line 466

Quoted 10 lines
drivers/ide/ide-cd_ioctl.c +4 -0
@@ -457,6 +457,10 @@ int ide_cdrom_packet(struct cdrom_device
 	   layer. the packet must be complete, as we do not
 	   touch it at all. */
 	ide_cd_init_rq(drive, &req);
+
+	if (cgc->data_direction == CGC_DATA_WRITE)
+		req.cmd_flags |= REQ_RW;
+
 	memcpy(req.cmd, cgc->cmd, CDROM_PACKET_SIZE);
 	if (cgc->sense)
 		memset(cgc->sense, 0, sizeof(struct request_sense));

Here's how i understand it (and correct me if i'm wrong):

Quoted 4 lines

req.cmd_flags & 0x1 (i.e. the least sig. bit) was remaining unset for write requests and that's why the ireason check in the interrupt handler was failing loudly. This sets it correctly so that rq_data_dir(rq) evaluates to the correct data direction of the request, no?

Yep.

Thanks for checking it, I'll push the patch to Linus.

#5 lore raw
Quoted 101 lines · Bartlomiej Zolnierkiewicz
Quoted 100 lines · Borislav Petkov
Quoted 86 lines · Bartlomiej Zolnierkiewicz
Quoted 34 lines · walt

Hi Bartolmiej,

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

When I revert this commit everything works properly again, including CD burning.

Here is the drive info:

hdc: TSSTcorpCD/DVDW SH-S182M, ATAPI CD/DVD-ROM drive
hdc: host max PIO5 wanted PIO255(auto-tune) selected PIO4
hdc: UDMA/33 mode selected
hdc: ATAPI 48X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache Uniform CD-ROM driver Revision: 3.20
hdc: UDMA/33 mode selected

Quoted 2 lines

I'm happy to try patches or supply more info, just let me know what you need.

Does the following patch help?

[ Borislav, please take a look and double check that it is OK. ]

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH] ide-cd: mark REQ_TYPE_ATA_PC write requests with REQ_RW flag

Quoted 20 lines · walt

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

Quoted 2 lines

When I revert this commit everything works properly again, including CD burning.

It turned out that REQ_TYPE_ATA_PC write requests were not marked as such (the previous commit assumed them to be).

---
 drivers/ide/ide-cd_ioctl.c |    4 ++++
 1 file changed, 4 insertions(+)

Index: b/drivers/ide/ide-cd_ioctl.c
===================================================================

drivers/ide/ide-cd_ioctl.c int ide_cdrom_packet(struct cdrom_device · line 466

Quoted 10 lines
drivers/ide/ide-cd_ioctl.c +4 -0
@@ -457,6 +457,10 @@ int ide_cdrom_packet(struct cdrom_device
 	   layer. the packet must be complete, as we do not
 	   touch it at all. */
 	ide_cd_init_rq(drive, &req);
+
+	if (cgc->data_direction == CGC_DATA_WRITE)
+		req.cmd_flags |= REQ_RW;
+
 	memcpy(req.cmd, cgc->cmd, CDROM_PACKET_SIZE);
 	if (cgc->sense)
 		memset(cgc->sense, 0, sizeof(struct request_sense));

Here's how i understand it (and correct me if i'm wrong):

Quoted 8 lines
Quoted 4 lines

req.cmd_flags & 0x1 (i.e. the least sig. bit) was remaining unset for write requests and that's why the ireason check in the interrupt handler was failing loudly. This sets it correctly so that rq_data_dir(rq) evaluates to the correct data direction of the request, no?

Yep.

Thanks for checking it, I'll push the patch to Linus.

Cool, can we close http://bugzilla.kernel.org/show_bug.cgi?id=10185 now?

Signature · 2 lines
Regards/Gruß,
    Boris.
#6 lore raw
Quoted 107 lines · Borislav Petkov
Quoted 106 lines · Bartlomiej Zolnierkiewicz
Quoted 104 lines · Borislav Petkov
Quoted 86 lines · Bartlomiej Zolnierkiewicz
Quoted 34 lines · walt

Hi Bartolmiej,

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

When I revert this commit everything works properly again, including CD burning.

Here is the drive info:

hdc: TSSTcorpCD/DVDW SH-S182M, ATAPI CD/DVD-ROM drive
hdc: host max PIO5 wanted PIO255(auto-tune) selected PIO4
hdc: UDMA/33 mode selected
hdc: ATAPI 48X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache Uniform CD-ROM driver Revision: 3.20
hdc: UDMA/33 mode selected

Quoted 2 lines

I'm happy to try patches or supply more info, just let me know what you need.

Does the following patch help?

[ Borislav, please take a look and double check that it is OK. ]

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH] ide-cd: mark REQ_TYPE_ATA_PC write requests with REQ_RW flag

Quoted 20 lines · walt

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

Quoted 2 lines

When I revert this commit everything works properly again, including CD burning.

It turned out that REQ_TYPE_ATA_PC write requests were not marked as such (the previous commit assumed them to be).

---
 drivers/ide/ide-cd_ioctl.c |    4 ++++
 1 file changed, 4 insertions(+)

Index: b/drivers/ide/ide-cd_ioctl.c
===================================================================

drivers/ide/ide-cd_ioctl.c int ide_cdrom_packet(struct cdrom_device · line 466

Quoted 10 lines
drivers/ide/ide-cd_ioctl.c +4 -0
@@ -457,6 +457,10 @@ int ide_cdrom_packet(struct cdrom_device
 	   layer. the packet must be complete, as we do not
 	   touch it at all. */
 	ide_cd_init_rq(drive, &req);
+
+	if (cgc->data_direction == CGC_DATA_WRITE)
+		req.cmd_flags |= REQ_RW;
+
 	memcpy(req.cmd, cgc->cmd, CDROM_PACKET_SIZE);
 	if (cgc->sense)
 		memset(cgc->sense, 0, sizeof(struct request_sense));

Here's how i understand it (and correct me if i'm wrong):

req.cmd_flags & 0x1 (i.e. the least sig. bit) was remaining unset for write requests and that's why the ireason check in the interrupt handler was failing loudly. This sets it correctly so that rq_data_dir(rq) evaluates to the correct data direction of the request, no?

Quoted 5 lines
Quoted 3 lines

Yep.

Thanks for checking it, I'll push the patch to Linus.

Cool, can we close http://bugzilla.kernel.org/show_bug.cgi?id=10185 now?

I suppose that Rafael would prefer to close it when patch hits the mainline.

Thanks,
Bart

#7 lore raw
Quoted 108 lines · Bartlomiej Zolnierkiewicz
Quoted 107 lines · Borislav Petkov
Quoted 106 lines · Bartlomiej Zolnierkiewicz
Quoted 104 lines · Borislav Petkov
Quoted 86 lines · Bartlomiej Zolnierkiewicz
Quoted 34 lines · walt

Hi Bartolmiej,

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

When I revert this commit everything works properly again, including CD burning.

Here is the drive info:

hdc: TSSTcorpCD/DVDW SH-S182M, ATAPI CD/DVD-ROM drive
hdc: host max PIO5 wanted PIO255(auto-tune) selected PIO4
hdc: UDMA/33 mode selected
hdc: ATAPI 48X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache Uniform CD-ROM driver Revision: 3.20
hdc: UDMA/33 mode selected

Quoted 2 lines

I'm happy to try patches or supply more info, just let me know what you need.

Does the following patch help?

[ Borislav, please take a look and double check that it is OK. ]

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH] ide-cd: mark REQ_TYPE_ATA_PC write requests with REQ_RW flag

Quoted 20 lines · walt

For me, this commit causes the problem it's intended to fix:

commit 9f10d9ee0ac6d79d7bc8b9a158bf4a29322d84d3

Author: Bartlomiej Zolnierkiewicz <[email protected]>
Date:   Tue Feb 26 21:50:35 2008 +0100

ide-cd: fix 'ireason' handling for REQ_TYPE_ATA_PC requests

This fixes some hangs caused by not finishing the transfer before ending the request and also makes use of 'ireason == 1' quirk for spurious IRQs.

When I mount a CD there is a long delay, and I see this error message:

hdc: ide_cd_check_ireason: wrong transfer direction!
cdrom: failed setting lba address space
hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }
ide: failed opcode was: unknown
hdc: drive not ready for command
<repeated many times>

Quoted 2 lines

When I revert this commit everything works properly again, including CD burning.

It turned out that REQ_TYPE_ATA_PC write requests were not marked as such (the previous commit assumed them to be).

---
 drivers/ide/ide-cd_ioctl.c |    4 ++++
 1 file changed, 4 insertions(+)

Index: b/drivers/ide/ide-cd_ioctl.c
===================================================================

drivers/ide/ide-cd_ioctl.c int ide_cdrom_packet(struct cdrom_device · line 466

Quoted 10 lines
drivers/ide/ide-cd_ioctl.c +4 -0
@@ -457,6 +457,10 @@ int ide_cdrom_packet(struct cdrom_device
 	   layer. the packet must be complete, as we do not
 	   touch it at all. */
 	ide_cd_init_rq(drive, &req);
+
+	if (cgc->data_direction == CGC_DATA_WRITE)
+		req.cmd_flags |= REQ_RW;
+
 	memcpy(req.cmd, cgc->cmd, CDROM_PACKET_SIZE);
 	if (cgc->sense)
 		memset(cgc->sense, 0, sizeof(struct request_sense));

Here's how i understand it (and correct me if i'm wrong):

req.cmd_flags & 0x1 (i.e. the least sig. bit) was remaining unset for write requests and that's why the ireason check in the interrupt handler was failing loudly. This sets it correctly so that rq_data_dir(rq) evaluates to the correct data direction of the request, no?

Quoted 7 lines
Quoted 5 lines
Quoted 3 lines

Yep.

Thanks for checking it, I'll push the patch to Linus.

Cool, can we close http://bugzilla.kernel.org/show_bug.cgi?id=10185 now?

I suppose that Rafael would prefer to close it when patch hits the mainline.

That's correct.

Thanks,
Rafael