[PATH 0/4] Memstick patches for 2.6.39

[PATH 0/4] Memstick patches for 2.6.39

[Maxim Levitsky]

Hi,

This is a repost of my patches for 2.6.39 inclusion, which I hope not to
miss this time.

I addressed the comments on the scatterlist issues.

Andrew, please note that my richoh memstick driver is standalone, unchanged from previos versions
has many users which use the version I posted at ubuntu's Launchpad and happy with it.
Please include it regardless of other patches.

The other half of my work is support for legacy memorysticks which consists of 2 patches,
first that adds few functions to scatterlist.c, and the other patch that adds the driver.
Driver is also stable and tested.

Best regards,
	Maxim Levitsky

[PATCH 1/4] scatterlist: new helper functions

[Maxim Levitsky]

While developing memstick driver for legacy memsticks
I found the need in few helpers that I think should be
in common scatterlist library

The functions that were added:

* sg_nents/sg_total_len - iterate over scatterlist to figure
out total length of memory it covers / number of entries.

Usefull for small sg lists where there is no prefomance advantage of
storing this info in a special variable.

* sg_copy/sg_truncate - Allow to break scatterlists apart into smaller chunks.
sg_copy creates smaller scatterlist, spanning first 'len' bytes, while
sg_truncate edits the scatterlist in such way that it skips over 'len' bytes.


* sg_compare_to_buffer - another function to hide gory details of access
to sg list by CPU.
Allows to transparently compare contents of the sg list to given linear
buffer.
If needed later, a function that compares two sgs can be added.

All of this code is used by my ms_block.c driver.

Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
---
 include/linux/scatterlist.h |    8 ++
 lib/scatterlist.c           |  152 +++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 160 insertions(+), 0 deletions(-)

diff --git a/include/linux/scatterlist.h b/include/linux/scatterlist.h
index 9aaf5bf..88fc7a5 100644
--- a/include/linux/scatterlist.h
+++ b/include/linux/scatterlist.h
@@ -199,6 +199,12 @@ static inline void *sg_virt(struct scatterlist *sg)
 	return page_address(sg_page(sg)) + sg->offset;
 }
 
+struct scatterlist *sg_truncate(struct scatterlist *sg, int consumed);
+int sg_nents(struct scatterlist *sg);
+int sg_total_len(struct scatterlist *sg);
+int sg_copy(struct scatterlist *sg_from, struct scatterlist *sg_to,
+							int to_nents, int len);
+
 struct scatterlist *sg_next(struct scatterlist *);
 struct scatterlist *sg_last(struct scatterlist *s, unsigned int);
 void sg_init_table(struct scatterlist *, unsigned int);
@@ -217,6 +223,8 @@ size_t sg_copy_from_buffer(struct scatterlist *sgl, unsigned int nents,
 			   void *buf, size_t buflen);
 size_t sg_copy_to_buffer(struct scatterlist *sgl, unsigned int nents,
 			 void *buf, size_t buflen);
+bool sg_compare_to_buffer(struct scatterlist *sg, unsigned int nents,
+							u8 *buffer, size_t len);
 
 /*
  * Maximum number of entries that will be allocated in one piece, if
diff --git a/lib/scatterlist.c b/lib/scatterlist.c
index 4ceb05d..941195d 100644
--- a/lib/scatterlist.c
+++ b/lib/scatterlist.c
@@ -39,6 +39,76 @@ struct scatterlist *sg_next(struct scatterlist *sg)
 EXPORT_SYMBOL(sg_next);
 
 /**
+ * sg_truncate - remove 'consumed' bytes from head of a scatterlist
+ * @sg:		The current sg entry
+ * @consumed:	How much bytes to remove
+ */
+struct scatterlist *sg_truncate(struct scatterlist *sg, int consumed)
+{
+	while (consumed >= sg->length) {
+		consumed -= sg->length;
+
+		sg = sg_next(sg);
+		if (!sg)
+			break;
+	}
+
+	WARN_ON(!sg && consumed);
+
+	if (!sg)
+		return NULL;
+
+	sg->offset += consumed;
+	sg->length -= consumed;
+
+	if (sg->offset >= PAGE_SIZE) {
+		struct page *page =
+			nth_page(sg_page(sg), sg->offset / PAGE_SIZE);
+		sg_set_page(sg, page, sg->length, sg->offset % PAGE_SIZE);
+	}
+
+	return sg;
+}
+EXPORT_SYMBOL(sg_truncate);
+
+/**
+ * sg_nents - calculate number of sg entries in sg list
+ * @sg:		The current sg entry
+ *
+ * Allows to calculate dynamically the length of the sg table, based on
+ * assumption that last entry is correctly marked by sg_mark_end
+ */
+int sg_nents(struct scatterlist *sg)
+{
+	int nents = 0;
+	while (sg) {
+		nents++;
+		sg = sg_next(sg);
+	}
+
+	return nents;
+}
+EXPORT_SYMBOL(sg_nents);
+
+/**
+ * sg_total_len - calculate total length of scatterlist
+ * @sg:		The current sg entry
+ *
+ * Dynamically calculate total number of bytes in an scatterlist
+ * based on assumption that last entry is correctly marked by sg_mark_end
+ */
+int sg_total_len(struct scatterlist *sg)
+{
+	int len = 0;
+	while (sg) {
+		len += sg->length;
+		sg = sg_next(sg);
+	}
+	return len;
+}
+EXPORT_SYMBOL(sg_total_len);
+
+/**
  * sg_last - return the last scatterlist entry in a list
  * @sgl:	First entry in the scatterlist
  * @nents:	Number of entries in the scatterlist
@@ -110,6 +180,47 @@ void sg_init_one(struct scatterlist *sg, const void *buf, unsigned int buflen)
 }
 EXPORT_SYMBOL(sg_init_one);
 
+/**
+ * sg_copy - copies sg entries from sg_from to sg_to, such
+ * as sg_to covers first 'len' bytes from sg_from.
+ * @sg_from:	SG list to copy entries from
+ * @sg_to:	SG list to write entries to
+ * @to_nents:	number of usable entries in 'sg_to'
+ * @len:	maximum number of bytes the 'sg_to' will cover
+ *
+ * Returns actual number of bytes covered by sg_to
+ */
+int sg_copy(struct scatterlist *sg_from, struct scatterlist *sg_to,
+							int to_nents, int len)
+{
+	int copied = 0;
+
+	while (len > sg_from->length && to_nents--) {
+
+		len -= sg_from->length;
+		copied += sg_from->length;
+
+		sg_set_page(sg_to, sg_page(sg_from),
+				sg_from->length, sg_from->offset);
+
+		if (sg_is_last(sg_from) || !len) {
+			sg_mark_end(sg_to);
+			return copied;
+		}
+
+		sg_from = sg_next(sg_from);
+		sg_to = sg_next(sg_to);
+	}
+
+	if (to_nents) {
+		sg_set_page(sg_to, sg_page(sg_from), len, sg_from->offset);
+		sg_mark_end(sg_to);
+	}
+
+	return copied;
+}
+EXPORT_SYMBOL(sg_copy);
+
 /*
  * The default behaviour of sg_alloc_table() is to use these kmalloc/kfree
  * helpers.
@@ -517,3 +628,44 @@ size_t sg_copy_to_buffer(struct scatterlist *sgl, unsigned int nents,
 	return sg_copy_buffer(sgl, nents, buf, buflen, 1);
 }
 EXPORT_SYMBOL(sg_copy_to_buffer);
+
+
+/**
+ * sg_compare_to_buffer - compare contents of the data pointed by sg table
+ * to a kernel ram buffer
+ *
+ * @sg:		The current sg entry
+ * @buffer:	Linear kernel buffer to compare with
+ * @len:	Length of that buffer
+ *
+ * Returns 0 if equal and memcmp compliant result otherwise
+ */
+bool sg_compare_to_buffer(struct scatterlist *sg, unsigned int nents,
+							u8 *buffer, size_t len)
+{
+	unsigned long flags;
+	int retval = 0;
+	struct sg_mapping_iter miter;
+
+	local_irq_save(flags);
+	sg_miter_start(&miter, sg, nents, SG_MITER_ATOMIC | SG_MITER_FROM_SG);
+
+	while (sg_miter_next(&miter) && len > 0) {
+
+		int cmplen = min(miter.length, len);
+		retval = memcmp(miter.addr, buffer, cmplen);
+		if (retval)
+			break;
+
+		buffer += cmplen;
+		len -= cmplen;
+	}
+
+	if (!retval && len)
+		retval = -1;
+
+	sg_miter_stop(&miter);
+	local_irq_restore(flags);
+	return retval;
+}
+EXPORT_SYMBOL(sg_compare_to_buffer);
-- 
1.7.1

Re: [PATCH 1/4] scatterlist: new helper functions

[FUJITA Tomonori]

On Fri,  4 Mar 2011 06:16:50 +0200
Maxim Levitsky <maximlevitsky@gmail.com> wrote:

> While developing memstick driver for legacy memsticks
> I found the need in few helpers that I think should be
> in common scatterlist library
> 
> The functions that were added:
> 
> * sg_nents/sg_total_len - iterate over scatterlist to figure
> out total length of memory it covers / number of entries.

You should invent a data structure per I/O request, something like
msb_request structure. Then you can store nents and total_len in
that.

That's what block subsystems and drivers do. I took a look at your
driver but I can't see why your driver can't do the same.

Re: [PATCH 1/4] scatterlist: new helper functions

[Maxim Levitsky]

On Sun, 2011-03-06 at 16:29 +0900, FUJITA Tomonori wrote:
> On Fri,  4 Mar 2011 06:16:50 +0200
> Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> 
> > While developing memstick driver for legacy memsticks
> > I found the need in few helpers that I think should be
> > in common scatterlist library
> > 
> > The functions that were added:
> > 
> > * sg_nents/sg_total_len - iterate over scatterlist to figure
> > out total length of memory it covers / number of entries.
> 
> You should invent a data structure per I/O request, something like
> msb_request structure. Then you can store nents and total_len in
> that.
> 
> That's what block subsystems and drivers do. I took a look at your
> driver but I can't see why your driver can't do the same.
I also need to break the request into small grained chunks.
If I invent such structure, I will end up writing these helpers for it.

The I have this lifetime of a request:

I get arbitary sized request from block layer (I can of course control
maximum size/number of segments in it, etc).

I break it into eraseblock sized chunks, and for each I translate the
the LBA, into flash address.

Then I break it into flash page sized requests (512 bytes), and yet its
better not to assume that such requests always contained in one sg
entry.

Worse than that, I have to pass an sg list that spans always one 512
page to lowlevel driver, because thats how Alex defined the interface.
 
Thats why I coded it this way.

Folks, really what the status of this, when to expect it to be merged?

If you think some of helper functions don't belong to scatterlist.c,
just tell me to move them back to ms_block.c.

Andrew, please note again that richoh lowlevel driver doesn't need any
helper functions, its patch is standalone and thus should be merged
regardless.

Best regards,
	Maxim Levitsky

Re: [PATCH 1/4] scatterlist: new helper functions

[FUJITA Tomonori]

On Sun, 06 Mar 2011 17:14:30 +0200
Maxim Levitsky <maximlevitsky@gmail.com> wrote:

> On Sun, 2011-03-06 at 16:29 +0900, FUJITA Tomonori wrote:
> > On Fri,  4 Mar 2011 06:16:50 +0200
> > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > 
> > > While developing memstick driver for legacy memsticks
> > > I found the need in few helpers that I think should be
> > > in common scatterlist library
> > > 
> > > The functions that were added:
> > > 
> > > * sg_nents/sg_total_len - iterate over scatterlist to figure
> > > out total length of memory it covers / number of entries.
> > 
> > You should invent a data structure per I/O request, something like
> > msb_request structure. Then you can store nents and total_len in
> > that.
> > 
> > That's what block subsystems and drivers do. I took a look at your
> > driver but I can't see why your driver can't do the same.
> I also need to break the request into small grained chunks.
> If I invent such structure, I will end up writing these helpers for it.
> 
> The I have this lifetime of a request:
> 
> I get arbitary sized request from block layer (I can of course control
> maximum size/number of segments in it, etc).
> 
> I break it into eraseblock sized chunks, and for each I translate the
> the LBA, into flash address.
> 
> Then I break it into flash page sized requests (512 bytes), and yet its
> better not to assume that such requests always contained in one sg
> entry.
> 
> Worse than that, I have to pass an sg list that spans always one 512
> page to lowlevel driver, because thats how Alex defined the interface.

This restriction is due to hardware specification or the software
design (e.g. memstick layer)? If it is due to the latter, why can't
you fix that?

Why can't the block layer split requests for you? It's better to let
the block layer handle that.


> Folks, really what the status of this, when to expect it to be merged?
> 
> If you think some of helper functions don't belong to scatterlist.c,
> just tell me to move them back to ms_block.c.
> 
> Andrew, please note again that richoh lowlevel driver doesn't need any
> helper functions, its patch is standalone and thus should be merged
> regardless.

I think that we need to make the design of the driver easily
understandable to kernel developers and maintainable by them. I don't
think that this is 'standalone or not' issue.

Adding a doc about why the driver is designed in such odd way would be
helpful. But I still think that we could design the driver in a better
way.

Re: [PATCH 1/4] scatterlist: new helper functions

[Maxim Levitsky]

On Mon, 2011-03-07 at 06:49 +0900, FUJITA Tomonori wrote:
> On Sun, 06 Mar 2011 17:14:30 +0200
> Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> 
> > On Sun, 2011-03-06 at 16:29 +0900, FUJITA Tomonori wrote:
> > > On Fri,  4 Mar 2011 06:16:50 +0200
> > > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > 
> > > > While developing memstick driver for legacy memsticks
> > > > I found the need in few helpers that I think should be
> > > > in common scatterlist library
> > > > 
> > > > The functions that were added:
> > > > 
> > > > * sg_nents/sg_total_len - iterate over scatterlist to figure
> > > > out total length of memory it covers / number of entries.
> > > 
> > > You should invent a data structure per I/O request, something like
> > > msb_request structure. Then you can store nents and total_len in
> > > that.
> > > 
> > > That's what block subsystems and drivers do. I took a look at your
> > > driver but I can't see why your driver can't do the same.
> > I also need to break the request into small grained chunks.
> > If I invent such structure, I will end up writing these helpers for it.
> > 
> > The I have this lifetime of a request:
> > 
> > I get arbitary sized request from block layer (I can of course control
> > maximum size/number of segments in it, etc).
> > 
> > I break it into eraseblock sized chunks, and for each I translate the
> > the LBA, into flash address.
> > 
> > Then I break it into flash page sized requests (512 bytes), and yet its
> > better not to assume that such requests always contained in one sg
> > entry.
> > 
> > Worse than that, I have to pass an sg list that spans always one 512
> > page to lowlevel driver, because thats how Alex defined the interface.
> 
> This restriction is due to hardware specification or the software
> design (e.g. memstick layer)? If it is due to the latter, why can't
> you fix that?

Yes.
I already tried addressing some shortcomings of memstick layer, no no, I
don't want to deal with its author, Alex Dubov again.
I think this code tries to be too clever/complex for the range of
devices/speeds it supports, but I rather leave it as is.


To be honest, the code in question is for >5 year old memstick standard
cards, thats hardly anybody uses.
It works, it is more or less simple, its not performance bound, its
testd, and thus I want to keep it as is _for_ now.


Why I break sg lists into chunks?
Because unlike vast majority of block devices, I need to do FTL in the
driver, thus its easier to work on eraseblock boundary.
Also unlike anything else, you can't just read/write a sector from a
memorystick (especially the legacy one), you have to perform full dance
of commands.

Not to mention error handling (like if you failed to write to block, you
must try to choose another one, etc...)

(Of course writes follow same rules as raw nand flash, thats is writes
only clear bits, and you can erase a eraseblock only).



> 
> Why can't the block layer split requests for you? It's better to let
> the block layer handle that.
You mean tell it not to give me more that one eraseblock to handle?
Could you explain that a bit more?


Anyway, could we merge the code?
I would happy to improve it later, but currently merge window is very
close, and the code is more or less agreed upon everyone, and written
more that 1/2 of year ago.

Andrew Morton, could you help me with this, Please?


Best regards,
	Maxim Levitsky

> 
> 
> > Folks, really what the status of this, when to expect it to be merged?
> > 
> > If you think some of helper functions don't belong to scatterlist.c,
> > just tell me to move them back to ms_block.c.
> > 
> > Andrew, please note again that richoh lowlevel driver doesn't need any
> > helper functions, its patch is standalone and thus should be merged
> > regardless.
> 
> I think that we need to make the design of the driver easily
> understandable to kernel developers and maintainable by them. I don't
> think that this is 'standalone or not' issue.
> 
> Adding a doc about why the driver is designed in such odd way would be
> helpful. But I still think that we could design the driver in a better
> way.

Re: [PATCH 1/4] scatterlist: new helper functions

[Maxim Levitsky]

On Mon, 2011-03-07 at 04:20 +0200, Maxim Levitsky wrote:
> On Mon, 2011-03-07 at 06:49 +0900, FUJITA Tomonori wrote:
> > On Sun, 06 Mar 2011 17:14:30 +0200
> > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > 
> > > On Sun, 2011-03-06 at 16:29 +0900, FUJITA Tomonori wrote:
> > > > On Fri,  4 Mar 2011 06:16:50 +0200
> > > > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > > 
> > > > > While developing memstick driver for legacy memsticks
> > > > > I found the need in few helpers that I think should be
> > > > > in common scatterlist library
> > > > > 
> > > > > The functions that were added:
> > > > > 
> > > > > * sg_nents/sg_total_len - iterate over scatterlist to figure
> > > > > out total length of memory it covers / number of entries.
> > > > 
> > > > You should invent a data structure per I/O request, something like
> > > > msb_request structure. Then you can store nents and total_len in
> > > > that.
> > > > 
> > > > That's what block subsystems and drivers do. I took a look at your
> > > > driver but I can't see why your driver can't do the same.
> > > I also need to break the request into small grained chunks.
> > > If I invent such structure, I will end up writing these helpers for it.
> > > 
> > > The I have this lifetime of a request:
> > > 
> > > I get arbitary sized request from block layer (I can of course control
> > > maximum size/number of segments in it, etc).
> > > 
> > > I break it into eraseblock sized chunks, and for each I translate the
> > > the LBA, into flash address.
> > > 
> > > Then I break it into flash page sized requests (512 bytes), and yet its
> > > better not to assume that such requests always contained in one sg
> > > entry.
> > > 
> > > Worse than that, I have to pass an sg list that spans always one 512
> > > page to lowlevel driver, because thats how Alex defined the interface.
> > 
> > This restriction is due to hardware specification or the software
> > design (e.g. memstick layer)? If it is due to the latter, why can't
> > you fix that?
> 
> Yes.
> I already tried addressing some shortcomings of memstick layer, no no, I
> don't want to deal with its author, Alex Dubov again.
> I think this code tries to be too clever/complex for the range of
> devices/speeds it supports, but I rather leave it as is.
> 
> 
> To be honest, the code in question is for >5 year old memstick standard
> cards, thats hardly anybody uses.
> It works, it is more or less simple, its not performance bound, its
> testd, and thus I want to keep it as is _for_ now.
> 
> 
> Why I break sg lists into chunks?
> Because unlike vast majority of block devices, I need to do FTL in the
> driver, thus its easier to work on eraseblock boundary.
> Also unlike anything else, you can't just read/write a sector from a
> memorystick (especially the legacy one), you have to perform full dance
> of commands.
> 
> Not to mention error handling (like if you failed to write to block, you
> must try to choose another one, etc...)
> 
> (Of course writes follow same rules as raw nand flash, thats is writes
> only clear bits, and you can erase a eraseblock only).
> 
> 
> 
> > 
> > Why can't the block layer split requests for you? It's better to let
> > the block layer handle that.
> You mean tell it not to give me more that one eraseblock to handle?
> Could you explain that a bit more?
> 
> 
> Anyway, could we merge the code?
> I would happy to improve it later, but currently merge window is very
> close, and the code is more or less agreed upon everyone, and written
> more that 1/2 of year ago.
> 
> Andrew Morton, could you help me with this, Please?
Ping (I really worry about this).

> 
> 
> Best regards,
> 	Maxim Levitsky
> 
> > 
> > 
> > > Folks, really what the status of this, when to expect it to be merged?
> > > 
> > > If you think some of helper functions don't belong to scatterlist.c,
> > > just tell me to move them back to ms_block.c.
> > > 
> > > Andrew, please note again that richoh lowlevel driver doesn't need any
> > > helper functions, its patch is standalone and thus should be merged
> > > regardless.
> > 
> > I think that we need to make the design of the driver easily
> > understandable to kernel developers and maintainable by them. I don't
> > think that this is 'standalone or not' issue.
> > 
> > Adding a doc about why the driver is designed in such odd way would be
> > helpful. But I still think that we could design the driver in a better
> > way.
> 
> 

Re: [PATCH 1/4] scatterlist: new helper functions

[Andrew Morton]

On Mon, 07 Mar 2011 04:20:37 +0200 Maxim Levitsky <maximlevitsky@gmail.com> wrote:

> On Mon, 2011-03-07 at 06:49 +0900, FUJITA Tomonori wrote:
> > On Sun, 06 Mar 2011 17:14:30 +0200
> > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > 
> > > On Sun, 2011-03-06 at 16:29 +0900, FUJITA Tomonori wrote:
> > > > On Fri,  4 Mar 2011 06:16:50 +0200
> > > > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > > 
> > > > > While developing memstick driver for legacy memsticks
> > > > > I found the need in few helpers that I think should be
> > > > > in common scatterlist library
> > > > > 
> > > > > The functions that were added:
> > > > > 
> > > > > * sg_nents/sg_total_len - iterate over scatterlist to figure
> > > > > out total length of memory it covers / number of entries.
> > > > 
> > > > You should invent a data structure per I/O request, something like
> > > > msb_request structure. Then you can store nents and total_len in
> > > > that.
> > > > 
> > > > That's what block subsystems and drivers do. I took a look at your
> > > > driver but I can't see why your driver can't do the same.
> > > I also need to break the request into small grained chunks.
> > > If I invent such structure, I will end up writing these helpers for it.
> > > 
> > > The I have this lifetime of a request:
> > > 
> > > I get arbitary sized request from block layer (I can of course control
> > > maximum size/number of segments in it, etc).
> > > 
> > > I break it into eraseblock sized chunks, and for each I translate the
> > > the LBA, into flash address.
> > > 
> > > Then I break it into flash page sized requests (512 bytes), and yet its
> > > better not to assume that such requests always contained in one sg
> > > entry.
> > > 
> > > Worse than that, I have to pass an sg list that spans always one 512
> > > page to lowlevel driver, because thats how Alex defined the interface.
> > 
> > This restriction is due to hardware specification or the software
> > design (e.g. memstick layer)? If it is due to the latter, why can't
> > you fix that?
> 
> Yes.
> I already tried addressing some shortcomings of memstick layer, no no, I
> don't want to deal with its author, Alex Dubov again.
> I think this code tries to be too clever/complex for the range of
> devices/speeds it supports, but I rather leave it as is.
> 

I have to say, these aren't very good reasons for a particular
implementation!

> To be honest, the code in question is for >5 year old memstick standard
> cards, thats hardly anybody uses.
> It works, it is more or less simple, its not performance bound, its
> testd, and thus I want to keep it as is _for_ now.
> 
> 
> Why I break sg lists into chunks?
> Because unlike vast majority of block devices, I need to do FTL in the
> driver, thus its easier to work on eraseblock boundary.
> Also unlike anything else, you can't just read/write a sector from a
> memorystick (especially the legacy one), you have to perform full dance
> of commands.
> 
> Not to mention error handling (like if you failed to write to block, you
> must try to choose another one, etc...)
> 
> (Of course writes follow same rules as raw nand flash, thats is writes
> only clear bits, and you can erase a eraseblock only).

hm.  If you think there's little likelihood that other drivers will
need the new sg functions in the future then perhaps they should be
made private to the memstick driver, rather than bloating everyone's
kernels.  Which is, I think, the exact opposite of what I suggested
last year :(

Fujita-san, you've gone all quiet.  Do you believe that these functions
should be added to the sg API?

Thanks.

Re: [PATCH 1/4] scatterlist: new helper functions

[FUJITA Tomonori]

On Tue, 15 Mar 2011 17:44:59 -0700
Andrew Morton <akpm@linux-foundation.org> wrote:

> > > This restriction is due to hardware specification or the software
> > > design (e.g. memstick layer)? If it is due to the latter, why can't
> > > you fix that?
> > 
> > Yes.
> > I already tried addressing some shortcomings of memstick layer, no no, I
> > don't want to deal with its author, Alex Dubov again.
> > I think this code tries to be too clever/complex for the range of
> > devices/speeds it supports, but I rather leave it as is.
> > 
> 
> I have to say, these aren't very good reasons for a particular
> implementation!

Agreed. We need to fix it.


> > To be honest, the code in question is for >5 year old memstick standard
> > cards, thats hardly anybody uses.
> > It works, it is more or less simple, its not performance bound, its
> > testd, and thus I want to keep it as is _for_ now.
> > 
> > 
> > Why I break sg lists into chunks?
> > Because unlike vast majority of block devices, I need to do FTL in the
> > driver, thus its easier to work on eraseblock boundary.
> > Also unlike anything else, you can't just read/write a sector from a
> > memorystick (especially the legacy one), you have to perform full dance
> > of commands.
> > 
> > Not to mention error handling (like if you failed to write to block, you
> > must try to choose another one, etc...)
> > 
> > (Of course writes follow same rules as raw nand flash, thats is writes
> > only clear bits, and you can erase a eraseblock only).
> 
> hm.  If you think there's little likelihood that other drivers will
> need the new sg functions in the future then perhaps they should be
> made private to the memstick driver, rather than bloating everyone's
> kernels.  Which is, I think, the exact opposite of what I suggested
> last year :(
> 
> Fujita-san, you've gone all quiet.  Do you believe that these functions
> should be added to the sg API?

I don't think so.

Splitting (or merging) a request and playing with sg lists inside a
driver is a bad idea. Such should be done in the block layer.

I still don't see why the block layer can't do that for the driver.

I believe that the helper functions for such should not be added.

Re: [PATCH 1/4] scatterlist: new helper functions

[Alex Dubov]

> 
> I don't think so.
> 
> Splitting (or merging) a request and playing with sg lists
> inside a
> driver is a bad idea. Such should be done in the block
> layer.
> 
> I still don't see why the block layer can't do that for the
> driver.
> 
> I believe that the helper functions for such should not be
> added.
> 


In a particular case of flash-like devices, letting the block layer
split requests into memory sequential blocks too often results in
unnecessary fragmentation of writes/erases.

If only one sg entry is requested from the block layer, it will be (more
often than not) only 1 or 2 pages in length, even if total size of
prospective write request spans multiple erase blocks.

So there are really only two options for legacy memorystick driver:
1. Play with scatterlists explicitly.
2. Make it an MTD backend, rather then stand-alone block device.

The second option makes more sense, but it is not necessarily the optimal
approach for implementation of this particular media format.



      

Re: [PATCH 1/4] scatterlist: new helper functions

[FUJITA Tomonori]

On Tue, 15 Mar 2011 21:18:14 -0700 (PDT)
Alex Dubov <oakad@yahoo.com> wrote:

> 
> > 
> > I don't think so.
> > 
> > Splitting (or merging) a request and playing with sg lists
> > inside a
> > driver is a bad idea. Such should be done in the block
> > layer.
> > 
> > I still don't see why the block layer can't do that for the
> > driver.
> > 
> > I believe that the helper functions for such should not be
> > added.
> > 
> 
> 
> In a particular case of flash-like devices, letting the block layer
> split requests into memory sequential blocks too often results in
> unnecessary fragmentation of writes/erases.

Why?

Why can't the block layer split a request in the way the driver wants
to do?

That is, why can't the driver tell the block layer how to split a
request?


> If only one sg entry is requested from the block layer, it will be (more
> often than not) only 1 or 2 pages in length, even if total size of
> prospective write request spans multiple erase blocks.

In this case, what does the driver do? Why can't the block layer do
the same?


> So there are really only two options for legacy memorystick driver:
> 1. Play with scatterlists explicitly.
> 2. Make it an MTD backend, rather then stand-alone block device.
> 
> The second option makes more sense, but it is not necessarily the optimal
> approach for implementation of this particular media format.
> 
> 
> 
>       
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 1/4] scatterlist: new helper functions

[Alex Dubov]

> 
> Why?
> 
> Why can't the block layer split a request in the way the
> driver wants
> to do?
> 
> That is, why can't the driver tell the block layer how to
> split a
> request?

What is needed is the ability to get fixed sized (in bytes) blocks from 
the block layer.

Last time I checked (it was a long time ago, admittedly) one could only
ask for a fixed number of sg entries, without any control on how many
bytes each sg entry references.

Is there a way to get data from the block layer in a fashion of:
"Give me 16k/32k/whatever in one sg entry if request is equal or larger
than this"?

If my knowledge is correct, MTD currently addresses this issue by
maintaining its own cache, which it uses to aggregate write requests until
it can write a whole erase block. While this is ok with old media
(legacy memory stick being an example of such), new flash chips can
have multi-megabyte sized erase blocks and can benefit from operations
(like copy and compare) directly on scatter list .



      

Re: [PATCH 1/4] scatterlist: new helper functions

[James Bottomley]

On Wed, 2011-03-16 at 20:46 -0700, Alex Dubov wrote:
> > 
> > Why?
> > 
> > Why can't the block layer split a request in the way the
> > driver wants
> > to do?
> > 
> > That is, why can't the driver tell the block layer how to
> > split a
> > request?
> 
> What is needed is the ability to get fixed sized (in bytes) blocks from 
> the block layer.
> 
> Last time I checked (it was a long time ago, admittedly) one could only
> ask for a fixed number of sg entries, without any control on how many
> bytes each sg entry references.
> 
> Is there a way to get data from the block layer in a fashion of:
> "Give me 16k/32k/whatever in one sg entry if request is equal or larger
> than this"?

Yes; we've always had it: it's blk_max_queue_hw_sectors().  No request
will go over that number times the sector size (of course, they may go
under).

> If my knowledge is correct, MTD currently addresses this issue by
> maintaining its own cache, which it uses to aggregate write requests until
> it can write a whole erase block. While this is ok with old media
> (legacy memory stick being an example of such), new flash chips can
> have multi-megabyte sized erase blocks and can benefit from operations
> (like copy and compare) directly on scatter list .

So this is where you want a minimum too.  What you likely want is to set
the logical block size to your erase block
(blk_queue_logical_block_size) and the physical block size to the actual
block size.  Then we'll try as hard as we can to send down blocks on an
erase boundary.  Of course, there are some that just won't fit (like fs
metadata) and you'll have to do a RMW for them.

James

Re: [PATCH 1/4] scatterlist: new helper functions

[Andrew Morton]

On Tue, 15 Mar 2011 21:18:14 -0700 (PDT)
Alex Dubov <oakad@yahoo.com> wrote:

> 
> > 
> > I don't think so.
> > 
> > Splitting (or merging) a request and playing with sg lists
> > inside a
> > driver is a bad idea. Such should be done in the block
> > layer.
> > 
> > I still don't see why the block layer can't do that for the
> > driver.
> > 
> > I believe that the helper functions for such should not be
> > added.
> > 
> 
> 
> In a particular case of flash-like devices, letting the block layer
> split requests into memory sequential blocks too often results in
> unnecessary fragmentation of writes/erases.
> 
> If only one sg entry is requested from the block layer, it will be (more
> often than not) only 1 or 2 pages in length, even if total size of
> prospective write request spans multiple erase blocks.
> 
> So there are really only two options for legacy memorystick driver:
> 1. Play with scatterlists explicitly.
> 2. Make it an MTD backend, rather then stand-alone block device.
> 
> The second option makes more sense, but it is not necessarily the optimal
> approach for implementation of this particular media format.
> 

Thanks.

Do you believe that any other driver is likely to ue the sg
infrastructure which this patch adds?  Should those additions be 
internal to the memstick code, at least for now?

Re: [PATCH 1/4] scatterlist: new helper functions

[Alex Dubov]

> 
> Do you believe that any other driver is likely to ue the
> sg
> infrastructure which this patch adds?  Should those
> additions be 
> internal to the memstick code, at least for now?
> 

Considering the comments from Fujita and James, I suppose it will be best
for now to fold the new sg functions into the memstick patch and
investigate how the block layer functionality can be used to a greater
utility at some later time.


      

Re: [PATCH 1/4] scatterlist: new helper functions

[Maxim Levitsky]

On Wed, 2011-03-16 at 23:41 -0700, Alex Dubov wrote:
> > 
> > Do you believe that any other driver is likely to ue the
> > sg
> > infrastructure which this patch adds?  Should those
> > additions be 
> > internal to the memstick code, at least for now?
> > 
> 
> Considering the comments from Fujita and James, I suppose it will be best
> for now to fold the new sg functions into the memstick patch and
> investigate how the block layer functionality can be used to a greater
> utility at some later time.

I am thinking the same thing.
If the unusual sg code is a problem, I'll think of something later to
improve it, heck I can even completely rip it off and just use kernel
pointers, relying on block core to bounce for high-mem situations.
Since its a driver for a legacy hardware, mostly provided for
completeness, the performance difference shouldn't be much.


I think its ok to merge the code with folded sg code _for_ now, and then
I will send patches to address this point you didn't like.


OK?

Best regards,
	Maxim Levitsky 

Re: [PATCH 1/4] scatterlist: new helper functions

[FUJITA Tomonori]

On Wed, 16 Mar 2011 23:41:13 -0700 (PDT)
Alex Dubov <oakad@yahoo.com> wrote:

> Considering the comments from Fujita and James, I suppose it will be best
> for now to fold the new sg functions into the memstick patch and
> investigate how the block layer functionality can be used to a greater
> utility at some later time.

I suspect that such 'some later time' never come (especially for old
hardware). So I would prefer to merge the properly designed driver
rather than the current hacky driver.

[PATCH 2/4] memstick: add support for legacy memorysticks

[Maxim Levitsky]

Based partially on MS standard spec quotes from Alex Dubov.

As any code that works with user data this driver isn't
recommended to use to write cards that contain valuable data.

It tries its best though to avoid data corruption
and possible damage to the card.

Tested on MS DUO 64 MB card on Ricoh R592 card reader.

Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
---
 MAINTAINERS                      |    5 +
 drivers/memstick/core/Kconfig    |   12 +
 drivers/memstick/core/Makefile   |    2 +-
 drivers/memstick/core/ms_block.c | 2288 ++++++++++++++++++++++++++++++++++++++
 drivers/memstick/core/ms_block.h |  245 ++++
 5 files changed, 2551 insertions(+), 1 deletions(-)
 create mode 100644 drivers/memstick/core/ms_block.c
 create mode 100644 drivers/memstick/core/ms_block.h

diff --git a/MAINTAINERS b/MAINTAINERS
index 8afba63..0269107 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -5799,6 +5799,11 @@ W:	http://tifmxx.berlios.de/
 S:	Maintained
 F:	drivers/memstick/host/tifm_ms.c
 
+SONY MEMORYSTICK STANDARD SUPPORT
+M:	Maxim Levitsky <maximlevitsky@gmail.com>
+S:	Maintained
+F:	drivers/memstick/core/ms_block.*
+
 SOUND
 M:	Jaroslav Kysela <perex@perex.cz>
 M:	Takashi Iwai <tiwai@suse.de>
diff --git a/drivers/memstick/core/Kconfig b/drivers/memstick/core/Kconfig
index 95f1814..f79f2a8 100644
--- a/drivers/memstick/core/Kconfig
+++ b/drivers/memstick/core/Kconfig
@@ -24,3 +24,15 @@ config MSPRO_BLOCK
 	  support. This provides a block device driver, which you can use
 	  to mount the filesystem. Almost everyone wishing MemoryStick
 	  support should say Y or M here.
+
+config MS_BLOCK
+	tristate "MemoryStick Standard device driver"
+	depends on BLOCK && EXPERIMENTAL
+	help
+	  Say Y here to enable the MemoryStick Standard device driver
+	  support. This provides a block device driver, which you can use
+	  to mount the filesystem.
+	  This driver works with old (bulky) MemoryStick and MemoryStick Duo
+	  but not PRO. Say Y if you have such card.
+	  Driver is new and not yet well tested, thus it can damage your card
+	  (even permanently)
diff --git a/drivers/memstick/core/Makefile b/drivers/memstick/core/Makefile
index 8b2b529..19d960b 100644
--- a/drivers/memstick/core/Makefile
+++ b/drivers/memstick/core/Makefile
@@ -7,5 +7,5 @@ ifeq ($(CONFIG_MEMSTICK_DEBUG),y)
 endif
 
 obj-$(CONFIG_MEMSTICK)		+= memstick.o
-
+obj-$(CONFIG_MS_BLOCK)		+= ms_block.o
 obj-$(CONFIG_MSPRO_BLOCK)	+= mspro_block.o
diff --git a/drivers/memstick/core/ms_block.c b/drivers/memstick/core/ms_block.c
new file mode 100644
index 0000000..97ebf90
--- /dev/null
+++ b/drivers/memstick/core/ms_block.c
@@ -0,0 +1,2288 @@
+/*
+ *  ms_block.c - Sony MemoryStick (legacy) storage support
+
+ *  Copyright (C) 2010 Maxim Levitsky <maximlevitsky@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * Minor portions of the driver were copied from mspro_block.c which is
+ * Copyright (C) 2007 Alex Dubov <oakad@yahoo.com>
+ *
+ */
+
+#include <linux/blkdev.h>
+#include <linux/mm.h>
+#include <linux/idr.h>
+#include <linux/hdreg.h>
+#include <linux/kthread.h>
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/random.h>
+#include <linux/memstick.h>
+#include <linux/bitmap.h>
+#include <linux/scatterlist.h>
+#include <linux/sched.h>
+#include <linux/jiffies.h>
+#include "ms_block.h"
+
+static int major;
+static int debug;
+static int cache_flush_timeout = 1000;
+static bool verify_writes;
+
+/* Get zone at which block with logical address 'lba' lives
+ * Flash is broken into zones.
+ * Each zone consists of 512 eraseblocks, out of which in first
+ * zone 494 are used and 496 are for all following zones.
+ * Therefore zone #0 hosts blocks 0-493, zone #1 blocks 494-988, etc...
+*/
+static int msb_get_zone_from_lba(int lba)
+{
+	if (lba < 494)
+		return 0;
+	return ((lba - 494) / 496) + 1;
+}
+
+/* Get zone of physical block. Trivial */
+static int msb_get_zone_from_pba(int pba)
+{
+	return pba / MS_BLOCKS_IN_ZONE;
+}
+
+/* Debug test to validate free block counts */
+#ifdef DEBUG
+static int msb_validate_used_block_bitmap(struct msb_data *msb)
+{
+	int total_free_blocks = 0;
+	int i;
+
+	for (i = 0 ; i < msb->zone_count ; i++)
+		total_free_blocks += msb->free_block_count[i];
+
+	if (msb->block_count - bitmap_weight(msb->used_blocks_bitmap,
+					msb->block_count) == total_free_blocks)
+		return 0;
+
+	ms_printk("BUG: free block counts don't match the bitmap");
+	msb->read_only = true;
+	return -EINVAL;
+}
+#endif
+
+/* Mark physical block as used */
+static void msb_mark_block_used(struct msb_data *msb, int pba)
+{
+	int zone = msb_get_zone_from_pba(pba);
+
+	if (test_bit(pba, msb->used_blocks_bitmap)) {
+		ms_printk("BUG: attempt to mark "
+			"already used pba %d as used", pba);
+		msb->read_only = true;
+		return;
+	}
+
+#ifdef DEBUG
+	if (msb_validate_used_block_bitmap(msb))
+		return;
+#endif
+	set_bit(pba, msb->used_blocks_bitmap);
+	msb->free_block_count[zone]--;
+}
+
+/* Mark physical block as free */
+static void msb_mark_block_unused(struct msb_data *msb, int pba)
+{
+	int zone = msb_get_zone_from_pba(pba);
+
+	if (!test_bit(pba, msb->used_blocks_bitmap)) {
+		ms_printk("BUG: attempt to mark "
+				"already unused pba %d as unused" , pba);
+		msb->read_only = true;
+		return;
+	}
+
+#ifdef DEBUG
+	if (msb_validate_used_block_bitmap(msb))
+		return;
+#endif
+	clear_bit(pba, msb->used_blocks_bitmap);
+	msb->free_block_count[zone]++;
+}
+
+/* Invalidate current register window*/
+static void msb_invalidate_reg_window(struct msb_data *msb)
+{
+	msb->reg_addr.w_offset = offsetof(struct ms_register, id);
+	msb->reg_addr.w_length = sizeof(struct ms_id_register);
+	msb->reg_addr.r_offset = offsetof(struct ms_register, id);
+	msb->reg_addr.r_length = sizeof(struct ms_id_register);
+	msb->addr_valid = false;
+}
+
+/* Sane way to start a state machine*/
+static int msb_run_state_machine(struct msb_data *msb, int   (*state_func)
+		(struct memstick_dev *card, struct memstick_request **req))
+{
+	struct memstick_dev *card = msb->card;
+
+	WARN_ON(msb->state != -1);
+	msb->int_polling = false;
+	msb->state = 0;
+	msb->exit_error = 0;
+
+	memset(&card->current_mrq, 0, sizeof(card->current_mrq));
+
+	card->next_request = state_func;
+	memstick_new_req(card->host);
+	wait_for_completion(&card->mrq_complete);
+
+	WARN_ON(msb->state != -1);
+	return msb->exit_error;
+}
+
+/* State machines call that to exit */
+int msb_exit_state_machine(struct msb_data *msb, int error)
+{
+	WARN_ON(msb->state == -1);
+
+	msb->state = -1;
+	msb->exit_error = error;
+	msb->card->next_request = h_msb_default_bad;
+
+	/* Invalidate reg window on errors */
+	if (error)
+		msb_invalidate_reg_window(msb);
+
+	complete(&msb->card->mrq_complete);
+	return -ENXIO;
+}
+
+/* read INT register */
+int msb_read_int_reg(struct msb_data *msb, long timeout)
+{
+	struct memstick_request *mrq = &msb->card->current_mrq;
+	WARN_ON(msb->state == -1);
+
+	if (!msb->int_polling) {
+		msb->int_timeout = jiffies +
+			msecs_to_jiffies(timeout == -1 ? 500 : timeout);
+		msb->int_polling = true;
+	} else if (time_after(jiffies, msb->int_timeout)) {
+		mrq->data[0] = MEMSTICK_INT_CMDNAK;
+		return 0;
+	}
+
+	if ((msb->caps & MEMSTICK_CAP_AUTO_GET_INT) &&
+				mrq->need_card_int && !mrq->error) {
+		mrq->data[0] = mrq->int_reg;
+		mrq->need_card_int = false;
+		return 0;
+	} else {
+		memstick_init_req(mrq, MS_TPC_GET_INT, NULL, 1);
+		return 1;
+	}
+}
+
+/* Read a register */
+int msb_read_regs(struct msb_data *msb, int offset, int len)
+{
+	struct memstick_request *req = &msb->card->current_mrq;
+
+	if (msb->reg_addr.r_offset != offset ||
+	    msb->reg_addr.r_length != len || !msb->addr_valid) {
+
+		msb->reg_addr.r_offset = offset;
+		msb->reg_addr.r_length = len;
+		msb->addr_valid = true;
+
+		memstick_init_req(req, MS_TPC_SET_RW_REG_ADRS,
+			&msb->reg_addr, sizeof(msb->reg_addr));
+		return 0;
+	}
+
+	memstick_init_req(req, MS_TPC_READ_REG, NULL, len);
+	return 1;
+}
+
+/* Write a card register */
+int msb_write_regs(struct msb_data *msb, int offset, int len, void *buf)
+{
+	struct memstick_request *req = &msb->card->current_mrq;
+
+	if (msb->reg_addr.w_offset != offset ||
+		msb->reg_addr.w_length != len  || !msb->addr_valid) {
+
+		msb->reg_addr.w_offset = offset;
+		msb->reg_addr.w_length = len;
+		msb->addr_valid = true;
+
+		memstick_init_req(req, MS_TPC_SET_RW_REG_ADRS,
+			&msb->reg_addr, sizeof(msb->reg_addr));
+		return 0;
+	}
+
+	memstick_init_req(req, MS_TPC_WRITE_REG, buf, len);
+	return 1;
+}
+
+/* Handler for absense of IO */
+static int h_msb_default_bad(struct memstick_dev *card,
+						struct memstick_request **mrq)
+{
+	return -ENXIO;
+}
+
+/*
+ * This function is a handler for reads of one page from device.
+ * Writes output to msb->current_sg, takes sector address from msb->reg.param
+ * Can also be used to read extra data only. Set params accordintly.
+ */
+static int h_msb_read_page(struct memstick_dev *card,
+					struct memstick_request **out_mrq)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	struct memstick_request *mrq = *out_mrq = &card->current_mrq;
+	u8 command, intreg;
+
+	if (mrq->error) {
+		dbg("read_page, unknown error");
+		return msb_exit_state_machine(msb, mrq->error);
+	}
+again:
+	switch (msb->state) {
+	case 0: /* Write the sector address */
+		if (!msb_write_regs(msb,
+			offsetof(struct ms_register, param),
+			sizeof(struct ms_param_register),
+			(unsigned char *)&msb->regs.param))
+			return 0;
+		break;
+
+	case 1: /* Execute the read command*/
+		command = MS_CMD_BLOCK_READ;
+		memstick_init_req(mrq, MS_TPC_SET_CMD, &command, 1);
+		break;
+
+	case 2: /* send INT request */
+		if (msb_read_int_reg(msb, -1))
+			break;
+		msb->state++;
+
+	case 3: /* get result of the INT request*/
+		intreg = mrq->data[0];
+		msb->regs.status.interrupt = intreg;
+
+		if (intreg & MEMSTICK_INT_CMDNAK)
+			return msb_exit_state_machine(msb, -EIO);
+
+		if (!(intreg & MEMSTICK_INT_CED)) {
+			msb->state--;
+			goto again;
+		}
+
+		msb->int_polling = false;
+
+		if (intreg & MEMSTICK_INT_ERR)
+			msb->state++;
+		else
+			msb->state = 6;
+
+		goto again;
+
+	case 4: /* read the status register
+				to understand source of the INT_ERR */
+		if (!msb_read_regs(msb,
+			offsetof(struct ms_register, status),
+			sizeof(struct ms_status_register)))
+			return 0;
+		break;
+
+	case 5: /* get results of status check */
+		msb->regs.status = *(struct ms_status_register *)mrq->data;
+		msb->state++;
+
+	case 6: /* Send extra data read request */
+		if (!msb_read_regs(msb,
+			offsetof(struct ms_register, extra_data),
+			sizeof(struct ms_extra_data_register)))
+			return 0;
+		break;
+
+	case 7: /* Save result of extra data request */
+		msb->regs.extra_data =
+			*(struct ms_extra_data_register *) mrq->data;
+		msb->state++;
+
+	case 8: /* Send the  MS_TPC_READ_LONG_DATA to read IO buffer */
+
+		/* Skip that state if we only read the oob */
+		if (msb->regs.param.cp == MEMSTICK_CP_EXTRA) {
+			msb->state++;
+			goto again;
+		}
+
+		memstick_init_req_sg(mrq, MS_TPC_READ_LONG_DATA, msb->current_sg);
+		break;
+
+	case 9: /* check validity of data buffer & done */
+
+		if (!(msb->regs.status.interrupt & MEMSTICK_INT_ERR))
+			return msb_exit_state_machine(msb, 0);
+
+		if (msb->regs.status.status1 & MEMSTICK_UNCORR_ERROR) {
+			dbg("read_page: uncorrectable error");
+			return msb_exit_state_machine(msb, -EBADMSG);
+		}
+
+		if (msb->regs.status.status1 & MEMSTICK_CORR_ERROR) {
+			dbg("read_page: correctable error");
+			return msb_exit_state_machine(msb, -EUCLEAN);
+		} else {
+			dbg("read_page: INT error, but no status error bits");
+			return msb_exit_state_machine(msb, -EIO);
+		}
+	default:
+		BUG();
+	}
+	msb->state++;
+	return 0;
+}
+
+/*
+ * Handler of writes of exactly one block.
+ * Takes address from msb->regs.param.
+ * Writes same extra data to blocks, also taken
+ * from msb->regs.extra
+ * Returns -EBADMSG if write fails due to uncorrectable error, or -EIO if
+ * device refuses to take the command or something else
+ */
+static int h_msb_write_block(struct memstick_dev *card,
+					struct memstick_request **out_mrq)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	struct memstick_request *mrq = *out_mrq = &card->current_mrq;
+	struct scatterlist sg[2];
+	u8 intreg, command;
+
+	if (mrq->error)
+		return msb_exit_state_machine(msb, mrq->error);
+
+again:
+	switch (msb->state) {
+
+	/* HACK: Jmicon handling of TPCs between 8 and
+	 *	sizeof(memstick_request.data) is broken due to hardware
+	 *	bug in PIO mode that is used for these TPCs
+	 *	Therefore split the write
+	 */
+
+	case 0: /* write param register*/
+		if (!msb_write_regs(msb,
+			offsetof(struct ms_register, param),
+			sizeof(struct ms_param_register),
+			&msb->regs.param))
+			return 0;
+		break;
+
+	case 1: /* write extra data */
+		if (!msb_write_regs(msb,
+			offsetof(struct ms_register, extra_data),
+			sizeof(struct ms_extra_data_register),
+			&msb->regs.extra_data))
+			return 0;
+		break;
+
+
+	case 2: /* execute the write command*/
+		command = MS_CMD_BLOCK_WRITE;
+		memstick_init_req(mrq, MS_TPC_SET_CMD, &command, 1);
+		break;
+
+	case 3: /* send INT request */
+		if (msb_read_int_reg(msb, -1))
+			break;
+		msb->state++;
+
+	case 4: /* read INT response */
+		intreg = mrq->data[0];
+		msb->regs.status.interrupt = intreg;
+
+		/* errors mean out of here, and fast... */
+		if (intreg & (MEMSTICK_INT_CMDNAK))
+			return msb_exit_state_machine(msb, -EIO);
+
+		if (intreg & MEMSTICK_INT_ERR)
+			return msb_exit_state_machine(msb, -EBADMSG);
+
+
+		/* for last page we need to poll CED */
+		if (msb->current_page == msb->pages_in_block) {
+			if (intreg & MEMSTICK_INT_CED)
+				return msb_exit_state_machine(msb, 0);
+			msb->state--;
+			goto again;
+
+		}
+
+		/* for non-last page we need BREQ before writing next chunk */
+		if (!(intreg & MEMSTICK_INT_BREQ)) {
+			msb->state--;
+			goto again;
+		}
+
+		msb->int_polling = false;
+		msb->state++;
+
+	case 5: /* send the MS_TPC_WRITE_LONG_DATA to perform the write*/
+		sg_init_table(sg, ARRAY_SIZE(sg));
+		sg_copy(msb->current_sg, sg, ARRAY_SIZE(sg), msb->page_size);
+		memstick_init_req_sg(mrq, MS_TPC_WRITE_LONG_DATA, sg);
+		mrq->need_card_int = 1;
+		break;
+
+	case 6: /* Switch to next page + go back to int polling */
+		msb->current_page++;
+
+		if (msb->current_page < msb->pages_in_block) {
+			msb->current_sg = sg_truncate(msb->current_sg, msb->page_size);
+
+			if (!msb->current_sg) {
+				ms_printk(
+				"BUG: out of data while writing block!");
+				return msb_exit_state_machine(msb, -EFAULT);
+			}
+		}
+		msb->state = 3;
+		goto again;
+	default:
+		BUG();
+	}
+	msb->state++;
+	return 0;
+}
+
+/*
+ * This function is used to send simple IO requests to device that consist
+ * of register write + command
+ */
+static int h_msb_send_command(struct memstick_dev *card,
+					struct memstick_request **out_mrq)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	struct memstick_request *mrq = *out_mrq = &card->current_mrq;
+
+	u8 intreg;
+
+	if (mrq->error) {
+		dbg("send_command: unknown error");
+		return msb_exit_state_machine(msb, mrq->error);
+	}
+again:
+	switch (msb->state) {
+
+		/* HACK: see h_msb_write_block */
+
+	case 0: /* write param register*/
+		if (!msb_write_regs(msb,
+			offsetof(struct ms_register, param),
+			sizeof(struct ms_param_register),
+			&msb->regs.param))
+			return 0;
+		break;
+
+	case 1: /* write extra data */
+		if (!msb->command_need_oob) {
+			msb->state++;
+			goto again;
+		}
+
+		if (!msb_write_regs(msb,
+			offsetof(struct ms_register, extra_data),
+			sizeof(struct ms_extra_data_register),
+			&msb->regs.extra_data))
+			return 0;
+		break;
+
+	case 2: /* execute the command*/
+		memstick_init_req(mrq, MS_TPC_SET_CMD, &msb->command_value, 1);
+		break;
+
+	case 3: /* send INT request */
+		if (msb_read_int_reg(msb, -1))
+			break;
+		msb->state++;
+
+	case 4: /* poll for int bits */
+		intreg = mrq->data[0];
+
+		if (intreg & MEMSTICK_INT_CMDNAK)
+			return msb_exit_state_machine(msb, -EIO);
+		if (intreg & MEMSTICK_INT_ERR)
+			return msb_exit_state_machine(msb, -EBADMSG);
+
+
+		if (!(intreg & MEMSTICK_INT_CED)) {
+			msb->state--;
+			goto again;
+		}
+
+		return msb_exit_state_machine(msb, 0);
+	}
+	msb->state++;
+	return 0;
+}
+
+/* Small handler for card reset */
+static int h_msb_reset(struct memstick_dev *card,
+					struct memstick_request **out_mrq)
+{
+	u8 command = MS_CMD_RESET;
+	struct msb_data *msb = memstick_get_drvdata(card);
+	struct memstick_request *mrq = *out_mrq = &card->current_mrq;
+
+	if (mrq->error)
+		return msb_exit_state_machine(msb, mrq->error);
+
+	switch (msb->state) {
+	case 0:
+		memstick_init_req(mrq, MS_TPC_SET_CMD, &command, 1);
+		mrq->need_card_int = 0;
+		break;
+	case 1:
+		return msb_exit_state_machine(msb, 0);
+	}
+	msb->state++;
+	return 0;
+}
+
+/* This handler is used to do serial->parallel switch */
+static int h_msb_parallel_switch(struct memstick_dev *card,
+					struct memstick_request **out_mrq)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	struct memstick_request *mrq = *out_mrq = &card->current_mrq;
+
+	struct memstick_host *host = card->host;
+
+	if (mrq->error) {
+		dbg("parallel_switch: error");
+		msb->regs.param.system &= ~MEMSTICK_SYS_PAM;
+		return msb_exit_state_machine(msb, mrq->error);
+	}
+
+	switch (msb->state) {
+	case 0: /* Set the parallel interface on memstick side */
+		msb->regs.param.system |= MEMSTICK_SYS_PAM;
+
+		if (!msb_write_regs(msb,
+			offsetof(struct ms_register, param),
+			1,
+			(unsigned char *)&msb->regs.param))
+			return 0;
+		break;
+
+	case 1: /* Set parallel interface on our side + send a dummy request
+			to see if card responds */
+		host->set_param(host, MEMSTICK_INTERFACE, MEMSTICK_PAR4);
+		memstick_init_req(mrq, MS_TPC_GET_INT, NULL, 1);
+		break;
+
+	case 2:
+		return msb_exit_state_machine(msb, 0);
+	}
+	msb->state++;
+	return 0;
+}
+
+static int msb_switch_to_parallel(struct msb_data *msb);
+
+/* Reset the card, to guard against hw errors beeing treated as bad blocks */
+static int msb_reset(struct msb_data *msb, bool full)
+{
+
+	bool was_parallel = msb->regs.param.system & MEMSTICK_SYS_PAM;
+	struct memstick_dev *card = msb->card;
+	struct memstick_host *host = card->host;
+	int error;
+
+	/* Reset the card */
+	msb->regs.param.system = MEMSTICK_SYS_BAMD;
+
+	if (full) {
+		error =  host->set_param(host, MEMSTICK_POWER, MEMSTICK_POWER_OFF);
+		if (error)
+			goto out_error;
+
+		msb_invalidate_reg_window(msb);
+
+		error = host->set_param(host, MEMSTICK_POWER, MEMSTICK_POWER_ON);
+		if (error)
+			goto out_error;
+
+		error = host->set_param(host, MEMSTICK_INTERFACE, MEMSTICK_SERIAL);
+		if (error) {
+out_error:
+			dbg("Failed to reset the host controller");
+			msb->read_only = true;
+			return -EFAULT;
+		}
+	}
+
+	error = msb_run_state_machine(msb, h_msb_reset);
+	if (error) {
+		dbg("Failed to reset the card");
+		msb->read_only = true;
+		return -ENODEV;
+	}
+
+	/* Set parallel mode */
+	if (was_parallel)
+		msb_switch_to_parallel(msb);
+	return 0;
+}
+
+/* Attempts to switch interface to parallel mode */
+static int msb_switch_to_parallel(struct msb_data *msb)
+{
+	int error;
+
+	error = msb_run_state_machine(msb, h_msb_parallel_switch);
+	if (error) {
+		ms_printk("Switch to parallel failed");
+		msb->regs.param.system &= ~MEMSTICK_SYS_PAM;
+		msb_reset(msb, true);
+		return -EFAULT;
+	}
+
+	msb->caps |= MEMSTICK_CAP_AUTO_GET_INT;
+	return 0;
+}
+
+/* Changes overwrite flag on a page */
+static int msb_set_overwrite_flag(struct msb_data *msb,
+						u16 pba, u8 page, u8 flag)
+{
+	if (msb->read_only)
+		return -EROFS;
+
+	msb->regs.param.block_address = cpu_to_be16(pba);
+	msb->regs.param.page_address = page;
+	msb->regs.param.cp = MEMSTICK_CP_OVERWRITE;
+	msb->regs.extra_data.overwrite_flag = flag;
+	msb->command_value = MS_CMD_BLOCK_WRITE;
+	msb->command_need_oob = true;
+
+	dbg_verbose("changing overwrite flag to %02x for sector %d, page %d",
+							flag, pba, page);
+	return msb_run_state_machine(msb, h_msb_send_command);
+}
+
+static int msb_mark_bad(struct msb_data *msb, int pba)
+{
+	ms_printk("marking pba %d as bad", pba);
+	msb_reset(msb, true);
+	return msb_set_overwrite_flag(
+			msb, pba, 0, 0xFF & ~MEMSTICK_OVERWRITE_BKST);
+}
+
+static int msb_mark_page_bad(struct msb_data *msb, int pba, int page)
+{
+	dbg("marking page %d of pba %d as bad", page, pba);
+	msb_reset(msb, true);
+	return msb_set_overwrite_flag(msb,
+		pba, page, ~MEMSTICK_OVERWRITE_PGST0);
+}
+
+/* Erases one physical block */
+static int msb_erase_block(struct msb_data *msb, u16 pba)
+{
+	int error, try;
+	if (msb->read_only)
+		return -EROFS;
+
+	dbg_verbose("erasing pba %d", pba);
+
+	for (try = 1 ; try < 3 ; try++) {
+		msb->regs.param.block_address = cpu_to_be16(pba);
+		msb->regs.param.page_address = 0;
+		msb->regs.param.cp = MEMSTICK_CP_BLOCK;
+		msb->command_value = MS_CMD_BLOCK_ERASE;
+		msb->command_need_oob = false;
+
+
+		error = msb_run_state_machine(msb, h_msb_send_command);
+		if (!error || msb_reset(msb, true))
+			break;
+	}
+
+	if (error) {
+		ms_printk("erase failed, marking pba %d as bad", pba);
+		msb_mark_bad(msb, pba);
+	}
+
+	dbg_verbose("erase success, marking pba %d as unused", pba);
+	msb_mark_block_unused(msb, pba);
+	set_bit(pba, msb->erased_blocks_bitmap);
+	return error;
+}
+
+/* Reads one page from device */
+static int msb_read_page(struct msb_data *msb,
+	u16 pba, u8 page, struct ms_extra_data_register *extra,
+						struct scatterlist *sg)
+{
+	int try, error;
+
+	if (sg && sg->length < msb->page_size) {
+		ms_printk(
+			"BUG: attempt to read pba %d page %d with too small sg",
+								pba, page);
+		return -EINVAL;
+	}
+
+	if (pba == MS_BLOCK_INVALID) {
+		unsigned long flags;
+		struct sg_mapping_iter miter;
+		size_t len = msb->page_size;
+
+		dbg_verbose("read unmapped sector. returning 0xFF");
+
+		local_irq_save(flags);
+		sg_miter_start(&miter, sg, sg_nents(sg),
+				SG_MITER_ATOMIC | SG_MITER_TO_SG);
+
+		while (sg_miter_next(&miter) && len > 0) {
+			int chunklen = min(miter.length, len);
+			memset(miter.addr, 0xFF, chunklen);
+			len -= chunklen;
+		}
+
+		sg_miter_stop(&miter);
+		local_irq_restore(flags);
+
+		if (extra)
+			memset(extra, 0xFF, sizeof(*extra));
+		return 0;
+	}
+
+	if (pba >= msb->block_count) {
+		ms_printk("BUG: attempt to read beyond"
+					" the end of the card at pba %d", pba);
+		return -EINVAL;
+	}
+
+	for (try = 1 ; try < 3 ; try++) {
+		msb->regs.param.block_address = cpu_to_be16(pba);
+		msb->regs.param.page_address = page;
+		msb->regs.param.cp = MEMSTICK_CP_PAGE;
+
+		msb->current_sg = msb->sg;
+		sg_init_table(msb->current_sg, MS_BLOCK_MAX_SEGS+1);
+		sg_copy(sg, msb->current_sg, MS_BLOCK_MAX_SEGS+1,
+							msb->page_size);
+		error = msb_run_state_machine(msb, h_msb_read_page);
+
+
+		if (error == -EUCLEAN) {
+			ms_printk("correctable error on pba %d, page %d",
+				pba, page);
+			error = 0;
+		}
+
+		if (!error && extra)
+			*extra = msb->regs.extra_data;
+
+		if (!error || msb_reset(msb, true))
+			break;
+
+	}
+
+	/* Mark bad pages */
+	if (error == -EBADMSG) {
+		ms_printk("uncorrectable error on read of pba %d, page %d",
+			pba, page);
+
+		if (msb->regs.extra_data.overwrite_flag &
+					MEMSTICK_OVERWRITE_PGST0)
+			msb_mark_page_bad(msb, pba, page);
+		return -EBADMSG;
+	}
+
+	if (error)
+		ms_printk("read of pba %d, page %d failed with error %d",
+			pba, page, error);
+	return error;
+}
+
+/* Reads oob of page only */
+static int msb_read_oob(struct msb_data *msb, u16 pba, u16 page,
+	struct ms_extra_data_register *extra)
+{
+	int error;
+	BUG_ON(!extra);
+
+	msb->regs.param.block_address = cpu_to_be16(pba);
+	msb->regs.param.page_address = page;
+	msb->regs.param.cp = MEMSTICK_CP_EXTRA;
+
+	if (pba > msb->block_count) {
+		ms_printk("BUG: attempt to read beyond"
+					" the end of card at pba %d", pba);
+		return -EINVAL;
+	}
+
+	error = msb_run_state_machine(msb, h_msb_read_page);
+	*extra = msb->regs.extra_data;
+
+	if (error == -EUCLEAN) {
+		ms_printk("correctable error on pba %d, page %d",
+			pba, page);
+		return 0;
+	}
+
+	return error;
+}
+
+
+/* Reads a block and compares it with data contained in scatterlist orig_sg */
+static bool msb_verify_block(struct msb_data *msb, u16 pba,
+					struct scatterlist *orig_sg)
+{
+	struct scatterlist sg;
+	int page = 0, error;
+
+	while (page < msb->pages_in_block) {
+		sg_init_one(&sg, msb->block_buffer +
+					page * msb->page_size, msb->page_size);
+
+		error = msb_read_page(msb, pba, page, NULL, &sg);
+		if (error)
+			return -EIO;
+		page++;
+	}
+
+	if (sg_compare_to_buffer(orig_sg, sg_nents(orig_sg),
+				msb->block_buffer, msb->block_size))
+		return -EIO;
+	return 0;
+}
+
+/* Writes exectly one block + oob */
+static int msb_write_block(struct msb_data *msb,
+			u16 pba, u32 lba, struct scatterlist *sg)
+{
+	int error, current_try = 1;
+	BUG_ON(sg->length < msb->page_size);
+
+	if (msb->read_only)
+		return -EROFS;
+
+	if (sg_total_len(sg) < msb->block_size) {
+		ms_printk("BUG: write: sg underrrun");
+		return -EINVAL;
+	}
+
+	if (pba == MS_BLOCK_INVALID) {
+		ms_printk(
+			"BUG: write: attempt to write MS_BLOCK_INVALID block");
+		return -EINVAL;
+	}
+
+	if (pba >= msb->block_count || lba >= msb->logical_block_count) {
+		ms_printk(
+		"BUG: write: attempt to write beyond the end of device");
+		return -EINVAL;
+	}
+
+	if (msb_get_zone_from_lba(lba) != msb_get_zone_from_pba(pba)) {
+		ms_printk("BUG: write: lba zone mismatch");
+		return -EINVAL;
+	}
+
+	if (pba == msb->boot_block_locations[0] ||
+		pba == msb->boot_block_locations[1]) {
+		ms_printk("BUG: write: attempt to write to boot blocks!");
+		return -EINVAL;
+	}
+
+	while (1) {
+
+		if (msb->read_only)
+			return -EROFS;
+
+		msb->regs.param.cp = MEMSTICK_CP_BLOCK;
+		msb->regs.param.page_address = 0;
+		msb->regs.param.block_address = cpu_to_be16(pba);
+
+		msb->regs.extra_data.management_flag = 0xFF;
+		msb->regs.extra_data.overwrite_flag = 0xF8;
+		msb->regs.extra_data.logical_address = cpu_to_be16(lba);
+
+		msb->current_sg = msb->sg;
+		sg_init_table(msb->current_sg, MS_BLOCK_MAX_SEGS+1);
+		sg_copy(sg, msb->current_sg, MS_BLOCK_MAX_SEGS+1,
+							msb->block_size);
+		msb->current_page = 0;
+
+		error = msb_run_state_machine(msb, h_msb_write_block);
+
+		/* Sector we just wrote to is assumed erased since its pba
+			was erased. If it wasn't erased, write will succeed
+			and will just clear the bits that were set in the block
+			thus test that what we have written,
+			matches what we expect.
+			We do trust the blocks that we erased */
+		if (!error && (verify_writes ||
+				!test_bit(pba, msb->erased_blocks_bitmap)))
+			error = msb_verify_block(msb, pba, sg);
+
+		if (!error)
+			break;
+
+		if (current_try > 1 || msb_reset(msb, true))
+			break;
+
+		ms_printk("write failed, trying to erase the pba %d", pba);
+		error = msb_erase_block(msb, pba);
+		if (error)
+			break;
+
+		current_try++;
+	}
+	return error;
+}
+
+/* Finds a free block for write replacement */
+static u16 msb_get_free_block(struct msb_data *msb, int zone)
+{
+	u16 pos;
+	int pba = zone * MS_BLOCKS_IN_ZONE;
+	int i;
+
+	get_random_bytes(&pos, sizeof(pos));
+
+	if (!msb->free_block_count[zone]) {
+		ms_printk("NO free blocks in the zone %d, to use for a write, "
+			"(media is WORN out) switching to RO mode", zone);
+		msb->read_only = true;
+		return MS_BLOCK_INVALID;
+	}
+
+	pos %= msb->free_block_count[zone];
+
+	dbg_verbose("have %d choices for a free block, selected randomally: %d",
+		msb->free_block_count[zone], pos);
+
+	pba = find_next_zero_bit(msb->used_blocks_bitmap,
+							msb->block_count, pba);
+	for (i = 0 ; i < pos ; ++i)
+		pba = find_next_zero_bit(msb->used_blocks_bitmap,
+						msb->block_count, pba + 1);
+
+	dbg_verbose("result of the free blocks scan: pba %d", pba);
+
+	if (pba == msb->block_count || (msb_get_zone_from_pba(pba)) != zone) {
+		ms_printk("BUG: cant get a free block");
+		msb->read_only = true;
+		return MS_BLOCK_INVALID;
+	}
+
+	msb_mark_block_used(msb, pba);
+	return pba;
+}
+
+static int msb_update_block(struct msb_data *msb, u16 lba,
+	struct scatterlist *sg)
+{
+	u16 pba, new_pba;
+	int error, try;
+
+	pba = msb->lba_to_pba_table[lba];
+	dbg_verbose("start of a block update at lba  %d, pba %d", lba, pba);
+
+	if (pba != MS_BLOCK_INVALID) {
+		dbg_verbose("setting the update flag on the block");
+		msb_set_overwrite_flag(msb, pba, 0,
+				0xFF & ~MEMSTICK_OVERWRITE_UDST);
+	}
+
+	for (try = 0 ; try < 3 ; try++) {
+		new_pba = msb_get_free_block(msb,
+			msb_get_zone_from_lba(lba));
+
+		if (new_pba == MS_BLOCK_INVALID) {
+			error = -EIO;
+			goto out;
+		}
+
+		dbg_verbose("block update: writing updated block to the pba %d",
+								new_pba);
+		error = msb_write_block(msb, new_pba, lba, sg);
+		if (error == -EBADMSG) {
+			msb_mark_bad(msb, new_pba);
+			continue;
+		}
+
+		if (error)
+			goto out;
+
+		dbg_verbose("block update: erasing the old block");
+		msb_erase_block(msb, pba);
+		msb->lba_to_pba_table[lba] = new_pba;
+		return 0;
+	}
+out:
+	if (error) {
+		ms_printk("block update error after %d tries, "
+						"switching to r/o mode", try);
+		msb->read_only = true;
+	}
+	return error;
+}
+
+/* Converts endiannes in the boot block for easy use */
+static void msb_fix_boot_page_endianness(struct ms_boot_page *p)
+{
+	p->header.block_id = be16_to_cpu(p->header.block_id);
+	p->header.format_reserved = be16_to_cpu(p->header.format_reserved);
+	p->entry.disabled_block.start_addr
+		= be32_to_cpu(p->entry.disabled_block.start_addr);
+	p->entry.disabled_block.data_size
+		= be32_to_cpu(p->entry.disabled_block.data_size);
+	p->entry.cis_idi.start_addr
+		= be32_to_cpu(p->entry.cis_idi.start_addr);
+	p->entry.cis_idi.data_size
+		= be32_to_cpu(p->entry.cis_idi.data_size);
+	p->attr.block_size = be16_to_cpu(p->attr.block_size);
+	p->attr.number_of_blocks = be16_to_cpu(p->attr.number_of_blocks);
+	p->attr.number_of_effective_blocks
+		= be16_to_cpu(p->attr.number_of_effective_blocks);
+	p->attr.page_size = be16_to_cpu(p->attr.page_size);
+	p->attr.memory_manufacturer_code
+		= be16_to_cpu(p->attr.memory_manufacturer_code);
+	p->attr.memory_device_code = be16_to_cpu(p->attr.memory_device_code);
+	p->attr.implemented_capacity
+		= be16_to_cpu(p->attr.implemented_capacity);
+	p->attr.controller_number = be16_to_cpu(p->attr.controller_number);
+	p->attr.controller_function = be16_to_cpu(p->attr.controller_function);
+}
+
+static int msb_read_boot_blocks(struct msb_data *msb)
+{
+	int pba = 0;
+	struct scatterlist sg;
+	struct ms_extra_data_register extra;
+	struct ms_boot_page *page;
+
+	msb->boot_block_locations[0] = MS_BLOCK_INVALID;
+	msb->boot_block_locations[1] = MS_BLOCK_INVALID;
+	msb->boot_block_count = 0;
+
+	dbg_verbose("Start of a scan for the boot blocks");
+
+	if (!msb->boot_page) {
+		page = kmalloc(sizeof(struct ms_boot_page)*2, GFP_KERNEL);
+		if (!page)
+			return -ENOMEM;
+
+		msb->boot_page = page;
+	}
+
+	msb->block_count = MS_BLOCK_MAX_BOOT_ADDR;
+
+	for (pba = 0 ; pba < MS_BLOCK_MAX_BOOT_ADDR ; pba++) {
+
+		sg_init_one(&sg, page, sizeof(*page));
+		if (msb_read_page(msb, pba, 0, &extra, &sg)) {
+			dbg("boot scan: can't read pba %d", pba);
+			continue;
+		}
+
+		if (extra.management_flag & MEMSTICK_MANAGEMENT_SYSFLG) {
+			dbg("managment flag doesn't indicate boot block %d",
+									pba);
+			continue;
+		}
+
+		if (be16_to_cpu(page->header.block_id) != MS_BLOCK_BOOT_ID) {
+			dbg("the pba at %d doesn' contain boot block ID", pba);
+			continue;
+		}
+
+		msb_fix_boot_page_endianness(page);
+		msb->boot_block_locations[msb->boot_block_count] = pba;
+
+		page++;
+		msb->boot_block_count++;
+
+		if (msb->boot_block_count == 2)
+			break;
+	}
+
+	if (!msb->boot_block_count) {
+		ms_printk("media doesn't contain master page, aborting");
+		return -EIO;
+	}
+
+	dbg_verbose("End of scan for boot blocks");
+	return 0;
+}
+
+static int msb_read_bad_block_table(struct msb_data *msb, int block_nr)
+{
+	struct ms_boot_page *boot_block;
+	struct scatterlist sg;
+	struct scatterlist *sg_ptr = &sg;
+	u16 *buffer = NULL;
+
+	int i, error = 0;
+	int data_size, data_offset, page, page_offset, size_to_read;
+	u16 pba;
+
+	BUG_ON(block_nr > 1);
+
+	boot_block = &msb->boot_page[block_nr];
+	pba = msb->boot_block_locations[block_nr];
+
+	if (msb->boot_block_locations[block_nr] == MS_BLOCK_INVALID)
+		return -EINVAL;
+
+	data_size = boot_block->entry.disabled_block.data_size;
+	data_offset = sizeof(struct ms_boot_page) +
+			boot_block->entry.disabled_block.start_addr;
+	if (!data_size)
+		return 0;
+
+	page = data_offset / msb->page_size;
+	page_offset = data_offset % msb->page_size;
+	size_to_read =
+		DIV_ROUND_UP(data_size + page_offset, msb->page_size) *
+			msb->page_size;
+
+	dbg("reading bad block of boot block at pba %d, offset %d len %d",
+		pba, data_offset, data_size);
+
+	buffer = kzalloc(size_to_read, GFP_KERNEL);
+	if (!buffer)
+		return -ENOMEM;
+
+	/* Read the buffer */
+	sg_init_one(&sg, buffer, size_to_read);
+
+	while (sg_ptr) {
+		error = msb_read_page(msb, pba, page, NULL, sg_ptr);
+		if (error)
+			goto out;
+
+		sg_ptr = sg_truncate(sg_ptr, msb->page_size);
+		page++;
+		if (page == msb->pages_in_block) {
+			ms_printk(
+			"bad block table extends beyond the boot block");
+			break;
+		}
+	}
+
+	/* Process the bad block table */
+	for (i = page_offset ; i < data_size / sizeof(u16) ; i++) {
+
+		u16 bad_block = be16_to_cpu(buffer[i]);
+
+		if (bad_block >= msb->block_count) {
+			dbg("bad block table contains invalid block %d",
+								bad_block);
+			continue;
+		}
+
+		if (test_bit(bad_block, msb->used_blocks_bitmap))  {
+			dbg("duplicate bad block %d in the table",
+				bad_block);
+			continue;
+		}
+
+		dbg("block %d is marked as factory bad", bad_block);
+		msb_mark_block_used(msb, bad_block);
+	}
+out:
+	kfree(buffer);
+	return error;
+}
+
+static int msb_ftl_initialize(struct msb_data *msb)
+{
+	int i;
+
+	if (msb->ftl_initialized)
+		return 0;
+
+	msb->zone_count = msb->block_count / MS_BLOCKS_IN_ZONE;
+	msb->logical_block_count = msb->zone_count * 496 - 2;
+
+	msb->used_blocks_bitmap = kzalloc(msb->block_count / 8, GFP_KERNEL);
+	msb->erased_blocks_bitmap = kzalloc(msb->block_count / 8, GFP_KERNEL);
+	msb->lba_to_pba_table =
+		kmalloc(msb->logical_block_count * sizeof(u16), GFP_KERNEL);
+
+	if (!msb->used_blocks_bitmap || !msb->lba_to_pba_table ||
+						!msb->erased_blocks_bitmap) {
+		kfree(msb->used_blocks_bitmap);
+		kfree(msb->lba_to_pba_table);
+		kfree(msb->erased_blocks_bitmap);
+		return -ENOMEM;
+	}
+
+	for (i = 0 ; i < msb->zone_count ; i++)
+		msb->free_block_count[i] = MS_BLOCKS_IN_ZONE;
+
+	memset(msb->lba_to_pba_table, MS_BLOCK_INVALID,
+			msb->logical_block_count * sizeof(u16));
+
+	dbg("initial FTL tables created. Zone count = %d, "
+					"Logical block count = %d",
+		msb->zone_count, msb->logical_block_count);
+
+	msb->ftl_initialized = true;
+	return 0;
+}
+
+static int msb_ftl_scan(struct msb_data *msb)
+{
+	u16 pba, lba, other_block;
+	u8 overwrite_flag, managment_flag, other_overwrite_flag;
+	int error;
+	struct ms_extra_data_register extra;
+	u8 *overwrite_flags = kzalloc(msb->block_count, GFP_KERNEL);
+
+	if (!overwrite_flags)
+		return -ENOMEM;
+
+	dbg("Start of media scanning");
+	for (pba = 0 ; pba < msb->block_count ; pba++) {
+
+		if (pba == msb->boot_block_locations[0] ||
+			pba == msb->boot_block_locations[1]) {
+			dbg_verbose("pba %05d -> [boot block]", pba);
+			msb_mark_block_used(msb, pba);
+			continue;
+		}
+
+		if (test_bit(pba, msb->used_blocks_bitmap)) {
+			dbg_verbose("pba %05d -> [factory bad]", pba);
+			continue;
+		}
+
+		error = msb_read_oob(msb, pba, 0, &extra);
+
+		/* can't trust the page if we can't read the oob */
+		if (error == -EBADMSG) {
+			ms_printk(
+			"oob of pba %d damaged, will try to erase it", pba);
+			msb_mark_block_used(msb, pba);
+			msb_erase_block(msb, pba);
+			continue;
+		} else if (error)
+			return error;
+
+		lba = be16_to_cpu(extra.logical_address);
+		managment_flag = extra.management_flag;
+		overwrite_flag = extra.overwrite_flag;
+		overwrite_flags[pba] = overwrite_flag;
+
+		/* Skip bad blocks */
+		if (!(overwrite_flag & MEMSTICK_OVERWRITE_BKST)) {
+			dbg("pba %05d -> [BAD]", pba);
+			msb_mark_block_used(msb, pba);
+			continue;
+		}
+
+		/* Skip system/drm blocks */
+		if ((managment_flag & MEMSTICK_MANAGMENT_FLAG_NORMAL) !=
+			MEMSTICK_MANAGMENT_FLAG_NORMAL) {
+			dbg("pba %05d -> [reserved managment flag %02x]",
+							pba, managment_flag);
+			msb_mark_block_used(msb, pba);
+			continue;
+		}
+
+		/* Erase temporary tables */
+		if (!(managment_flag & MEMSTICK_MANAGEMENT_ATFLG)) {
+			dbg("pba %05d -> [temp table] - will erase", pba);
+
+			msb_mark_block_used(msb, pba);
+			msb_erase_block(msb, pba);
+			continue;
+		}
+
+		if (lba == MS_BLOCK_INVALID) {
+			dbg_verbose("pba %05d -> [free]", pba);
+			continue;
+		}
+
+		msb_mark_block_used(msb, pba);
+
+		/* Block has LBA not according to zoning*/
+		if (msb_get_zone_from_lba(lba) != msb_get_zone_from_pba(pba)) {
+			ms_printk("pba %05d -> [bad lba %05d] - will erase",
+								pba, lba);
+			msb_erase_block(msb, pba);
+			continue;
+		}
+
+		/* No collisions - great */
+		if (msb->lba_to_pba_table[lba] == MS_BLOCK_INVALID) {
+			dbg_verbose("pba %05d -> [lba %05d]", pba, lba);
+			msb->lba_to_pba_table[lba] = pba;
+			continue;
+		}
+
+		other_block = msb->lba_to_pba_table[lba];
+		other_overwrite_flag = overwrite_flags[other_block];
+
+		ms_printk("Collision between pba %d and pba %d",
+			pba, other_block);
+
+		if (!(overwrite_flag & MEMSTICK_OVERWRITE_UDST)) {
+			ms_printk("pba %d is marked as stable, use it", pba);
+			msb_erase_block(msb, other_block);
+			msb->lba_to_pba_table[lba] = pba;
+			continue;
+		}
+
+		if (!(other_overwrite_flag & MEMSTICK_OVERWRITE_UDST)) {
+			ms_printk("pba %d is marked as stable, use it",
+								other_block);
+			msb_erase_block(msb, pba);
+			continue;
+		}
+
+		ms_printk("collision between blocks %d and %d,"
+		" without stable flag set on both, erasing pba %d",
+				pba, other_block, other_block);
+
+		msb_erase_block(msb, other_block);
+		msb->lba_to_pba_table[lba] = pba;
+	}
+
+	dbg("End of media scanning");
+	kfree(overwrite_flags);
+	return 0;
+}
+
+static void msb_cache_flush_timer(unsigned long data)
+{
+	struct msb_data *msb = (struct msb_data *)data;
+	msb->need_flush_cache = true;
+	wake_up_process(msb->io_thread);
+}
+
+
+static void msb_cache_discard(struct msb_data *msb)
+{
+	if (msb->cache_block_lba == MS_BLOCK_INVALID)
+		return;
+
+	del_timer_sync(&msb->cache_flush_timer);
+
+	dbg_verbose("Discarding the write cache");
+	msb->cache_block_lba = MS_BLOCK_INVALID;
+	bitmap_zero(&msb->valid_cache_bitmap, msb->pages_in_block);
+}
+
+static int msb_cache_init(struct msb_data *msb)
+{
+	setup_timer(&msb->cache_flush_timer, msb_cache_flush_timer,
+		(unsigned long)msb);
+
+	if (!msb->cache)
+		msb->cache = kzalloc(msb->block_size, GFP_KERNEL);
+	if (!msb->cache)
+		return -ENOMEM;
+
+	msb_cache_discard(msb);
+	return 0;
+}
+
+static int msb_cache_flush(struct msb_data *msb)
+{
+	struct scatterlist sg;
+	struct ms_extra_data_register extra;
+	int page, offset, error;
+	u16 pba, lba;
+
+	if (msb->read_only)
+		return -EROFS;
+
+	if (msb->cache_block_lba == MS_BLOCK_INVALID)
+		return 0;
+
+	lba = msb->cache_block_lba;
+	pba = msb->lba_to_pba_table[lba];
+
+	dbg_verbose("Flushing the write cache of pba %d (LBA %d)",
+						pba, msb->cache_block_lba);
+
+	/* Read all missing pages in cache */
+	for (page = 0 ; page < msb->pages_in_block ; page++) {
+
+		if (test_bit(page, &msb->valid_cache_bitmap))
+			continue;
+
+		offset = page * msb->page_size;
+		sg_init_one(&sg, msb->cache + offset , msb->page_size);
+
+
+		dbg_verbose("reading non-present sector %d of cache block %d",
+			page, lba);
+		error = msb_read_page(msb, pba, page, &extra, &sg);
+
+		/* Bad pages are copied with 00 page status */
+		if (error == -EBADMSG) {
+			ms_printk("read error on sector %d, contents probably"
+				" damaged", page);
+			continue;
+		}
+
+		if (error)
+			return error;
+
+		if ((extra.overwrite_flag & MEMSTICK_OV_PG_NORMAL) !=
+							MEMSTICK_OV_PG_NORMAL) {
+			dbg("page %d is marked as bad", page);
+			continue;
+		}
+
+		set_bit(page, &msb->valid_cache_bitmap);
+	}
+
+	/* Write the cache now */
+	sg_init_one(&sg, msb->cache , msb->block_size);
+	error = msb_update_block(msb, msb->cache_block_lba, &sg);
+	pba = msb->lba_to_pba_table[msb->cache_block_lba];
+
+	/* Mark invalid pages */
+	if (!error) {
+		for (page = 0 ; page < msb->pages_in_block ; page++) {
+
+			if (test_bit(page, &msb->valid_cache_bitmap))
+				continue;
+
+			dbg("marking page %d as containing damaged data",
+				page);
+			msb_set_overwrite_flag(msb,
+				pba , page, 0xFF & ~MEMSTICK_OV_PG_NORMAL);
+		}
+	}
+
+	msb_cache_discard(msb);
+	return error;
+}
+
+static int msb_cache_write(struct msb_data *msb, int lba,
+		int page, bool add_to_cache_only, struct scatterlist *sg)
+{
+	int error;
+	if (msb->read_only)
+		return -EROFS;
+
+	if (msb->cache_block_lba == MS_BLOCK_INVALID ||
+						lba != msb->cache_block_lba)
+		if (add_to_cache_only)
+			return 0;
+
+	/* If we need to write different block */
+	if (msb->cache_block_lba != MS_BLOCK_INVALID &&
+						lba != msb->cache_block_lba) {
+		dbg_verbose("first flush the cache");
+		error = msb_cache_flush(msb);
+		if (error)
+			return error;
+	}
+
+	if (msb->cache_block_lba  == MS_BLOCK_INVALID) {
+		msb->cache_block_lba  = lba;
+		mod_timer(&msb->cache_flush_timer,
+			jiffies + msecs_to_jiffies(cache_flush_timeout));
+	}
+
+	dbg_verbose("Write of LBA %d page %d to cache ", lba, page);
+
+	sg_copy_to_buffer(sg, 1, msb->cache + page * msb->page_size,
+			msb->page_size);
+	set_bit(page, &msb->valid_cache_bitmap);
+	return 0;
+}
+
+static int msb_cache_read(struct msb_data *msb, int lba,
+					int page, struct scatterlist *sg)
+{
+	int pba = msb->lba_to_pba_table[lba];
+	int error = 0;
+
+	if (lba == msb->cache_block_lba &&
+			test_bit(page, &msb->valid_cache_bitmap)) {
+
+		dbg_verbose("Read of LBA %d (pba %d) sector %d from cache",
+							lba, pba, page);
+		sg_copy_from_buffer(sg, 1,
+			msb->cache + msb->page_size * page,
+							msb->page_size);
+	} else {
+		dbg_verbose("Read of LBA %d (pba %d) sector %d from device",
+							lba, pba, page);
+
+		error = msb_read_page(msb, pba, page, NULL, sg);
+		if (error)
+			return error;
+
+		msb_cache_write(msb, lba, page, true, sg);
+	}
+	return error;
+}
+
+/* Emulated geometry table
+ * This table content isn't that importaint,
+ * One could put here different values, providing that they still
+ * cover whole disk.
+ * 64 MB entry is what windows reports for my 64M memstick */
+
+static const struct chs_entry chs_table[] = {
+/*        size sectors cylynders  heads */
+	{ 4,    16,    247,       2  },
+	{ 8,    16,    495,       2  },
+	{ 16,   16,    495,       4  },
+	{ 32,   16,    991,       4  },
+	{ 64,   16,    991,       8  },
+	{128,   16,    991,       16 },
+	{ 0 }
+};
+
+/* Load information about the card */
+static int msb_init_card(struct memstick_dev *card)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	struct memstick_host *host = card->host;
+	struct ms_boot_page *boot_block;
+	int error = 0, i, raw_size_in_megs;
+
+	msb->caps = 0;
+
+	if (card->id.class >= MEMSTICK_CLASS_ROM &&
+				card->id.class <= MEMSTICK_CLASS_ROM)
+		msb->read_only = true;
+
+	msb->state = -1;
+	error = msb_reset(msb, false);
+	if (error)
+		return error;
+
+	/* Due to a bug in Jmicron driver written by Alex Dubov,
+	 its serial mode barely works,
+	 so we switch to parallel mode right away */
+	if (host->caps & MEMSTICK_CAP_PAR4)
+		msb_switch_to_parallel(msb);
+
+	msb->page_size = sizeof(struct ms_boot_page);
+
+	/* Read the boot page */
+	error = msb_read_boot_blocks(msb);
+	if (error)
+		return -EIO;
+
+	boot_block = &msb->boot_page[0];
+
+	/* Save intersting attributes from boot page */
+	msb->block_count = boot_block->attr.number_of_blocks;
+	msb->page_size = boot_block->attr.page_size;
+
+	msb->pages_in_block = boot_block->attr.block_size * 2;
+	msb->block_size = msb->page_size * msb->pages_in_block;
+
+	if (msb->page_size > PAGE_SIZE) {
+		/* this isn't supported by linux at all, anyway*/
+		dbg("device page %d size isn't supported", msb->page_size);
+		return -EINVAL;
+	}
+
+	msb->block_buffer = kzalloc(msb->block_size, GFP_KERNEL);
+	if (!msb->block_buffer)
+		return -ENOMEM;
+
+	raw_size_in_megs = (msb->block_size * msb->block_count) >> 20;
+
+	for (i = 0 ; chs_table[i].size ; i++) {
+
+		if (chs_table[i].size != raw_size_in_megs)
+			continue;
+
+		msb->geometry.cylinders = chs_table[i].cyl;
+		msb->geometry.heads = chs_table[i].head;
+		msb->geometry.sectors = chs_table[i].sec;
+		break;
+	}
+
+	if (boot_block->attr.transfer_supporting == 1)
+		msb->caps |= MEMSTICK_CAP_PAR4;
+
+	if (boot_block->attr.device_type & 0x03)
+		msb->read_only = true;
+
+	dbg("Total block count = %d", msb->block_count);
+	dbg("Each block consists of %d pages", msb->pages_in_block);
+	dbg("Page size = %d bytes", msb->page_size);
+	dbg("Parallel mode supported: %d", !!(msb->caps & MEMSTICK_CAP_PAR4));
+	dbg("Read only: %d", msb->read_only);
+
+#if 0
+	/* Now we can switch the interface */
+	if (host->caps & msb->caps & MEMSTICK_CAP_PAR4)
+		msb_switch_to_parallel(msb);
+#endif
+
+	error = msb_cache_init(msb);
+	if (error)
+		return error;
+
+	error = msb_ftl_initialize(msb);
+	if (error)
+		return error;
+
+
+	/* Read the bad block table */
+	error = msb_read_bad_block_table(msb, 0);
+
+	if (error && error != -ENOMEM) {
+		dbg("failed to read bad block table from primary boot block,"
+							" trying from backup");
+		error = msb_read_bad_block_table(msb, 1);
+	}
+
+	if (error)
+		return error;
+
+	/* *drum roll* Scan the media */
+	error = msb_ftl_scan(msb);
+	if (error) {
+		ms_printk("Scan of media failed");
+		return error;
+	}
+
+	return 0;
+
+}
+
+static int msb_do_write_request(struct msb_data *msb, int lba,
+		int page, struct scatterlist *sg, int *sucessfuly_written)
+{
+	int error = 0;
+	*sucessfuly_written = 0;
+
+	while (sg) {
+		if (page == 0 && sg_total_len(sg) >= msb->block_size) {
+
+			if (msb->cache_block_lba == lba)
+				msb_cache_discard(msb);
+
+			dbg_verbose("Writing whole lba %d", lba);
+			error = msb_update_block(msb, lba, sg);
+			if (error)
+				return error;
+
+			sg = sg_truncate(sg, msb->block_size);
+			*sucessfuly_written += msb->block_size;
+			lba++;
+			continue;
+		}
+
+		error = msb_cache_write(msb, lba, page, false, sg);
+		if (error)
+			return error;
+
+		sg = sg_truncate(sg, msb->page_size);
+		*sucessfuly_written += msb->page_size;
+
+		page++;
+		if (page == msb->pages_in_block) {
+			page = 0;
+			lba++;
+		}
+	}
+	return 0;
+}
+
+static int msb_do_read_request(struct msb_data *msb, int lba,
+		int page, struct scatterlist *sg, int *sucessfuly_read)
+{
+	int error = 0;
+	*sucessfuly_read = 0;
+
+	while (sg) {
+
+		error = msb_cache_read(msb, lba, page, sg);
+		if (error)
+			return error;
+
+		sg = sg_truncate(sg, msb->page_size);
+		*sucessfuly_read += msb->page_size;
+
+		page++;
+		if (page == msb->pages_in_block) {
+			page = 0;
+			lba++;
+		}
+	}
+	return 0;
+}
+
+static int msb_io_thread(void *data)
+{
+	struct msb_data *msb = data;
+	int page, error, len;
+	sector_t lba;
+	unsigned long flags;
+	struct scatterlist *sg;
+
+
+	sg = kmalloc((MS_BLOCK_MAX_SEGS+1) *
+				sizeof(struct scatterlist), GFP_KERNEL);
+	if (!sg)
+		return -ENOMEM;
+	sg_init_table(sg, MS_BLOCK_MAX_SEGS+1);
+
+	dbg("IO: thread started");
+
+	while (1) {
+
+		if (kthread_should_stop()) {
+			if (msb->req)
+				blk_requeue_request(msb->queue, msb->req);
+			break;
+		}
+
+		spin_lock_irqsave(&msb->q_lock, flags);
+
+		if (msb->need_flush_cache) {
+			msb->need_flush_cache = false;
+			spin_unlock_irqrestore(&msb->q_lock, flags);
+			msb_cache_flush(msb);
+			continue;
+		}
+
+		if (!msb->req) {
+			msb->req = blk_fetch_request(msb->queue);
+
+			if (!msb->req) {
+				dbg_verbose("IO: no more requests, sleeping");
+				set_current_state(TASK_INTERRUPTIBLE);
+				spin_unlock_irqrestore(&msb->q_lock, flags);
+				schedule();
+				dbg_verbose("IO: thread woken up");
+				continue;
+			}
+		}
+
+		spin_unlock_irqrestore(&msb->q_lock, flags);
+
+		/* If card was removed meanwhile */
+		if (!msb->req)
+			continue;
+
+		/* process the request */
+		dbg_verbose("IO: thread processing new request");
+		blk_rq_map_sg(msb->queue, msb->req, sg);
+
+		lba = blk_rq_pos(msb->req);
+
+		sector_div(lba, msb->page_size / 512);
+		page = do_div(lba, msb->pages_in_block);
+
+		if (rq_data_dir(msb->req) == READ)
+			error = msb_do_read_request(
+					msb, lba, page, sg, &len);
+		else
+			error = msb_do_write_request(
+					msb, lba, page, sg, &len);
+
+		spin_lock_irqsave(&msb->q_lock, flags);
+
+		if (len)
+			if (!__blk_end_request(msb->req, 0, len))
+				msb->req = NULL;
+
+		if (error && msb->req) {
+			dbg_verbose("IO: ending one sector "
+					"of the request with error");
+			if (!__blk_end_request(msb->req, error, msb->page_size))
+				msb->req = NULL;
+		}
+
+		if (msb->req)
+			dbg_verbose("IO: request still pending");
+
+		spin_unlock_irqrestore(&msb->q_lock, flags);
+	}
+
+	kfree(sg);
+	return 0;
+}
+
+static DEFINE_IDR(msb_disk_idr);
+static DEFINE_MUTEX(msb_disk_lock);
+
+static int msb_bd_open(struct block_device *bdev, fmode_t mode)
+{
+	struct gendisk *disk = bdev->bd_disk;
+	struct msb_data *msb = disk->private_data;
+
+	dbg_verbose("block device open");
+
+	mutex_lock(&msb_disk_lock);
+
+	if (msb && msb->card)
+		msb->usage_count++;
+
+	mutex_unlock(&msb_disk_lock);
+	return 0;
+}
+
+static void msb_data_clear(struct msb_data *msb)
+{
+	kfree(msb->boot_page);
+	kfree(msb->used_blocks_bitmap);
+	kfree(msb->lba_to_pba_table);
+	kfree(msb->cache);
+	msb->card = NULL;
+}
+
+static int msb_disk_release(struct gendisk *disk)
+{
+	struct msb_data *msb = disk->private_data;
+	int disk_id = MINOR(disk_devt(disk)) >> MS_BLOCK_PART_SHIFT;
+
+	dbg_verbose("block device release");
+
+	mutex_lock(&msb_disk_lock);
+
+	if (msb) {
+		if (msb->usage_count)
+			msb->usage_count--;
+
+		if (!msb->usage_count) {
+			kfree(msb);
+			disk->private_data = NULL;
+			idr_remove(&msb_disk_idr, disk_id);
+			put_disk(disk);
+		}
+	}
+	mutex_unlock(&msb_disk_lock);
+	return 0;
+}
+
+static int msb_bd_release(struct gendisk *disk, fmode_t mode)
+{
+	return msb_disk_release(disk);
+}
+
+static int msb_bd_getgeo(struct block_device *bdev,
+				 struct hd_geometry *geo)
+{
+	struct msb_data *msb = bdev->bd_disk->private_data;
+	*geo = msb->geometry;
+	return 0;
+}
+
+static int msb_prepare_req(struct request_queue *q, struct request *req)
+{
+	if (req->cmd_type != REQ_TYPE_FS &&
+				req->cmd_type != REQ_TYPE_BLOCK_PC) {
+		blk_dump_rq_flags(req, "MS unsupported request");
+		return BLKPREP_KILL;
+	}
+	req->cmd_flags |= REQ_DONTPREP;
+	return BLKPREP_OK;
+}
+
+static void msb_submit_req(struct request_queue *q)
+{
+	struct memstick_dev *card = q->queuedata;
+	struct msb_data *msb = memstick_get_drvdata(card);
+	struct request *req = NULL;
+
+	dbg_verbose("Submit request");
+
+	if (msb->card_dead) {
+		dbg("Refusing requests on removed card");
+
+		WARN_ON(msb->io_thread);
+
+		while ((req = blk_fetch_request(q)) != NULL)
+			__blk_end_request_all(req, -ENODEV);
+		return;
+	}
+
+	if (msb->req)
+		return;
+
+	if (msb->io_thread)
+		wake_up_process(msb->io_thread);
+}
+
+static int msb_check_card(struct memstick_dev *card)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	return (msb->card_dead == 0);
+}
+
+static void msb_stop(struct memstick_dev *card)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	unsigned long flags;
+	struct task_struct *io_thread;
+
+	dbg("Stopping all msblock IO");
+
+	/* Just stop the IO thread.
+	   Be carefull not to race against submit_request
+	   If it is called, all pending requests will be processed by
+	   the IO thread as soon as msb_start is called */
+
+	spin_lock_irqsave(&msb->q_lock, flags);
+	blk_stop_queue(msb->queue);
+	io_thread = msb->io_thread;
+	msb->io_thread = NULL;
+	spin_unlock_irqrestore(&msb->q_lock, flags);
+
+	del_timer_sync(&msb->cache_flush_timer);
+
+	if (io_thread)
+		kthread_stop(io_thread);
+}
+
+static void msb_start(struct memstick_dev *card)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	unsigned long flags;
+	int disk_id = MINOR(disk_devt(msb->disk)) >> MS_BLOCK_PART_SHIFT;
+
+	dbg("Resuming IO from msblock");
+
+	msb_invalidate_reg_window(msb);
+
+	spin_lock_irqsave(&msb->q_lock, flags);
+	if (msb->io_thread || msb->card_dead) {
+		spin_unlock_irqrestore(&msb->q_lock, flags);
+		return;
+	}
+	spin_unlock_irqrestore(&msb->q_lock, flags);
+
+	/* Kick cache flush anyway, its harmless */
+	msb->need_flush_cache = true;
+
+	msb->io_thread = kthread_run(msb_io_thread, msb, "kms_block%d",
+		disk_id);
+
+	spin_lock_irqsave(&msb->q_lock, flags);
+	blk_start_queue(msb->queue);
+	spin_unlock_irqrestore(&msb->q_lock, flags);
+}
+
+static const struct block_device_operations msb_bdops = {
+	.open    = msb_bd_open,
+	.release = msb_bd_release,
+	.getgeo  = msb_bd_getgeo,
+	.owner   = THIS_MODULE
+};
+
+/* Registers the block device */
+static int msb_init_disk(struct memstick_dev *card)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	struct memstick_host *host = card->host;
+	int rc, disk_id;
+	u64 limit = BLK_BOUNCE_HIGH;
+	unsigned long capacity;
+
+	if (host->dev.dma_mask && *(host->dev.dma_mask))
+		limit = *(host->dev.dma_mask);
+
+	mutex_lock(&msb_disk_lock);
+
+	if (!idr_pre_get(&msb_disk_idr, GFP_KERNEL)) {
+		mutex_unlock(&msb_disk_lock);
+		return -ENOMEM;
+	}
+
+	rc = idr_get_new(&msb_disk_idr, card, &disk_id);
+	mutex_unlock(&msb_disk_lock);
+
+	if (rc)
+		return rc;
+
+	if ((disk_id << MS_BLOCK_PART_SHIFT) > 255) {
+		rc = -ENOSPC;
+		goto out_release_id;
+	}
+
+	msb->disk = alloc_disk(1 << MS_BLOCK_PART_SHIFT);
+	if (!msb->disk) {
+		rc = -ENOMEM;
+		goto out_release_id;
+	}
+
+	msb->queue = blk_init_queue(msb_submit_req, &msb->q_lock);
+	if (!msb->queue) {
+		rc = -ENOMEM;
+		goto out_put_disk;
+	}
+
+	msb->queue->queuedata = card;
+	blk_queue_prep_rq(msb->queue, msb_prepare_req);
+
+	blk_queue_bounce_limit(msb->queue, limit);
+	blk_queue_max_hw_sectors(msb->queue, MS_BLOCK_MAX_PAGES);
+	blk_queue_max_segments(msb->queue, MS_BLOCK_MAX_SEGS);
+	blk_queue_max_segment_size(msb->queue,
+				   MS_BLOCK_MAX_PAGES * msb->page_size);
+	msb->disk->major = major;
+	msb->disk->first_minor = disk_id << MS_BLOCK_PART_SHIFT;
+	msb->disk->fops = &msb_bdops;
+	msb->usage_count = 1;
+	msb->disk->private_data = msb;
+	msb->disk->queue = msb->queue;
+	msb->disk->driverfs_dev = &card->dev;
+
+	sprintf(msb->disk->disk_name, "msblk%d", disk_id);
+
+	blk_queue_logical_block_size(msb->queue, msb->page_size);
+
+	capacity = msb->pages_in_block * msb->logical_block_count;
+	capacity *= (msb->page_size / 512);
+
+	set_capacity(msb->disk, capacity);
+	dbg("Set total disk size to %lu sectors", capacity);
+
+	if (msb->read_only)
+		set_disk_ro(msb->disk, 1);
+
+	msb_start(card);
+	add_disk(msb->disk);
+	dbg("Disk added");
+	return 0;
+
+out_put_disk:
+	put_disk(msb->disk);
+out_release_id:
+	mutex_lock(&msb_disk_lock);
+	idr_remove(&msb_disk_idr, disk_id);
+	mutex_unlock(&msb_disk_lock);
+	return rc;
+}
+
+static int msb_probe(struct memstick_dev *card)
+{
+	struct msb_data *msb;
+	int rc = 0;
+
+	msb = kzalloc(sizeof(struct msb_data), GFP_KERNEL);
+	if (!msb)
+		return -ENOMEM;
+	memstick_set_drvdata(card, msb);
+	msb->card = card;
+	spin_lock_init(&msb->q_lock);
+
+	rc = msb_init_card(card);
+	if (rc)
+		goto out_free;
+
+	rc = msb_init_disk(card);
+	if (!rc) {
+		card->check = msb_check_card;
+		card->stop = msb_stop;
+		card->start = msb_start;
+		return 0;
+	}
+out_free:
+	memstick_set_drvdata(card, NULL);
+	msb_data_clear(msb);
+	kfree(msb);
+	return rc;
+}
+
+static void msb_remove(struct memstick_dev *card)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	unsigned long flags;
+
+	if (msb->io_thread)
+		msb_stop(card);
+
+	dbg("Removing the disk device");
+
+	/* Take care of unhandled + new requests from now on */
+	spin_lock_irqsave(&msb->q_lock, flags);
+	msb->card_dead = true;
+	blk_start_queue(msb->queue);
+	spin_unlock_irqrestore(&msb->q_lock, flags);
+
+	/* Remove the disk */
+	del_gendisk(msb->disk);
+	blk_cleanup_queue(msb->queue);
+	msb->queue = NULL;
+
+	mutex_lock(&msb_disk_lock);
+	msb_data_clear(msb);
+	mutex_unlock(&msb_disk_lock);
+
+	msb_disk_release(msb->disk);
+	memstick_set_drvdata(card, NULL);
+}
+
+#ifdef CONFIG_PM
+
+static int msb_suspend(struct memstick_dev *card, pm_message_t state)
+{
+	msb_stop(card);
+	return 0;
+}
+
+static int msb_resume(struct memstick_dev *card)
+{
+	struct msb_data *msb = memstick_get_drvdata(card);
+	struct msb_data *new_msb = NULL;
+	bool card_dead = true;
+
+#ifndef CONFIG_MEMSTICK_UNSAFE_RESUME
+	msb->card_dead = true;
+	return 0;
+#endif
+	mutex_lock(&card->host->lock);
+
+	new_msb = kzalloc(sizeof(struct msb_data), GFP_KERNEL);
+	if (!new_msb)
+		goto out;
+
+	new_msb->card = card;
+	memstick_set_drvdata(card, new_msb);
+	spin_lock_init(&new_msb->q_lock);
+
+	if (msb_init_card(card))
+		goto out;
+
+	if (msb->block_size != new_msb->block_size)
+		goto out;
+
+	if (memcmp(msb->boot_page, new_msb->boot_page,
+					sizeof(struct ms_boot_page)))
+		goto out;
+
+	if (msb->logical_block_count != new_msb->logical_block_count ||
+		memcmp(msb->lba_to_pba_table, new_msb->lba_to_pba_table,
+						msb->logical_block_count))
+		goto out;
+
+	if (msb->block_count != new_msb->block_count ||
+		memcmp(msb->used_blocks_bitmap, new_msb->used_blocks_bitmap,
+							msb->block_count / 8))
+		goto out;
+
+	card_dead = false;
+out:
+	if (card_dead)
+		dbg("Card was removed/replaced during suspend");
+
+	msb->card_dead = card_dead;
+	memstick_set_drvdata(card, msb);
+
+	if (new_msb) {
+		msb_data_clear(new_msb);
+		kfree(new_msb);
+	}
+
+	msb_start(card);
+	mutex_unlock(&card->host->lock);
+	return 0;
+}
+#else
+
+#define msb_suspend NULL
+#define msb_resume NULL
+
+#endif /* CONFIG_PM */
+
+static struct memstick_device_id msb_id_tbl[] = {
+	{MEMSTICK_MATCH_ALL, MEMSTICK_TYPE_LEGACY, MEMSTICK_CATEGORY_STORAGE,
+	 MEMSTICK_CLASS_FLASH},
+
+	{MEMSTICK_MATCH_ALL, MEMSTICK_TYPE_LEGACY, MEMSTICK_CATEGORY_STORAGE,
+	 MEMSTICK_CLASS_ROM},
+
+	{MEMSTICK_MATCH_ALL, MEMSTICK_TYPE_LEGACY, MEMSTICK_CATEGORY_STORAGE,
+	 MEMSTICK_CLASS_RO},
+
+	{MEMSTICK_MATCH_ALL, MEMSTICK_TYPE_LEGACY, MEMSTICK_CATEGORY_STORAGE,
+	 MEMSTICK_CLASS_WP},
+
+	{MEMSTICK_MATCH_ALL, MEMSTICK_TYPE_DUO, MEMSTICK_CATEGORY_STORAGE_DUO,
+	 MEMSTICK_CLASS_DUO},
+	{}
+};
+MODULE_DEVICE_TABLE(memstick, msb_id_tbl);
+
+
+static struct memstick_driver msb_driver = {
+	.driver = {
+		.name  = DRIVER_NAME,
+		.owner = THIS_MODULE
+	},
+	.id_table = msb_id_tbl,
+	.probe    = msb_probe,
+	.remove   = msb_remove,
+	.suspend  = msb_suspend,
+	.resume   = msb_resume
+};
+
+static int __init msb_init(void)
+{
+	int rc = -ENOMEM;
+
+	rc = register_blkdev(major, DRIVER_NAME);
+	if (rc < 0) {
+		printk(KERN_ERR DRIVER_NAME ": failed to register "
+		       "major %d, error %d\n", major, rc);
+		return rc;
+	}
+	if (!major)
+		major = rc;
+
+	rc = memstick_register_driver(&msb_driver);
+	if (rc)
+		unregister_blkdev(major, DRIVER_NAME);
+	return rc;
+}
+
+static void __exit msb_exit(void)
+{
+	memstick_unregister_driver(&msb_driver);
+	unregister_blkdev(major, DRIVER_NAME);
+	idr_destroy(&msb_disk_idr);
+}
+
+module_init(msb_init);
+module_exit(msb_exit);
+
+module_param(major, int, S_IRUGO);
+MODULE_PARM_DESC(major, "Major to use for block device (default auto)");
+
+module_param(cache_flush_timeout, int, S_IRUGO);
+MODULE_PARM_DESC(cache_flush_timeout,
+				"Cache flush timeout in msec (1000 default)");
+module_param(debug, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(debug, "Debug level (0-3)");
+
+module_param(verify_writes, bool, S_IRUGO);
+MODULE_PARM_DESC(verify_writes, "Read back and check all data that is written");
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Maxim Levitsky");
+MODULE_DESCRIPTION("Sony MemoryStick block device driver");
diff --git a/drivers/memstick/core/ms_block.h b/drivers/memstick/core/ms_block.h
new file mode 100644
index 0000000..4de7be0
--- /dev/null
+++ b/drivers/memstick/core/ms_block.h
@@ -0,0 +1,245 @@
+/*
+ *  ms_block.c - Sony MemoryStick (legacy) storage support
+
+ *  Copyright (C) 2010 Maxim Levitsky <maximlevitsky@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * Minor portions of the driver are copied from mspro_block.c which is
+ * Copyright (C) 2007 Alex Dubov <oakad@yahoo.com>
+ *
+ * Also ms structures were copied from old broken driver by same author
+ * These probably come from MS spec
+ *
+ */
+
+#ifndef MS_BLOCK_NEW_H
+#define MS_BLOCK_NEW_H
+
+#define MS_BLOCK_MAX_SEGS      32
+#define MS_BLOCK_MAX_PAGES     ((2 << 16) - 1)
+
+#define MS_BLOCK_MAX_BOOT_ADDR 0x000c
+#define MS_BLOCK_BOOT_ID       0x0001
+#define MS_BLOCK_INVALID       0xffff
+#define MS_MAX_ZONES           16
+#define MS_BLOCKS_IN_ZONE      512
+
+#define MS_BLOCK_MAP_LINE_SZ   16
+#define MS_BLOCK_PART_SHIFT    3
+
+
+#define MEMSTICK_UNCORR_ERROR (MEMSTICK_STATUS1_UCFG | \
+		MEMSTICK_STATUS1_UCEX | MEMSTICK_STATUS1_UCDT)
+
+#define MEMSTICK_CORR_ERROR (MEMSTICK_STATUS1_FGER | MEMSTICK_STATUS1_EXER | \
+	MEMSTICK_STATUS1_DTER)
+
+#define MEMSTICK_INT_ERROR (MEMSTICK_INT_CMDNAK | MEMSTICK_INT_ERR)
+
+#define MEMSTICK_OVERWRITE_FLAG_NORMAL \
+	(MEMSTICK_OVERWRITE_PGST1 | \
+	MEMSTICK_OVERWRITE_PGST0  | \
+	MEMSTICK_OVERWRITE_BKST)
+
+#define MEMSTICK_OV_PG_NORMAL \
+	(MEMSTICK_OVERWRITE_PGST1 | MEMSTICK_OVERWRITE_PGST0)
+
+#define MEMSTICK_MANAGMENT_FLAG_NORMAL \
+	(MEMSTICK_MANAGEMENT_SYSFLG |  \
+	MEMSTICK_MANAGEMENT_SCMS1   |  \
+	MEMSTICK_MANAGEMENT_SCMS0)     \
+
+struct ms_boot_header {
+	unsigned short block_id;
+	unsigned short format_reserved;
+	unsigned char  reserved0[184];
+	unsigned char  data_entry;
+	unsigned char  reserved1[179];
+} __packed;
+
+
+struct ms_system_item {
+	unsigned int  start_addr;
+	unsigned int  data_size;
+	unsigned char data_type_id;
+	unsigned char reserved[3];
+} __packed;
+
+struct ms_system_entry {
+	struct ms_system_item disabled_block;
+	struct ms_system_item cis_idi;
+	unsigned char         reserved[24];
+} __packed;
+
+struct ms_boot_attr_info {
+	unsigned char      memorystick_class;
+	unsigned char      format_unique_value1;
+	unsigned short     block_size;
+	unsigned short     number_of_blocks;
+	unsigned short     number_of_effective_blocks;
+	unsigned short     page_size;
+	unsigned char      extra_data_size;
+	unsigned char      format_unique_value2;
+	unsigned char      assembly_time[8];
+	unsigned char      format_unique_value3;
+	unsigned char      serial_number[3];
+	unsigned char      assembly_manufacturer_code;
+	unsigned char      assembly_model_code[3];
+	unsigned short     memory_manufacturer_code;
+	unsigned short     memory_device_code;
+	unsigned short     implemented_capacity;
+	unsigned char      format_unique_value4[2];
+	unsigned char      vcc;
+	unsigned char      vpp;
+	unsigned short     controller_number;
+	unsigned short     controller_function;
+	unsigned char      reserved0[9];
+	unsigned char      transfer_supporting;
+	unsigned short     format_unique_value5;
+	unsigned char      format_type;
+	unsigned char      memorystick_application;
+	unsigned char      device_type;
+	unsigned char      reserved1[22];
+	unsigned char      format_uniqure_value6[2];
+	unsigned char      reserved2[15];
+} __packed;
+
+struct ms_cis_idi {
+	unsigned short general_config;
+	unsigned short logical_cylinders;
+	unsigned short reserved0;
+	unsigned short logical_heads;
+	unsigned short track_size;
+	unsigned short page_size;
+	unsigned short pages_per_track;
+	unsigned short msw;
+	unsigned short lsw;
+	unsigned short reserved1;
+	unsigned char  serial_number[20];
+	unsigned short buffer_type;
+	unsigned short buffer_size_increments;
+	unsigned short long_command_ecc;
+	unsigned char  firmware_version[28];
+	unsigned char  model_name[18];
+	unsigned short reserved2[5];
+	unsigned short pio_mode_number;
+	unsigned short dma_mode_number;
+	unsigned short field_validity;
+	unsigned short current_logical_cylinders;
+	unsigned short current_logical_heads;
+	unsigned short current_pages_per_track;
+	unsigned int   current_page_capacity;
+	unsigned short mutiple_page_setting;
+	unsigned int   addressable_pages;
+	unsigned short single_word_dma;
+	unsigned short multi_word_dma;
+	unsigned char  reserved3[128];
+} __packed;
+
+
+struct ms_boot_page {
+	struct ms_boot_header    header;
+	struct ms_system_entry   entry;
+	struct ms_boot_attr_info attr;
+} __packed;
+
+struct msb_data {
+	unsigned int			usage_count;
+	struct memstick_dev		*card;
+	struct gendisk			*disk;
+	struct request_queue		*queue;
+	spinlock_t			q_lock;
+	struct hd_geometry		geometry;
+	struct attribute_group		attr_group;
+	struct request			*req;
+	int				caps;
+
+	/* IO */
+	struct task_struct		*io_thread;
+	bool				card_dead;
+
+	/* Media properties */
+	struct ms_boot_page		*boot_page;
+	u16				boot_block_locations[2];
+	int				boot_block_count;
+
+	bool				read_only;
+	unsigned short			page_size;
+	int				block_size;
+	int				pages_in_block;
+	int				zone_count;
+	int				block_count;
+	int				logical_block_count;
+
+	/* FTL tables */
+	unsigned long			*used_blocks_bitmap;
+	unsigned long			*erased_blocks_bitmap;
+	u16				*lba_to_pba_table;
+	int				free_block_count[MS_MAX_ZONES];
+	bool				ftl_initialized;
+
+	/* Cache */
+	unsigned char			*cache;
+	unsigned long			valid_cache_bitmap;
+	int				cache_block_lba;
+	bool				need_flush_cache;
+	struct timer_list		cache_flush_timer;
+
+	/* Preallocated buffers */
+	unsigned char			*block_buffer;
+	struct scatterlist		sg[MS_BLOCK_MAX_SEGS+1];
+
+
+	/* handler's local data */
+	struct ms_register_addr		reg_addr;
+	bool				addr_valid;
+
+	u8				command_value;
+	bool				command_need_oob;
+	struct scatterlist		*current_sg;
+
+	struct ms_register		regs;
+	int				current_page;
+
+	int				state;
+	int				exit_error;
+	bool				int_polling;
+	unsigned long			int_timeout;
+
+};
+
+
+struct chs_entry {
+	unsigned long size;
+	unsigned char sec;
+	unsigned short cyl;
+	unsigned char head;
+};
+
+static int msb_reset(struct msb_data *msb, bool full);
+
+static int h_msb_default_bad(struct memstick_dev *card,
+						struct memstick_request **mrq);
+
+
+
+#define DRIVER_NAME "ms_block"
+
+#define ms_printk(format, ...) \
+	printk(KERN_INFO DRIVER_NAME ": " format "\n", ## __VA_ARGS__)
+
+#define __dbg(level, format, ...) \
+	do { \
+		if (debug >= level) \
+			printk(KERN_DEBUG DRIVER_NAME \
+				": " format "\n", ## __VA_ARGS__); \
+	} while (0)
+
+
+#define dbg(format, ...)		__dbg(1, format, ## __VA_ARGS__)
+#define dbg_verbose(format, ...)	__dbg(2, format, ## __VA_ARGS__)
+
+#endif
-- 
1.7.1

[PATCH 3/4] memstick: Add driver for Ricoh R5C592 card reader.

[Maxim Levitsky]

Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
---
 MAINTAINERS                    |    5 +
 drivers/memstick/host/Kconfig  |   12 +
 drivers/memstick/host/Makefile |    1 +
 drivers/memstick/host/r592.c   |  908 ++++++++++++++++++++++++++++++++++++++++
 drivers/memstick/host/r592.h   |  175 ++++++++
 5 files changed, 1101 insertions(+), 0 deletions(-)
 create mode 100644 drivers/memstick/host/r592.c
 create mode 100644 drivers/memstick/host/r592.h

diff --git a/MAINTAINERS b/MAINTAINERS
index 0269107..08dfc7e 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -5243,6 +5243,11 @@ S:	Maintained
 F:	drivers/mtd/nand/r852.c
 F:	drivers/mtd/nand/r852.h
 
+RICOH R5C592 MEMORYSTICK DRIVER
+M:	Maxim Levitsky <maximlevitsky@gmail.com>
+S:	Maintained
+F:	drivers/memstick/host/r592.*
+
 RISCOM8 DRIVER
 S:	Orphan
 F:	Documentation/serial/riscom8.txt
diff --git a/drivers/memstick/host/Kconfig b/drivers/memstick/host/Kconfig
index 4ce5c8d..cc0997a 100644
--- a/drivers/memstick/host/Kconfig
+++ b/drivers/memstick/host/Kconfig
@@ -30,3 +30,15 @@ config MEMSTICK_JMICRON_38X
 
           To compile this driver as a module, choose M here: the
 	  module will be called jmb38x_ms.
+
+config MEMSTICK_R592
+	tristate "Ricoh R5C592 MemoryStick interface support (EXPERIMENTAL)"
+	depends on EXPERIMENTAL && PCI
+
+	help
+	  Say Y here if you want to be able to access MemoryStick cards with
+	  the Ricoh R5C592 MemoryStick card reader (which is part of 5 in one
+		multifunction reader)
+
+	  To compile this driver as a module, choose M here: the module will
+	  be called r592.
diff --git a/drivers/memstick/host/Makefile b/drivers/memstick/host/Makefile
index 12530e4..ad63c16 100644
--- a/drivers/memstick/host/Makefile
+++ b/drivers/memstick/host/Makefile
@@ -8,3 +8,4 @@ endif
 
 obj-$(CONFIG_MEMSTICK_TIFM_MS)		+= tifm_ms.o
 obj-$(CONFIG_MEMSTICK_JMICRON_38X)	+= jmb38x_ms.o
+obj-$(CONFIG_MEMSTICK_R592)		+= r592.o
diff --git a/drivers/memstick/host/r592.c b/drivers/memstick/host/r592.c
new file mode 100644
index 0000000..767406c
--- /dev/null
+++ b/drivers/memstick/host/r592.c
@@ -0,0 +1,908 @@
+/*
+ * Copyright (C) 2010 - Maxim Levitsky
+ * driver for Ricoh memstick readers
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/freezer.h>
+#include <linux/jiffies.h>
+#include <linux/interrupt.h>
+#include <linux/pci.h>
+#include <linux/pci_ids.h>
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/kthread.h>
+#include <linux/sched.h>
+#include <linux/highmem.h>
+#include <asm/byteorder.h>
+#include <linux/swab.h>
+#include "r592.h"
+
+static int enable_dma = 1;
+static int debug;
+
+static const char *tpc_names[] = {
+	"MS_TPC_READ_MG_STATUS",
+	"MS_TPC_READ_LONG_DATA",
+	"MS_TPC_READ_SHORT_DATA",
+	"MS_TPC_READ_REG",
+	"MS_TPC_READ_QUAD_DATA",
+	"INVALID",
+	"MS_TPC_GET_INT",
+	"MS_TPC_SET_RW_REG_ADRS",
+	"MS_TPC_EX_SET_CMD",
+	"MS_TPC_WRITE_QUAD_DATA",
+	"MS_TPC_WRITE_REG",
+	"MS_TPC_WRITE_SHORT_DATA",
+	"MS_TPC_WRITE_LONG_DATA",
+	"MS_TPC_SET_CMD",
+};
+
+/**
+ * memstick_debug_get_tpc_name - debug helper that returns string for
+ * a TPC number
+ */
+const char *memstick_debug_get_tpc_name(int tpc)
+{
+	return tpc_names[tpc-1];
+}
+EXPORT_SYMBOL(memstick_debug_get_tpc_name);
+
+
+/* Read a register*/
+static inline u32 r592_read_reg(struct r592_device *dev, int address)
+{
+	u32 value = readl(dev->mmio + address);
+	dbg_reg("reg #%02d == 0x%08x", address, value);
+	return value;
+}
+
+/* Write a register */
+static inline void r592_write_reg(struct r592_device *dev,
+							int address, u32 value)
+{
+	dbg_reg("reg #%02d <- 0x%08x", address, value);
+	writel(value, dev->mmio + address);
+}
+
+/* Reads a big endian DWORD register */
+static inline u32 r592_read_reg_raw_be(struct r592_device *dev, int address)
+{
+	u32 value = __raw_readl(dev->mmio + address);
+	dbg_reg("reg #%02d == 0x%08x", address, value);
+	return be32_to_cpu(value);
+}
+
+/* Writes a big endian DWORD register */
+static inline void r592_write_reg_raw_be(struct r592_device *dev,
+							int address, u32 value)
+{
+	dbg_reg("reg #%02d <- 0x%08x", address, value);
+	__raw_writel(cpu_to_be32(value), dev->mmio + address);
+}
+
+/* Set specific bits in a register (little endian) */
+static inline void r592_set_reg_mask(struct r592_device *dev,
+							int address, u32 mask)
+{
+	u32 reg = readl(dev->mmio + address);
+	dbg_reg("reg #%02d |= 0x%08x (old =0x%08x)", address, mask, reg);
+	writel(reg | mask , dev->mmio + address);
+}
+
+/* Clear specific bits in a register (little endian) */
+static inline void r592_clear_reg_mask(struct r592_device *dev,
+						int address, u32 mask)
+{
+	u32 reg = readl(dev->mmio + address);
+	dbg_reg("reg #%02d &= 0x%08x (old = 0x%08x, mask = 0x%08x)",
+						address, ~mask, reg, mask);
+	writel(reg & ~mask, dev->mmio + address);
+}
+
+
+/* Wait for status bits while checking for errors */
+static int r592_wait_status(struct r592_device *dev, u32 mask, u32 wanted_mask)
+{
+	unsigned long timeout = jiffies + msecs_to_jiffies(1000);
+	u32 reg = r592_read_reg(dev, R592_STATUS);
+
+	if ((reg & mask) == wanted_mask)
+		return 0;
+
+	while (time_before(jiffies, timeout)) {
+
+		reg = r592_read_reg(dev, R592_STATUS);
+
+		if ((reg & mask) == wanted_mask)
+			return 0;
+
+		if (reg & (R592_STATUS_SEND_ERR | R592_STATUS_RECV_ERR))
+			return -EIO;
+
+		cpu_relax();
+	}
+	return -ETIME;
+}
+
+
+/* Enable/disable device */
+static int r592_enable_device(struct r592_device *dev, bool enable)
+{
+	dbg("%sabling the device", enable ? "en" : "dis");
+
+	if (enable) {
+
+		/* Power up the card */
+		r592_write_reg(dev, R592_POWER, R592_POWER_0 | R592_POWER_1);
+
+		/* Perform a reset */
+		r592_set_reg_mask(dev, R592_IO, R592_IO_RESET);
+
+		msleep(100);
+	} else
+		/* Power down the card */
+		r592_write_reg(dev, R592_POWER, 0);
+
+	return 0;
+}
+
+/* Set serial/parallel mode */
+static int r592_set_mode(struct r592_device *dev, bool parallel_mode)
+{
+	if (!parallel_mode) {
+		dbg("switching to serial mode");
+
+		/* Set serial mode */
+		r592_write_reg(dev, R592_IO_MODE, R592_IO_MODE_SERIAL);
+
+		r592_clear_reg_mask(dev, R592_POWER, R592_POWER_20);
+
+	} else {
+		dbg("switching to parallel mode");
+
+		/* This setting should be set _before_ switch TPC */
+		r592_set_reg_mask(dev, R592_POWER, R592_POWER_20);
+
+		r592_clear_reg_mask(dev, R592_IO,
+			R592_IO_SERIAL1 | R592_IO_SERIAL2);
+
+		/* Set the parallel mode now */
+		r592_write_reg(dev, R592_IO_MODE, R592_IO_MODE_PARALLEL);
+	}
+
+	dev->parallel_mode = parallel_mode;
+	return 0;
+}
+
+/* Perform a controller reset without powering down the card */
+static void r592_host_reset(struct r592_device *dev)
+{
+	r592_set_reg_mask(dev, R592_IO, R592_IO_RESET);
+	msleep(100);
+	r592_set_mode(dev, dev->parallel_mode);
+}
+
+/* Disable all hardware interrupts */
+static void r592_clear_interrupts(struct r592_device *dev)
+{
+	/* Disable & ACK all interrupts */
+	r592_clear_reg_mask(dev, R592_REG_MSC, IRQ_ALL_ACK_MASK);
+	r592_clear_reg_mask(dev, R592_REG_MSC, IRQ_ALL_EN_MASK);
+}
+
+/* Tests if there is an CRC error */
+static int r592_test_io_error(struct r592_device *dev)
+{
+	if (!(r592_read_reg(dev, R592_STATUS) &
+		(R592_STATUS_SEND_ERR | R592_STATUS_RECV_ERR)))
+		return 0;
+
+	return -EIO;
+}
+
+/* Ensure that FIFO is ready for use */
+static int r592_test_fifo_empty(struct r592_device *dev)
+{
+	if (r592_read_reg(dev, R592_REG_MSC) & R592_REG_MSC_FIFO_EMPTY)
+		return 0;
+
+	dbg("FIFO not ready, trying to reset the device");
+	r592_host_reset(dev);
+
+	if (r592_read_reg(dev, R592_REG_MSC) & R592_REG_MSC_FIFO_EMPTY)
+		return 0;
+
+	message("FIFO still not ready, giving up");
+	return -EIO;
+}
+
+/* Activates the DMA transfer from to FIFO */
+static void r592_start_dma(struct r592_device *dev, bool is_write)
+{
+	unsigned long flags;
+	u32 reg;
+	spin_lock_irqsave(&dev->irq_lock, flags);
+
+	/* Ack interrupts (just in case) + enable them */
+	r592_clear_reg_mask(dev, R592_REG_MSC, DMA_IRQ_ACK_MASK);
+	r592_set_reg_mask(dev, R592_REG_MSC, DMA_IRQ_EN_MASK);
+
+	/* Set DMA address */
+	r592_write_reg(dev, R592_FIFO_DMA, sg_dma_address(&dev->req->sg));
+
+	/* Enable the DMA */
+	reg = r592_read_reg(dev, R592_FIFO_DMA_SETTINGS);
+	reg |= R592_FIFO_DMA_SETTINGS_EN;
+
+	if (!is_write)
+		reg |= R592_FIFO_DMA_SETTINGS_DIR;
+	else
+		reg &= ~R592_FIFO_DMA_SETTINGS_DIR;
+	r592_write_reg(dev, R592_FIFO_DMA_SETTINGS, reg);
+
+	spin_unlock_irqrestore(&dev->irq_lock, flags);
+}
+
+/* Cleanups DMA related settings */
+static void r592_stop_dma(struct r592_device *dev, int error)
+{
+	r592_clear_reg_mask(dev, R592_FIFO_DMA_SETTINGS,
+		R592_FIFO_DMA_SETTINGS_EN);
+
+	/* This is only a precation */
+	r592_write_reg(dev, R592_FIFO_DMA,
+			dev->dummy_dma_page_physical_address);
+
+	r592_clear_reg_mask(dev, R592_REG_MSC, DMA_IRQ_EN_MASK);
+	r592_clear_reg_mask(dev, R592_REG_MSC, DMA_IRQ_ACK_MASK);
+	dev->dma_error = error;
+}
+
+/* Test if hardware supports DMA */
+static void r592_check_dma(struct r592_device *dev)
+{
+	dev->dma_capable = enable_dma &&
+		(r592_read_reg(dev, R592_FIFO_DMA_SETTINGS) &
+			R592_FIFO_DMA_SETTINGS_CAP);
+}
+
+/* Transfers fifo contents in/out using DMA */
+static int r592_transfer_fifo_dma(struct r592_device *dev)
+{
+	int len, sg_count;
+	bool is_write;
+
+	if (!dev->dma_capable || !dev->req->long_data)
+		return -EINVAL;
+
+	len = dev->req->sg.length;
+	is_write = dev->req->data_dir == WRITE;
+
+	if (len != R592_LFIFO_SIZE)
+		return -EINVAL;
+
+	dbg_verbose("doing dma transfer");
+
+	dev->dma_error = 0;
+	INIT_COMPLETION(dev->dma_done);
+
+	/* TODO: hidden assumption about nenth beeing always 1 */
+	sg_count = dma_map_sg(&dev->pci_dev->dev, &dev->req->sg, 1, is_write ?
+		PCI_DMA_TODEVICE : PCI_DMA_FROMDEVICE);
+
+	if (sg_count != 1 ||
+			(sg_dma_len(&dev->req->sg) < dev->req->sg.length)) {
+		message("problem in dma_map_sg");
+		return -EIO;
+	}
+
+	r592_start_dma(dev, is_write);
+
+	/* Wait for DMA completion */
+	if (!wait_for_completion_timeout(
+			&dev->dma_done, msecs_to_jiffies(1000))) {
+		message("DMA timeout");
+		r592_stop_dma(dev, -ETIMEDOUT);
+	}
+
+	dma_unmap_sg(&dev->pci_dev->dev, &dev->req->sg, 1, is_write ?
+		PCI_DMA_TODEVICE : PCI_DMA_FROMDEVICE);
+
+
+	return dev->dma_error;
+}
+
+/*
+ * Writes the FIFO in 4 byte chunks.
+ * If length isn't 4 byte aligned, rest of the data if put to a fifo
+ * to be written later
+ * Use r592_flush_fifo_write to flush that fifo when writing for the
+ * last time
+ */
+static void r592_write_fifo_pio(struct r592_device *dev,
+					unsigned char *buffer, int len)
+{
+	/* flush spill from former write */
+	if (!kfifo_is_empty(&dev->pio_fifo)) {
+
+		u8 tmp[4] = {0};
+		int copy_len = kfifo_in(&dev->pio_fifo, buffer, len);
+
+		if (!kfifo_is_full(&dev->pio_fifo))
+			return;
+		len -= copy_len;
+		buffer += copy_len;
+
+		copy_len = kfifo_out(&dev->pio_fifo, tmp, 4);
+		WARN_ON(copy_len != 4);
+		r592_write_reg_raw_be(dev, R592_FIFO_PIO, *(u32 *)tmp);
+	}
+
+	WARN_ON(!kfifo_is_empty(&dev->pio_fifo));
+
+	/* write full dwords */
+	while (len >= 4) {
+		r592_write_reg_raw_be(dev, R592_FIFO_PIO, *(u32 *)buffer);
+		buffer += 4;
+		len -= 4;
+	}
+
+	/* put remaining bytes to the spill */
+	if (len)
+		kfifo_in(&dev->pio_fifo, buffer, len);
+}
+
+/* Flushes the temporary FIFO used to make aligned DWORD writes */
+static void r592_flush_fifo_write(struct r592_device *dev)
+{
+	u8 buffer[4] = { 0 };
+	int len;
+
+	if (kfifo_is_empty(&dev->pio_fifo))
+		return;
+
+	len = kfifo_out(&dev->pio_fifo, buffer, 4);
+	r592_write_reg_raw_be(dev, R592_FIFO_PIO, *(u32 *)buffer);
+}
+
+/*
+ * Read a fifo in 4 bytes chunks.
+ * If input doesn't fit the buffer, it places bytes of last dword in spill
+ * buffer, so that they don't get lost on last read, just throw these away.
+ */
+static void r592_read_fifo_pio(struct r592_device *dev,
+						unsigned char *buffer, int len)
+{
+	u8 tmp[4];
+
+	/* Read from last spill */
+	if (!kfifo_is_empty(&dev->pio_fifo)) {
+		int bytes_copied =
+			kfifo_out(&dev->pio_fifo, buffer, min(4, len));
+		buffer += bytes_copied;
+		len -= bytes_copied;
+
+		if (!kfifo_is_empty(&dev->pio_fifo))
+			return;
+	}
+
+	/* Reads dwords from FIFO */
+	while (len >= 4) {
+		*(u32 *)buffer = r592_read_reg_raw_be(dev, R592_FIFO_PIO);
+		buffer += 4;
+		len -= 4;
+	}
+
+	if (len) {
+		*(u32 *)tmp = r592_read_reg_raw_be(dev, R592_FIFO_PIO);
+		kfifo_in(&dev->pio_fifo, tmp, 4);
+		len -= kfifo_out(&dev->pio_fifo, buffer, len);
+	}
+
+	WARN_ON(len);
+	return;
+}
+
+/* Transfers actual data using PIO. */
+static int r592_transfer_fifo_pio(struct r592_device *dev)
+{
+	unsigned long flags;
+
+	bool is_write = dev->req->tpc >= MS_TPC_SET_RW_REG_ADRS;
+	struct sg_mapping_iter miter;
+
+	kfifo_reset(&dev->pio_fifo);
+
+	if (!dev->req->long_data) {
+		if (is_write) {
+			r592_write_fifo_pio(dev, dev->req->data,
+							dev->req->data_len);
+			r592_flush_fifo_write(dev);
+		} else
+			r592_read_fifo_pio(dev, dev->req->data,
+							dev->req->data_len);
+		return 0;
+	}
+
+	local_irq_save(flags);
+	sg_miter_start(&miter, &dev->req->sg, 1, SG_MITER_ATOMIC |
+		(is_write ? SG_MITER_FROM_SG : SG_MITER_TO_SG));
+
+	/* Do the transfer fifo<->memory*/
+	while (sg_miter_next(&miter))
+		if (is_write)
+			r592_write_fifo_pio(dev, miter.addr, miter.length);
+		else
+			r592_read_fifo_pio(dev, miter.addr, miter.length);
+
+
+	/* Write last few non aligned bytes*/
+	if (is_write)
+		r592_flush_fifo_write(dev);
+
+	sg_miter_stop(&miter);
+	local_irq_restore(flags);
+	return 0;
+}
+
+/* Executes one TPC (data is read/written from small or large fifo) */
+static void r592_execute_tpc(struct r592_device *dev)
+{
+	bool is_write = dev->req->tpc >= MS_TPC_SET_RW_REG_ADRS;
+	int len, error;
+	u32 status, reg;
+
+	if (!dev->req) {
+		message("BUG: tpc execution without request!");
+		return;
+	}
+
+	len = dev->req->long_data ?
+		dev->req->sg.length : dev->req->data_len;
+
+	/* Ensure that FIFO can hold the input data */
+	if (len > R592_LFIFO_SIZE) {
+		message("IO: hardware doesn't support TPCs longer that 512");
+		error = -ENOSYS;
+		goto out;
+	}
+
+	if (!(r592_read_reg(dev, R592_REG_MSC) & R592_REG_MSC_PRSNT)) {
+		dbg("IO: refusing to send TPC because card is absent");
+		error = -ENODEV;
+		goto out;
+	}
+
+	dbg("IO: executing %s LEN=%d",
+			memstick_debug_get_tpc_name(dev->req->tpc), len);
+
+	/* Set IO direction */
+	if (is_write)
+		r592_set_reg_mask(dev, R592_IO, R592_IO_DIRECTION);
+	else
+		r592_clear_reg_mask(dev, R592_IO, R592_IO_DIRECTION);
+
+
+	error = r592_test_fifo_empty(dev);
+	if (error)
+		goto out;
+
+	/* Transfer write data */
+	if (is_write) {
+		error = r592_transfer_fifo_dma(dev);
+		if (error == -EINVAL)
+			error = r592_transfer_fifo_pio(dev);
+	}
+
+	if (error)
+		goto out;
+
+	/* Trigger the TPC */
+	reg = (len << R592_TPC_EXEC_LEN_SHIFT) |
+		(dev->req->tpc << R592_TPC_EXEC_TPC_SHIFT) |
+			R592_TPC_EXEC_BIG_FIFO;
+
+	r592_write_reg(dev, R592_TPC_EXEC, reg);
+
+	/* Wait for TPC completion */
+	status = R592_STATUS_RDY;
+	if (dev->req->need_card_int)
+		status |= R592_STATUS_CED;
+
+	error = r592_wait_status(dev, status, status);
+	if (error) {
+		message("card didn't respond");
+		goto out;
+	}
+
+	/* Test IO errors */
+	error = r592_test_io_error(dev);
+	if (error) {
+		dbg("IO error");
+		goto out;
+	}
+
+	/* Read data from FIFO */
+	if (!is_write) {
+		error = r592_transfer_fifo_dma(dev);
+		if (error == -EINVAL)
+			error = r592_transfer_fifo_pio(dev);
+	}
+
+	/* read INT reg. This can be shortened with shifts, but that way
+		its more readable */
+	if (dev->parallel_mode && dev->req->need_card_int) {
+
+		dev->req->int_reg = 0;
+		status = r592_read_reg(dev, R592_STATUS);
+
+		if (status & R592_STATUS_P_CMDNACK)
+			dev->req->int_reg |= MEMSTICK_INT_CMDNAK;
+		if (status & R592_STATUS_P_BREQ)
+			dev->req->int_reg |= MEMSTICK_INT_BREQ;
+		if (status & R592_STATUS_P_INTERR)
+			dev->req->int_reg |= MEMSTICK_INT_ERR;
+		if (status & R592_STATUS_P_CED)
+			dev->req->int_reg |= MEMSTICK_INT_CED;
+	}
+
+	if (error)
+		dbg("FIFO read error");
+out:
+	dev->req->error = error;
+	r592_clear_reg_mask(dev, R592_REG_MSC, R592_REG_MSC_LED);
+	return;
+}
+
+/* Main request processing thread */
+static int r592_process_thread(void *data)
+{
+	int error;
+	struct r592_device *dev = (struct r592_device *)data;
+	unsigned long flags;
+
+	while (!kthread_should_stop()) {
+		spin_lock_irqsave(&dev->io_thread_lock, flags);
+		set_current_state(TASK_INTERRUPTIBLE);
+		error = memstick_next_req(dev->host, &dev->req);
+		spin_unlock_irqrestore(&dev->io_thread_lock, flags);
+
+		if (error) {
+			if (error == -ENXIO || error == -EAGAIN) {
+				dbg_verbose("IO: done IO, sleeping");
+			} else {
+				dbg("IO: unknown error from "
+					"memstick_next_req %d", error);
+			}
+
+			if (kthread_should_stop())
+				set_current_state(TASK_RUNNING);
+
+			schedule();
+		} else {
+			set_current_state(TASK_RUNNING);
+			r592_execute_tpc(dev);
+		}
+	}
+	return 0;
+}
+
+/* Reprogram chip to detect change in card state */
+/* eg, if card is detected, arm it to detect removal, and vice versa */
+static void r592_update_card_detect(struct r592_device *dev)
+{
+	u32 reg = r592_read_reg(dev, R592_REG_MSC);
+	bool card_detected = reg & R592_REG_MSC_PRSNT;
+
+	dbg("update card detect. card state: %s", card_detected ?
+		"present" : "absent");
+
+	reg &= ~((R592_REG_MSC_IRQ_REMOVE | R592_REG_MSC_IRQ_INSERT) << 16);
+
+	if (card_detected)
+		reg |= (R592_REG_MSC_IRQ_REMOVE << 16);
+	else
+		reg |= (R592_REG_MSC_IRQ_INSERT << 16);
+
+	r592_write_reg(dev, R592_REG_MSC, reg);
+}
+
+/* Timer routine that fires 1 second after last card detection event, */
+static void r592_detect_timer(long unsigned int data)
+{
+	struct r592_device *dev = (struct r592_device *)data;
+	r592_update_card_detect(dev);
+	memstick_detect_change(dev->host);
+}
+
+/* Interrupt handler */
+static irqreturn_t r592_irq(int irq, void *data)
+{
+	struct r592_device *dev = (struct r592_device *)data;
+	irqreturn_t ret = IRQ_NONE;
+	u32 reg;
+	u16 irq_enable, irq_status;
+	unsigned long flags;
+	int error;
+
+	spin_lock_irqsave(&dev->irq_lock, flags);
+
+	reg = r592_read_reg(dev, R592_REG_MSC);
+	irq_enable = reg >> 16;
+	irq_status = reg & 0xFFFF;
+
+	/* Ack the interrupts */
+	reg &= ~irq_status;
+	r592_write_reg(dev, R592_REG_MSC, reg);
+
+	/* Get the IRQ status minus bits that aren't enabled */
+	irq_status &= (irq_enable);
+
+	/* Due to limitation of memstick core, we don't look at bits that
+		indicate that card was removed/inserted and/or present */
+	if (irq_status & (R592_REG_MSC_IRQ_INSERT | R592_REG_MSC_IRQ_REMOVE)) {
+
+		bool card_was_added = irq_status & R592_REG_MSC_IRQ_INSERT;
+		ret = IRQ_HANDLED;
+
+		message("IRQ: card %s", card_was_added ? "added" : "removed");
+
+		mod_timer(&dev->detect_timer,
+			jiffies + msecs_to_jiffies(card_was_added ? 500 : 50));
+	}
+
+	if (irq_status &
+		(R592_REG_MSC_FIFO_DMA_DONE | R592_REG_MSC_FIFO_DMA_ERR)) {
+		ret = IRQ_HANDLED;
+
+		if (irq_status & R592_REG_MSC_FIFO_DMA_ERR) {
+			message("IRQ: DMA error");
+			error = -EIO;
+		} else {
+			dbg_verbose("IRQ: dma done");
+			error = 0;
+		}
+
+		r592_stop_dma(dev, error);
+		complete(&dev->dma_done);
+	}
+
+	spin_unlock_irqrestore(&dev->irq_lock, flags);
+	return ret;
+}
+
+/* External inteface: set settings */
+static int r592_set_param(struct memstick_host *host,
+			enum memstick_param param, int value)
+{
+	struct r592_device *dev = memstick_priv(host);
+
+	switch (param) {
+	case MEMSTICK_POWER:
+		switch (value) {
+		case MEMSTICK_POWER_ON:
+			return r592_enable_device(dev, true);
+		case MEMSTICK_POWER_OFF:
+			return r592_enable_device(dev, false);
+		default:
+			return -EINVAL;
+		}
+	case MEMSTICK_INTERFACE:
+		switch (value) {
+		case MEMSTICK_SERIAL:
+			return r592_set_mode(dev, 0);
+		case MEMSTICK_PAR4:
+			return r592_set_mode(dev, 1);
+		default:
+			return -EINVAL;
+		}
+	default:
+		return -EINVAL;
+	}
+}
+
+/* External interface: submit requests */
+static void r592_submit_req(struct memstick_host *host)
+{
+	struct r592_device *dev = memstick_priv(host);
+	unsigned long flags;
+
+	if (dev->req)
+		return;
+
+	spin_lock_irqsave(&dev->io_thread_lock, flags);
+	if (wake_up_process(dev->io_thread))
+		dbg_verbose("IO thread woken to process requests");
+	spin_unlock_irqrestore(&dev->io_thread_lock, flags);
+}
+
+static const struct pci_device_id r592_pci_id_tbl[] = {
+
+	{ PCI_VDEVICE(RICOH, 0x0592), },
+	{ },
+};
+
+/* Main entry */
+static int r592_probe(struct pci_dev *pdev, const struct pci_device_id *id)
+{
+	int error = -ENOMEM;
+	struct memstick_host *host;
+	struct r592_device *dev;
+
+	/* Allocate memory */
+	host = memstick_alloc_host(sizeof(struct r592_device), &pdev->dev);
+	if (!host)
+		goto error1;
+
+	dev = memstick_priv(host);
+	dev->host = host;
+	dev->pci_dev = pdev;
+	pci_set_drvdata(pdev, dev);
+
+	/* pci initialization */
+	error = pci_enable_device(pdev);
+	if (error)
+		goto error2;
+
+	pci_set_master(pdev);
+	error = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
+	if (error)
+		goto error3;
+
+	error = pci_request_regions(pdev, DRV_NAME);
+	if (error)
+		goto error3;
+
+	dev->mmio = pci_ioremap_bar(pdev, 0);
+	if (!dev->mmio)
+		goto error4;
+
+	dev->irq = pdev->irq;
+	spin_lock_init(&dev->irq_lock);
+	spin_lock_init(&dev->io_thread_lock);
+	init_completion(&dev->dma_done);
+	INIT_KFIFO(dev->pio_fifo);
+	setup_timer(&dev->detect_timer,
+		r592_detect_timer, (long unsigned int)dev);
+
+	/* Host initialization */
+	host->caps = MEMSTICK_CAP_PAR4;
+	host->request = r592_submit_req;
+	host->set_param = r592_set_param;
+	r592_check_dma(dev);
+
+	dev->io_thread = kthread_run(r592_process_thread, dev, "r592_io");
+	if (IS_ERR(dev->io_thread)) {
+		error = PTR_ERR(dev->io_thread);
+		goto error5;
+	}
+
+	/* This is just a precation, so don't fail */
+	dev->dummy_dma_page = pci_alloc_consistent(pdev, PAGE_SIZE,
+		&dev->dummy_dma_page_physical_address);
+	r592_stop_dma(dev , 0);
+
+	if (request_irq(dev->irq, &r592_irq, IRQF_SHARED,
+			  DRV_NAME, dev))
+		goto error6;
+
+	r592_update_card_detect(dev);
+	if (memstick_add_host(host))
+		goto error7;
+
+	message("driver succesfully loaded");
+	return 0;
+error7:
+	free_irq(dev->irq, dev);
+error6:
+	if (dev->dummy_dma_page)
+		pci_free_consistent(pdev, PAGE_SIZE, dev->dummy_dma_page,
+			dev->dummy_dma_page_physical_address);
+
+	kthread_stop(dev->io_thread);
+error5:
+	iounmap(dev->mmio);
+error4:
+	pci_release_regions(pdev);
+error3:
+	pci_disable_device(pdev);
+error2:
+	memstick_free_host(host);
+error1:
+	return error;
+}
+
+static void r592_remove(struct pci_dev *pdev)
+{
+	int error = 0;
+	struct r592_device *dev = pci_get_drvdata(pdev);
+
+	/* Stop the processing thread.
+	That ensures that we won't take any more requests */
+	kthread_stop(dev->io_thread);
+
+	r592_enable_device(dev, false);
+
+	while (!error && dev->req) {
+		dev->req->error = -ETIME;
+		error = memstick_next_req(dev->host, &dev->req);
+	}
+	memstick_remove_host(dev->host);
+
+	free_irq(dev->irq, dev);
+	iounmap(dev->mmio);
+	pci_release_regions(pdev);
+	pci_disable_device(pdev);
+	memstick_free_host(dev->host);
+
+	if (dev->dummy_dma_page)
+		pci_free_consistent(pdev, PAGE_SIZE, dev->dummy_dma_page,
+			dev->dummy_dma_page_physical_address);
+}
+
+#ifdef CONFIG_PM
+static int r592_suspend(struct device *core_dev)
+{
+	struct pci_dev *pdev = to_pci_dev(core_dev);
+	struct r592_device *dev = pci_get_drvdata(pdev);
+
+	r592_clear_interrupts(dev);
+	memstick_suspend_host(dev->host);
+	del_timer_sync(&dev->detect_timer);
+	return 0;
+}
+
+static int r592_resume(struct device *core_dev)
+{
+	struct pci_dev *pdev = to_pci_dev(core_dev);
+	struct r592_device *dev = pci_get_drvdata(pdev);
+
+	r592_clear_interrupts(dev);
+	r592_enable_device(dev, false);
+	memstick_resume_host(dev->host);
+	r592_update_card_detect(dev);
+	return 0;
+}
+
+SIMPLE_DEV_PM_OPS(r592_pm_ops, r592_suspend, r592_resume);
+#endif
+
+MODULE_DEVICE_TABLE(pci, r592_pci_id_tbl);
+
+static struct pci_driver r852_pci_driver = {
+	.name		= DRV_NAME,
+	.id_table	= r592_pci_id_tbl,
+	.probe		= r592_probe,
+	.remove		= r592_remove,
+#ifdef CONFIG_PM
+	.driver.pm	= &r592_pm_ops,
+#endif
+};
+
+static __init int r592_module_init(void)
+{
+	return pci_register_driver(&r852_pci_driver);
+}
+
+static void __exit r592_module_exit(void)
+{
+	pci_unregister_driver(&r852_pci_driver);
+}
+
+module_init(r592_module_init);
+module_exit(r592_module_exit);
+
+module_param(enable_dma, bool, S_IRUGO);
+MODULE_PARM_DESC(enable_dma, "Enable usage of the DMA (default)");
+module_param(debug, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(debug, "Debug level (0-3)");
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Maxim Levitsky <maximlevitsky@gmail.com>");
+MODULE_DESCRIPTION("Ricoh R5C592 Memstick/Memstick PRO card reader driver");
diff --git a/drivers/memstick/host/r592.h b/drivers/memstick/host/r592.h
new file mode 100644
index 0000000..eee264e
--- /dev/null
+++ b/drivers/memstick/host/r592.h
@@ -0,0 +1,175 @@
+/*
+ * Copyright (C) 2010 - Maxim Levitsky
+ * driver for Ricoh memstick readers
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef R592_H
+
+#include <linux/memstick.h>
+#include <linux/spinlock.h>
+#include <linux/interrupt.h>
+#include <linux/workqueue.h>
+#include <linux/kfifo.h>
+#include <linux/ctype.h>
+
+/* write to this reg (number,len) triggers TPC execution */
+#define R592_TPC_EXEC			0x00
+#define R592_TPC_EXEC_LEN_SHIFT		16		/* Bits 16..25 are TPC len */
+#define R592_TPC_EXEC_BIG_FIFO		(1 << 26)	/* If bit 26 is set, large fifo is used (reg 48) */
+#define R592_TPC_EXEC_TPC_SHIFT		28		/* Bits 28..31 are the TPC number */
+
+
+/* Window for small TPC fifo (big endian)*/
+/* reads and writes always are done in  8 byte chunks */
+/* Not used in driver, because large fifo does better job */
+#define R592_SFIFO			0x08
+
+
+/* Status register (ms int, small fifo, IO)*/
+#define R592_STATUS			0x10
+							/* Parallel INT bits */
+#define R592_STATUS_P_CMDNACK		(1 << 16)	/* INT reg: NACK (parallel mode) */
+#define R592_STATUS_P_BREQ		(1 << 17)	/* INT reg: card ready (parallel mode)*/
+#define R592_STATUS_P_INTERR		(1 << 18)	/* INT reg: int error (parallel mode)*/
+#define R592_STATUS_P_CED		(1 << 19)	/* INT reg: command done (parallel mode) */
+
+							/* Fifo status */
+#define R592_STATUS_SFIFO_FULL		(1 << 20)	/* Small Fifo almost full (last chunk is written) */
+#define R592_STATUS_SFIFO_EMPTY		(1 << 21)	/* Small Fifo empty */
+
+							/* Error detection via CRC */
+#define R592_STATUS_SEND_ERR		(1 << 24)	/* Send failed */
+#define R592_STATUS_RECV_ERR		(1 << 25)	/* Recieve failed */
+
+							/* Card state */
+#define R592_STATUS_RDY			(1 << 28)	/* RDY signal recieved */
+#define R592_STATUS_CED			(1 << 29)	/* INT: Command done (serial mode)*/
+#define R592_STATUS_SFIFO_INPUT		(1 << 30)	/* Small fifo recieved data*/
+
+#define R592_SFIFO_SIZE			32		/* total size of small fifo is 32 bytes */
+#define R592_SFIFO_PACKET		8		/* packet size of small fifo */
+
+/* IO control */
+#define R592_IO				0x18
+#define	R592_IO_16			(1 << 16)	/* Set by default, can be cleared */
+#define	R592_IO_18			(1 << 18)	/* Set by default, can be cleared */
+#define	R592_IO_SERIAL1			(1 << 20)	/* Set by default, can be cleared, (cleared on parallel) */
+#define	R592_IO_22			(1 << 22)	/* Set by default, can be cleared */
+#define R592_IO_DIRECTION		(1 << 24)	/* TPC direction (1 write 0 read) */
+#define	R592_IO_26			(1 << 26)	/* Set by default, can be cleared */
+#define	R592_IO_SERIAL2			(1 << 30)	/* Set by default, can be cleared (cleared on parallel), serial doesn't work if unset */
+#define R592_IO_RESET			(1 << 31)	/* Reset, sets defaults*/
+
+
+/* Turns hardware on/off */
+#define R592_POWER			0x20		/* bits 0-7 writeable */
+#define R592_POWER_0			(1 << 0)	/* set on start, cleared on stop - must be set*/
+#define R592_POWER_1			(1 << 1)	/* set on start, cleared on stop - must be set*/
+#define R592_POWER_3			(1 << 3)	/* must be clear */
+#define R592_POWER_20			(1 << 5)	/* set before switch to parallel */
+
+/* IO mode*/
+#define R592_IO_MODE			0x24
+#define R592_IO_MODE_SERIAL		1
+#define R592_IO_MODE_PARALLEL		3
+
+
+/* IRQ,card detection,large fifo (first word irq status, second enable) */
+/* IRQs are ACKed by clearing the bits */
+#define R592_REG_MSC			0x28
+#define R592_REG_MSC_PRSNT		(1 << 1)	/* card present (only status)*/
+#define R592_REG_MSC_IRQ_INSERT		(1 << 8)	/* detect insert / card insered */
+#define R592_REG_MSC_IRQ_REMOVE		(1 << 9)	/* detect removal / card removed */
+#define R592_REG_MSC_FIFO_EMPTY		(1 << 10)	/* fifo is empty */
+#define R592_REG_MSC_FIFO_DMA_DONE	(1 << 11)	/* dma enable / dma done */
+
+#define R592_REG_MSC_FIFO_USER_ORN	(1 << 12)	/* set if software reads empty fifo (if R592_REG_MSC_FIFO_EMPTY is set) */
+#define R592_REG_MSC_FIFO_MISMATH	(1 << 13)	/* set if amount of data in fifo doesn't match amount in TPC */
+#define R592_REG_MSC_FIFO_DMA_ERR	(1 << 14)	/* IO failure */
+#define R592_REG_MSC_LED		(1 << 15)	/* clear to turn led off (only status)*/
+
+#define DMA_IRQ_ACK_MASK \
+	(R592_REG_MSC_FIFO_DMA_DONE | R592_REG_MSC_FIFO_DMA_ERR)
+
+#define DMA_IRQ_EN_MASK (DMA_IRQ_ACK_MASK << 16)
+
+#define IRQ_ALL_ACK_MASK 0x00007F00
+#define IRQ_ALL_EN_MASK (IRQ_ALL_ACK_MASK << 16)
+
+/* DMA address for large FIFO read/writes*/
+#define R592_FIFO_DMA			0x2C
+
+/* PIO access to large FIFO (512 bytes) (big endian)*/
+#define R592_FIFO_PIO			0x30
+#define R592_LFIFO_SIZE			512		/* large fifo size */
+
+
+/* large FIFO DMA settings */
+#define R592_FIFO_DMA_SETTINGS		0x34
+#define R592_FIFO_DMA_SETTINGS_EN	(1 << 0)	/* DMA enabled */
+#define R592_FIFO_DMA_SETTINGS_DIR	(1 << 1)	/* Dma direction (1 read, 0 write) */
+#define R592_FIFO_DMA_SETTINGS_CAP	(1 << 24)	/* Dma is aviable */
+
+/* Maybe just an delay */
+/* Bits 17..19 are just number */
+/* bit 16 is set, then bit 20 is waited */
+/* time to wait is about 50 spins * 2 ^ (bits 17..19) */
+/* seems to be possible just to ignore */
+/* Probably debug register */
+#define R592_REG38			0x38
+#define R592_REG38_CHANGE		(1 << 16)	/* Start bit */
+#define R592_REG38_DONE			(1 << 20)	/* HW set this after the delay */
+#define R592_REG38_SHIFT		17
+
+/* Debug register, written (0xABCDEF00) when error happens - not used*/
+#define R592_REG_3C			0x3C
+
+struct r592_device {
+	struct pci_dev *pci_dev;
+	struct memstick_host	*host;		/* host backpointer */
+	struct memstick_request *req;		/* current request */
+
+	/* Registers, IRQ */
+	void __iomem *mmio;
+	int irq;
+	spinlock_t irq_lock;
+	spinlock_t io_thread_lock;
+	struct timer_list detect_timer;
+
+	struct task_struct *io_thread;
+	bool parallel_mode;
+
+	DECLARE_KFIFO(pio_fifo, u8, sizeof(u32));
+
+	/* DMA area */
+	int dma_capable;
+	int dma_error;
+	struct completion dma_done;
+	void *dummy_dma_page;
+	dma_addr_t dummy_dma_page_physical_address;
+
+};
+
+#define DRV_NAME "r592"
+
+
+#define message(format, ...) \
+	printk(KERN_INFO DRV_NAME ": " format "\n", ## __VA_ARGS__)
+
+#define __dbg(level, format, ...) \
+	do { \
+		if (debug >= level) \
+			printk(KERN_DEBUG DRV_NAME \
+				": " format "\n", ## __VA_ARGS__); \
+	} while (0)
+
+
+#define dbg(format, ...)		__dbg(1, format, ## __VA_ARGS__)
+#define dbg_verbose(format, ...)	__dbg(2, format, ## __VA_ARGS__)
+#define dbg_reg(format, ...)		__dbg(3, format, ## __VA_ARGS__)
+
+#endif
-- 
1.7.1

Re: [PATCH 3/4] memstick: Add driver for Ricoh R5C592 card reader.

[Alex Dubov]

> From: Maxim Levitsky <maximlevitsky@gmail.com>
> Subject: [PATCH 3/4] memstick: Add driver for Ricoh R5C592 card reader.
> To: "Andrew Morton" <akpm@linux-foundation.org>
> Cc: "James Bottomley" <James.Bottomley@HansenPartnership.com>, "FUJITA Tomonori" <fujita.tomonori@lab.ntt.co.jp>, linux-kernel@vger.kernel.org, oakad@yahoo.com, "Maxim Levitsky" <maximlevitsky@gmail.com>
> Received: Friday, 4 March, 2011, 3:16 PM
> Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>

Acked-by: Alex Dubov <oakad@yahoo.com>
---

I understand, this patch can be applied as it is, and has no dependency
on the discussed scatter list functionality.



      

[PATCH 4/4] memstick: add Alex Dubov to MAINTAINERS of the memstick core

[Maxim Levitsky]

This code is supposed to be maintained by him and unless
he doesn't want to do so, MAINTAINERS should reflect
the state of things, and bugreporters should know
whom to blame for bugs.

Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
---
 MAINTAINERS |   13 ++++++++++++-
 1 files changed, 12 insertions(+), 1 deletions(-)

diff --git a/MAINTAINERS b/MAINTAINERS
index 08dfc7e..75ae9a9 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -5798,7 +5798,18 @@ F:	drivers/char/sonypi.c
 F:	drivers/platform/x86/sony-laptop.c
 F:	include/linux/sony-laptop.h
 
-SONY MEMORYSTICK CARD SUPPORT
+SONY MEMORYSTICK SUBSYSTEM
+M:	Alex Dubov <oakad@yahoo.com>
+S:	Maintained
+F:	drivers/memstick/memstick.c
+F:	drivers/memstick/mspro_block.c
+
+SONY MEMORYSTICK DRIVER FOR JMICRON CONTROLLERS
+M:	Alex Dubov <oakad@yahoo.com>
+S:	Maintained
+F:	drivers/memstick/host/jmb38x_ms.c
+
+SONY MEMORYSTICK DRIVER FOR TI CONTROLLERS
 M:	Alex Dubov <oakad@yahoo.com>
 W:	http://tifmxx.berlios.de/
 S:	Maintained
-- 
1.7.1

Re: [PATH 0/4] Memstick patches for 2.6.39

[Maxim Levitsky]

On Fri, 2011-03-04 at 06:16 +0200, Maxim Levitsky wrote:
> Hi,
> 
> This is a repost of my patches for 2.6.39 inclusion, which I hope not to
> miss this time.
> 
> I addressed the comments on the scatterlist issues.
> 
> Andrew, please note that my richoh memstick driver is standalone, unchanged from previos versions
> has many users which use the version I posted at ubuntu's Launchpad and happy with it.
> Please include it regardless of other patches.
> 
> The other half of my work is support for legacy memorysticks which consists of 2 patches,
> first that adds few functions to scatterlist.c, and the other patch that adds the driver.
> Driver is also stable and tested.
> 
> Best regards,
> 	Maxim Levitsky


Any update?

Best regards,
	Maxim Levitsky

Re: [PATH 0/4] Memstick patches for 2.6.39

[Maxim Levitsky]

On Sat, 2011-03-12 at 18:23 +0200, Maxim Levitsky wrote:
> On Fri, 2011-03-04 at 06:16 +0200, Maxim Levitsky wrote:
> > Hi,
> > 
> > This is a repost of my patches for 2.6.39 inclusion, which I hope not to
> > miss this time.
> > 
> > I addressed the comments on the scatterlist issues.
> > 
> > Andrew, please note that my richoh memstick driver is standalone, unchanged from previos versions
> > has many users which use the version I posted at ubuntu's Launchpad and happy with it.
> > Please include it regardless of other patches.
> > 
> > The other half of my work is support for legacy memorysticks which consists of 2 patches,
> > first that adds few functions to scatterlist.c, and the other patch that adds the driver.
> > Driver is also stable and tested.
> > 
> > Best regards,
> > 	Maxim Levitsky
> 
> 
> Any update?

Any update?


Best regards,
	Maxim Levitsky

Re: [PATH 0/4] Memstick patches for 2.6.39

[Andrew Morton]

On Tue, 15 Mar 2011 22:00:10 +0200
Maxim Levitsky <maximlevitsky@gmail.com> wrote:

> On Sat, 2011-03-12 at 18:23 +0200, Maxim Levitsky wrote:
> > On Fri, 2011-03-04 at 06:16 +0200, Maxim Levitsky wrote:
> > > Hi,
> > > 
> > > This is a repost of my patches for 2.6.39 inclusion, which I hope not to
> > > miss this time.
> > > 
> > > I addressed the comments on the scatterlist issues.
> > > 
> > > Andrew, please note that my richoh memstick driver is standalone, unchanged from previos versions
> > > has many users which use the version I posted at ubuntu's Launchpad and happy with it.
> > > Please include it regardless of other patches.
> > > 
> > > The other half of my work is support for legacy memorysticks which consists of 2 patches,
> > > first that adds few functions to scatterlist.c, and the other patch that adds the driver.
> > > Driver is also stable and tested.
> > > 
> > > Best regards,
> > > 	Maxim Levitsky
> > 
> > 
> > Any update?
> 
> Any update?
> 

I'm hoping that Alex will soon have time to (re)review these patches.

Re: [PATH 0/4] Memstick patches for 2.6.39

[Maxim Levitsky]

On Tue, 2011-03-15 at 14:04 -0700, Andrew Morton wrote:
> On Tue, 15 Mar 2011 22:00:10 +0200
> Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> 
> > On Sat, 2011-03-12 at 18:23 +0200, Maxim Levitsky wrote:
> > > On Fri, 2011-03-04 at 06:16 +0200, Maxim Levitsky wrote:
> > > > Hi,
> > > > 
> > > > This is a repost of my patches for 2.6.39 inclusion, which I hope not to
> > > > miss this time.
> > > > 
> > > > I addressed the comments on the scatterlist issues.
> > > > 
> > > > Andrew, please note that my richoh memstick driver is standalone, unchanged from previos versions
> > > > has many users which use the version I posted at ubuntu's Launchpad and happy with it.
> > > > Please include it regardless of other patches.
> > > > 
> > > > The other half of my work is support for legacy memorysticks which consists of 2 patches,
> > > > first that adds few functions to scatterlist.c, and the other patch that adds the driver.
> > > > Driver is also stable and tested.
> > > > 
> > > > Best regards,
> > > > 	Maxim Levitsky
> > > 
> > > 
> > > Any update?
> > 
> > Any update?
> > 
> 
> I'm hoping that Alex will soon have time to (re)review these patches.
Andrew Morton, I am so glad to hear something, really was frustrated to
see no responses, really thanks.

Alex already did gave green light for these patches, and they are pretty
much unchanged from that point, except trivial adoptions of changes in
scatterlist code. 

Both drivers I send are completely standalone, patch just adds the .c
and .h file.

And I add few functions to scatterlist.c, which is more or less agreed.
(If some of these functions don't fit there, I can just put in in the
driver locally).

It is very important to me to merge that code now.
There is absolutely to risk of regressions, I don't touch any code
there.

Thanks in advance,

Best regards,
	Maxim Levitsky

Re: [PATH 0/4] Memstick patches for 2.6.39

[Maxim Levitsky]

On Tue, 2011-03-15 at 14:04 -0700, Andrew Morton wrote:
> On Tue, 15 Mar 2011 22:00:10 +0200
> Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> 
> > On Sat, 2011-03-12 at 18:23 +0200, Maxim Levitsky wrote:
> > > On Fri, 2011-03-04 at 06:16 +0200, Maxim Levitsky wrote:
> > > > Hi,
> > > > 
> > > > This is a repost of my patches for 2.6.39 inclusion, which I hope not to
> > > > miss this time.
> > > > 
> > > > I addressed the comments on the scatterlist issues.
> > > > 
> > > > Andrew, please note that my richoh memstick driver is standalone, unchanged from previos versions
> > > > has many users which use the version I posted at ubuntu's Launchpad and happy with it.
> > > > Please include it regardless of other patches.
> > > > 
> > > > The other half of my work is support for legacy memorysticks which consists of 2 patches,
> > > > first that adds few functions to scatterlist.c, and the other patch that adds the driver.
> > > > Driver is also stable and tested.
> > > > 
> > > > Best regards,
> > > > 	Maxim Levitsky
> > > 
> > > 
> > > Any update?
> > 
> > Any update?
> > 
> 
> I'm hoping that Alex will soon have time to (re)review these patches.

Andrew Morton, what the current state now?


-- 
Best regards,
        Maxim Levitsky

Visit my blog: http://maximlevitsky.wordpress.com
Warning: Above blog contains rants.

Re: [PATH 0/4] Memstick patches for 2.6.39

[Andrew Morton]

On Sun, 20 Mar 2011 05:09:05 +0200 Maxim Levitsky <maximlevitsky@gmail.com> wrote:

> On Tue, 2011-03-15 at 14:04 -0700, Andrew Morton wrote:
> > On Tue, 15 Mar 2011 22:00:10 +0200
> > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > 
> > > On Sat, 2011-03-12 at 18:23 +0200, Maxim Levitsky wrote:
> > > > On Fri, 2011-03-04 at 06:16 +0200, Maxim Levitsky wrote:
> > > > > Hi,
> > > > > 
> > > > > This is a repost of my patches for 2.6.39 inclusion, which I hope not to
> > > > > miss this time.
> > > > > 
> > > > > I addressed the comments on the scatterlist issues.
> > > > > 
> > > > > Andrew, please note that my richoh memstick driver is standalone, unchanged from previos versions
> > > > > has many users which use the version I posted at ubuntu's Launchpad and happy with it.
> > > > > Please include it regardless of other patches.
> > > > > 
> > > > > The other half of my work is support for legacy memorysticks which consists of 2 patches,
> > > > > first that adds few functions to scatterlist.c, and the other patch that adds the driver.
> > > > > Driver is also stable and tested.
> > > > > 
> > > > > Best regards,
> > > > > 	Maxim Levitsky
> > > > 
> > > > 
> > > > Any update?
> > > 
> > > Any update?
> > > 
> > 
> > I'm hoping that Alex will soon have time to (re)review these patches.
> 
> Andrew Morton, what the current state now?
> 

Technical discussion is ongoing.  James has described what appears to
be the architecturally preferred way of implementing this and there is
as yet no followup to his suggestion.

Re: [PATH 0/4] Memstick patches for 2.6.39

[Maxim Levitsky]

On Sat, 2011-03-19 at 23:47 -0700, Andrew Morton wrote:
> On Sun, 20 Mar 2011 05:09:05 +0200 Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> 
> > On Tue, 2011-03-15 at 14:04 -0700, Andrew Morton wrote:
> > > On Tue, 15 Mar 2011 22:00:10 +0200
> > > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > 
> > > > On Sat, 2011-03-12 at 18:23 +0200, Maxim Levitsky wrote:
> > > > > On Fri, 2011-03-04 at 06:16 +0200, Maxim Levitsky wrote:
> > > > > > Hi,
> > > > > > 
> > > > > > This is a repost of my patches for 2.6.39 inclusion, which I hope not to
> > > > > > miss this time.
> > > > > > 
> > > > > > I addressed the comments on the scatterlist issues.
> > > > > > 
> > > > > > Andrew, please note that my richoh memstick driver is standalone, unchanged from previos versions
> > > > > > has many users which use the version I posted at ubuntu's Launchpad and happy with it.
> > > > > > Please include it regardless of other patches.
> > > > > > 
> > > > > > The other half of my work is support for legacy memorysticks which consists of 2 patches,
> > > > > > first that adds few functions to scatterlist.c, and the other patch that adds the driver.
> > > > > > Driver is also stable and tested.
> > > > > > 
> > > > > > Best regards,
> > > > > > 	Maxim Levitsky
> > > > > 
> > > > > 
> > > > > Any update?
> > > > 
> > > > Any update?
> > > > 
> > > 
> > > I'm hoping that Alex will soon have time to (re)review these patches.
> > 
> > Andrew Morton, what the current state now?
> > 
> 
> Technical discussion is ongoing.  James has described what appears to
> be the architecturally preferred way of implementing this and there is
> as yet no followup to his suggestion.
> 

But I will at least see my r592 driver in kernel?
It doesn't depend on ether ms legacy driver nor on changes in
scatterlist.c

Also, I don't have much time now to improve the ms_block driver till
this summer (studying).
The driver works. Yes it has a flaw in regard to scatterlist processing,
because I didn't find a better way to deal with this monster, but I will
fix that later. I am not the kind of guy that runs away after a merge.
It would be nice to just see my code in kernel, code I wrote more that a
year ago.

This flaw is purely theoretical. Driver does work.

One of the ways to fix this is just use plain good kernel pointers.
Yes that means bouncing of high mem, but like they say "to hell with
that". The driver deals with legacy, and quite slow devices, so there
won't be any performance difference.
Besides, my other driver for that card reader, for xD portion, does
precisely that (more correctly common code in FTL frontend, the
mtd_blkdev.c does that) and the end result is quite good.


So, I am waiting for a word from you,


-- 
Thanks in advance,
Best regards,
        Maxim Levitsky

Visit my blog: http://maximlevitsky.wordpress.com
Warning: Above blog contains rants.

Re: [PATH 0/4] Memstick patches for 2.6.39

[Maxim Levitsky]

On Sun, 2011-03-20 at 13:42 +0200, Maxim Levitsky wrote:
> On Sat, 2011-03-19 at 23:47 -0700, Andrew Morton wrote:
> > On Sun, 20 Mar 2011 05:09:05 +0200 Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > 
> > > On Tue, 2011-03-15 at 14:04 -0700, Andrew Morton wrote:
> > > > On Tue, 15 Mar 2011 22:00:10 +0200
> > > > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > > 
> > > > > On Sat, 2011-03-12 at 18:23 +0200, Maxim Levitsky wrote:
> > > > > > On Fri, 2011-03-04 at 06:16 +0200, Maxim Levitsky wrote:
> > > > > > > Hi,
> > > > > > > 
> > > > > > > This is a repost of my patches for 2.6.39 inclusion, which I hope not to
> > > > > > > miss this time.
> > > > > > > 
> > > > > > > I addressed the comments on the scatterlist issues.
> > > > > > > 
> > > > > > > Andrew, please note that my richoh memstick driver is standalone, unchanged from previos versions
> > > > > > > has many users which use the version I posted at ubuntu's Launchpad and happy with it.
> > > > > > > Please include it regardless of other patches.
> > > > > > > 
> > > > > > > The other half of my work is support for legacy memorysticks which consists of 2 patches,
> > > > > > > first that adds few functions to scatterlist.c, and the other patch that adds the driver.
> > > > > > > Driver is also stable and tested.
> > > > > > > 
> > > > > > > Best regards,
> > > > > > > 	Maxim Levitsky
> > > > > > 
> > > > > > 
> > > > > > Any update?
> > > > > 
> > > > > Any update?
> > > > > 
> > > > 
> > > > I'm hoping that Alex will soon have time to (re)review these patches.
> > > 
> > > Andrew Morton, what the current state now?
> > > 
> > 
> > Technical discussion is ongoing.  James has described what appears to
> > be the architecturally preferred way of implementing this and there is
> > as yet no followup to his suggestion.
> > 
> 
> But I will at least see my r592 driver in kernel?
> It doesn't depend on ether ms legacy driver nor on changes in
> scatterlist.c
> 
> Also, I don't have much time now to improve the ms_block driver till
> this summer (studying).
> The driver works. Yes it has a flaw in regard to scatterlist processing,
> because I didn't find a better way to deal with this monster, but I will
> fix that later. I am not the kind of guy that runs away after a merge.
> It would be nice to just see my code in kernel, code I wrote more that a
> year ago.
> 
> This flaw is purely theoretical. Driver does work.
> 
> One of the ways to fix this is just use plain good kernel pointers.
> Yes that means bouncing of high mem, but like they say "to hell with
> that". The driver deals with legacy, and quite slow devices, so there
> won't be any performance difference.
> Besides, my other driver for that card reader, for xD portion, does
> precisely that (more correctly common code in FTL frontend, the
> mtd_blkdev.c does that) and the end result is quite good.
> 
> 
> So, I am waiting for a word from you,
Any update?


-- 
Best regards,
        Maxim Levitsky

Visit my blog: http://maximlevitsky.wordpress.com
Warning: Above blog contains rants.

Re: [PATH 0/4] Memstick patches for 2.6.39

[Andrew Morton]

On Sun, 20 Mar 2011 13:42:51 +0200
Maxim Levitsky <maximlevitsky@gmail.com> wrote:

> On Sat, 2011-03-19 at 23:47 -0700, Andrew Morton wrote:
> > On Sun, 20 Mar 2011 05:09:05 +0200 Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > 
> 
> ...
>
> Also, I don't have much time now to improve the ms_block driver till
> this summer (studying).
> The driver works. Yes it has a flaw in regard to scatterlist processing,
> because I didn't find a better way to deal with this monster, but I will
> fix that later. I am not the kind of guy that runs away after a merge.
> It would be nice to just see my code in kernel, code I wrote more that a
> year ago.
> 
> This flaw is purely theoretical. Driver does work.

Lots of code is "flawed but works".  The place for such code is
drivers/staging/ - it gets put in there so the code is available for
those who need the driver and the code is later moved over into
drivers/ once the flaws have been addressed.

So a path forward here would be for us to put the driver and the
sglist extensions into a directory under drivers/staging/.

Re: [PATH 0/4] Memstick patches for 2.6.39

[Maxim Levitsky]

On Tue, 2011-03-22 at 16:21 -0700, Andrew Morton wrote:
> On Sun, 20 Mar 2011 13:42:51 +0200
> Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> 
> > On Sat, 2011-03-19 at 23:47 -0700, Andrew Morton wrote:
> > > On Sun, 20 Mar 2011 05:09:05 +0200 Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > 
> > 
> > ...
> >
> > Also, I don't have much time now to improve the ms_block driver till
> > this summer (studying).
> > The driver works. Yes it has a flaw in regard to scatterlist processing,
> > because I didn't find a better way to deal with this monster, but I will
> > fix that later. I am not the kind of guy that runs away after a merge.
> > It would be nice to just see my code in kernel, code I wrote more that a
> > year ago.
> > 
> > This flaw is purely theoretical. Driver does work.
> 
> Lots of code is "flawed but works".  The place for such code is
> drivers/staging/ - it gets put in there so the code is available for
> those who need the driver and the code is later moved over into
> drivers/ once the flaws have been addressed.
> 
> So a path forward here would be for us to put the driver and the
> sglist extensions into a directory under drivers/staging/.
Ok, I won't be fighting with you over this one.

However, I ask for this:


1. Please merge r592.c, it doesn't depend on anything.

2. Please review ms_block.c for other problems that might prevent merge.
For example when I published the sg list helpers, nobody told me that I
am not allowed to add them. Actually the opposite, I was told to put
them in scatterlist.c.
When I did so, again I was told that I didn't do the kerneldoc comments
right, and also was told to improve few of the functions.
Now I did all that, and I am told that scatterlist usage in my driver is
no-go. OK. But what else is there that you don't like?

Best regards,
	Maxim Levitsky

Re: [PATH 0/4] Memstick patches for 2.6.39

[Greg KH]

On Wed, Mar 23, 2011 at 02:59:34AM +0200, Maxim Levitsky wrote:
> On Tue, 2011-03-22 at 16:21 -0700, Andrew Morton wrote:
> > On Sun, 20 Mar 2011 13:42:51 +0200
> > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > 
> > > On Sat, 2011-03-19 at 23:47 -0700, Andrew Morton wrote:
> > > > On Sun, 20 Mar 2011 05:09:05 +0200 Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > > 
> > > 
> > > ...
> > >
> > > Also, I don't have much time now to improve the ms_block driver till
> > > this summer (studying).
> > > The driver works. Yes it has a flaw in regard to scatterlist processing,
> > > because I didn't find a better way to deal with this monster, but I will
> > > fix that later. I am not the kind of guy that runs away after a merge.
> > > It would be nice to just see my code in kernel, code I wrote more that a
> > > year ago.
> > > 
> > > This flaw is purely theoretical. Driver does work.
> > 
> > Lots of code is "flawed but works".  The place for such code is
> > drivers/staging/ - it gets put in there so the code is available for
> > those who need the driver and the code is later moved over into
> > drivers/ once the flaws have been addressed.
> > 
> > So a path forward here would be for us to put the driver and the
> > sglist extensions into a directory under drivers/staging/.
> Ok, I won't be fighting with you over this one.
> 
> However, I ask for this:
> 
> 
> 1. Please merge r592.c, it doesn't depend on anything.

What type of driver is that?

> 2. Please review ms_block.c for other problems that might prevent merge.
> For example when I published the sg list helpers, nobody told me that I
> am not allowed to add them. Actually the opposite, I was told to put
> them in scatterlist.c.
> When I did so, again I was told that I didn't do the kerneldoc comments
> right, and also was told to improve few of the functions.
> Now I did all that, and I am told that scatterlist usage in my driver is
> no-go. OK. But what else is there that you don't like?

Ick, that's not nice, but formatting issues are usually the easiest to
pick on, sorry for that happening.

So the result is to use the block layer functions instead, right?  How
much work do you imagine that would take to do?

If it's a bunch, care to put the code in drivers/staging/ now?  I'll
gladly take the patches to queue them up for .40.

thanks,

greg k-h

Re: [PATH 0/4] Memstick patches for 2.6.39

[Maxim Levitsky]

On Tue, 2011-03-22 at 20:42 -0700, Greg KH wrote:
> On Wed, Mar 23, 2011 at 02:59:34AM +0200, Maxim Levitsky wrote:
> > On Tue, 2011-03-22 at 16:21 -0700, Andrew Morton wrote:
> > > On Sun, 20 Mar 2011 13:42:51 +0200
> > > Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > 
> > > > On Sat, 2011-03-19 at 23:47 -0700, Andrew Morton wrote:
> > > > > On Sun, 20 Mar 2011 05:09:05 +0200 Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > > > 
> > > > 
> > > > ...
> > > >
> > > > Also, I don't have much time now to improve the ms_block driver till
> > > > this summer (studying).
> > > > The driver works. Yes it has a flaw in regard to scatterlist processing,
> > > > because I didn't find a better way to deal with this monster, but I will
> > > > fix that later. I am not the kind of guy that runs away after a merge.
> > > > It would be nice to just see my code in kernel, code I wrote more that a
> > > > year ago.
> > > > 
> > > > This flaw is purely theoretical. Driver does work.
> > > 
> > > Lots of code is "flawed but works".  The place for such code is
> > > drivers/staging/ - it gets put in there so the code is available for
> > > those who need the driver and the code is later moved over into
> > > drivers/ once the flaws have been addressed.
> > > 
> > > So a path forward here would be for us to put the driver and the
> > > sglist extensions into a directory under drivers/staging/.
> > Ok, I won't be fighting with you over this one.
> > 
> > However, I ask for this:
> > 
> > 
> > 1. Please merge r592.c, it doesn't depend on anything.
> 
> What type of driver is that?
A driver for card reader that reads Memstick cards.
Two such drivers from Alex Dubov were already merged.


> 
> > 2. Please review ms_block.c for other problems that might prevent merge.
> > For example when I published the sg list helpers, nobody told me that I
> > am not allowed to add them. Actually the opposite, I was told to put
> > them in scatterlist.c.
> > When I did so, again I was told that I didn't do the kerneldoc comments
> > right, and also was told to improve few of the functions.
> > Now I did all that, and I am told that scatterlist usage in my driver is
> > no-go. OK. But what else is there that you don't like?
> 
> Ick, that's not nice, but formatting issues are usually the easiest to
> pick on, sorry for that happening.

Folk, could you actually review the driver once again, and really tell
me what is wrong with it?
I did another review of it now, and it looks nice and clean.

Let me explain exactly the controversial point:

I get a request from block core to read or write a specific amount of
512 blocks.
I turn that request into scatterlist.

For reads (look at msb_do_read_request), I read one page into first 512
bytes of pages that are pointed by that scatterlist, and then modify it
such as it now shorter and points to next 512 bytes.

For writes, I choose for 2 cases.
If I was told to write whole erase block or more, I create new scatter
list that covers the 'erase block size' bytes of original one and just
write it.

For smaller writes, I have a eraseblock sized cache buffer that I write
to until I ased to write to different erase block. Then I flush current
eraseblock doing copyonwrite.

Basicly the controversy is around my usage of scatterlist.
I just added ability to create a copy of a scatterlistby which will
cover n bytes, and ability to truncate a scatterlist  by removing n
bytes from its head.



> 
> So the result is to use the block layer functions instead, right?  How
> much work do you imagine that would take to do?
I could just stop using scatterlists at all, and just do all
writes/reads 512 bytes a time (using the copy on write cache for writes
of course).
I did that successfully in my sm_ftl.c



> 
> If it's a bunch, care to put the code in drivers/staging/ now?  I'll
> gladly take the patches to queue them up for .40.
> 
> thanks,
> 
> greg k-h

-- 
Best regards,
        Maxim Levitsky

Visit my blog: http://maximlevitsky.wordpress.com
Warning: Above blog contains rants.

Re: [PATH 0/4] Memstick patches for 2.6.39

[Alex Dubov]

--- On Thu, 24/3/11, Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > > 
> > > 1. Please merge r592.c, it doesn't depend on
> anything.
> > 
> > What type of driver is that?
> A driver for card reader that reads Memstick cards.
> Two such drivers from Alex Dubov were already merged.

I would like to point out that the discussion is not about r592.
We all agreed, quite some time ago, that it can be merged and were you,
Maxim, submitted it separately 4 month ago as per my advice it would be
in the tree already.