[PATCH] flashrom: Reflect generic match status more accurately

[Carl-Daniel Hailfinger]

All "unknown xy SPI chip" entries claim to have status UNTESTED for
probe/read/erase/write. That is incorrect.

A bit of confusion comes from how the #defines are named. We call them
TEST_BAD_*, but the message printed by flashrom says:
"This flash part has status NOT WORKING for operations:"

Something that is unimplemented is definitely not working.

Neither of the chip entries mentioned above has erase or write functions
implemented, so erase and write are not working.
Since their size is unknown, we can't read them in. That means read is
not working as well.
Probing is a different matter. If a chip-specific probe function had
matched, we wouldn't have to handle the chip with the "unknown xy SPI
chip" fallback. I'm tempted to call that "not working" as well, but I'm
open to discussion on this point.

Signed-off-by: Carl-Daniel Hailfinger

Index: flashrom-test_bad/flashchips.c
===================================================================
--- flashrom-test_bad/flashchips.c	(Revision 4162)
+++ flashrom-test_bad/flashchips.c	(Arbeitskopie)
@@ -2167,7 +2167,7 @@
    	.model_id	= GENERIC_DEVICE_ID,
    	.total_size	= 0,
    	.page_size	= 256,
-		.tested		= TEST_UNTESTED,
+		.tested		= TEST_BAD_PREW,
    	.probe		= probe_spi_rdid,
    	.erase		= NULL,
    	.write		= NULL,
@@ -2180,7 +2180,7 @@
    	.model_id	= GENERIC_DEVICE_ID,
    	.total_size	= 0,
    	.page_size	= 256,
-		.tested		= TEST_UNTESTED,
+		.tested		= TEST_BAD_PREW,
    	.probe		= probe_spi_rdid,
    	.erase		= NULL,
    	.write		= NULL,
@@ -2193,7 +2193,7 @@
    	.model_id	= GENERIC_DEVICE_ID,
    	.total_size	= 0,
    	.page_size	= 256,
-		.tested		= TEST_UNTESTED,
+		.tested		= TEST_BAD_PREW,
    	.probe		= probe_spi_rdid,
    	.erase		= NULL,
    	.write		= NULL,
@@ -2206,7 +2206,7 @@
    	.model_id	= GENERIC_DEVICE_ID,
    	.total_size	= 0,
    	.page_size	= 256,
-		.tested		= TEST_UNTESTED,
+		.tested		= TEST_BAD_PREW,
    	.probe		= probe_spi_rdid,
    	.erase		= NULL,
    	.write		= NULL,
@@ -2219,7 +2219,7 @@
    	.model_id	= GENERIC_DEVICE_ID,
    	.total_size	= 0,
    	.page_size	= 256,
-		.tested		= TEST_UNTESTED,
+		.tested		= TEST_BAD_PREW,
    	.probe		= probe_spi_rdid,
    	.erase		= NULL,
    	.write		= NULL,
@@ -2232,7 +2232,7 @@
    	.model_id	= GENERIC_DEVICE_ID,
    	.total_size	= 0,
    	.page_size	= 256,
-		.tested		= TEST_UNTESTED,
+		.tested		= TEST_BAD_PREW,
    	.probe		= probe_spi_rdid,
    	.erase		= NULL,
    	.write		= NULL,

[PATCH] flashrom: Group probe function together with associated IDs (and unknown chips)

[Peter Stuge]

Carl-Daniel Hailfinger wrote:
> Group each probe function together with the associated IDs in the flash
> chip description.

Please add a ticket for this in trac to avoid that the patch is
forgotten.

Carl-Daniel Hailfinger wrote:
> All "unknown xy SPI chip" entries claim to have status UNTESTED for
> probe/read/erase/write.

I think that's the only test status that makes sense for them.
> Something that is unimplemented is definitely not working.

TEST_BAD_ means that the operation has been confirmed to fail. Since
these unknown chip entries will match any number of chips never seen
by flashrom before I do not think it makes much sense to claim that
anything at all has been confirmed.

I'd like to shift focus to how we can take out the unknown chip
entries.

What purpose do these entries serve? Is it strictly to inform the
user that flashrom was able to communicate with the flash chip but
does not yet support the particular model? I think that is a fairly
special case, which needs special handling.

//Peter

[PATCH] flashrom: Reflect generic match status more accurately

[Uwe Hermann]

On Wed, Apr 22, 2009 at 01:43:14AM +0200, Carl-Daniel Hailfinger wrote:
> All "unknown xy SPI chip" entries claim to have status UNTESTED for
> probe/read/erase/write. That is incorrect.
>
> A bit of confusion comes from how the #defines are named. We call them
> TEST_BAD_*, but the message printed by flashrom says:
> "This flash part has status NOT WORKING for operations:"
>
> Something that is unimplemented is definitely not working.
>
> Neither of the chip entries mentioned above has erase or write functions
> implemented, so erase and write are not working.
> Since their size is unknown, we can't read them in. That means read is
> not working as well.
> Probing is a different matter. If a chip-specific probe function had
> matched, we wouldn't have to handle the chip with the "unknown xy SPI
> chip" fallback. I'm tempted to call that "not working" as well, but I'm
> open to discussion on this point.
>
> Signed-off-by: Carl-Daniel Hailfinger

Sounds reasonable.

Acked-by: Uwe Hermann

Uwe.

[PATCH] flashrom: Reflect generic match status more accurately

[Carl-Daniel Hailfinger]

On 22.04.2009 14:41, Uwe Hermann wrote:
> On Wed, Apr 22, 2009 at 01:43:14AM +0200, Carl-Daniel Hailfinger wrote:
> > All "unknown xy SPI chip" entries claim to have status UNTESTED for
> > probe/read/erase/write. That is incorrect.
> >
> > A bit of confusion comes from how the #defines are named. We call them
> > TEST_BAD_*, but the message printed by flashrom says:
> > "This flash part has status NOT WORKING for operations:"
> >
> > Something that is unimplemented is definitely not working.
> >
> > Neither of the chip entries mentioned above has erase or write functions
> > implemented, so erase and write are not working.
> > Since their size is unknown, we can't read them in. That means read is
> > not working as well.
> > Probing is a different matter. If a chip-specific probe function had
> > matched, we wouldn't have to handle the chip with the "unknown xy SPI
> > chip" fallback. I'm tempted to call that "not working" as well, but I'm
> > open to discussion on this point.
> >
> > Signed-off-by: Carl-Daniel Hailfinger
>
> Sounds reasonable.
>
> Acked-by: Uwe Hermann

Thanks, r4177.

Regards,
Carl-Daniel

[PATCH] flashrom: Reflect generic match status more accurately

[Peter Stuge]

Uwe Hermann wrote:
> Sounds reasonable.

Does it really sound reasonable for flashrom to claim that a chip
which can only be matched by a wildcard has a known test status?

//Peter