Hi Artem,
On Mon, 30 Jul 2012 16:56:50 +0300 Artem Bityutskiy wrote:
> Hi Shmulik, I've separated out the defconfig changes and pushed patches
> 1,2, and 3 to the UBI tree (the master branch). Patches 4 and 5 are
> already merged upstream. I did a couple of minor modifications in
> commentaries
Hi Artem,
On Mon, 30 Jul 2012 16:56:50 +0300 Artem Bityutskiy dedeki...@gmail.com wrote:
Hi Shmulik, I've separated out the defconfig changes and pushed patches
1,2, and 3 to the UBI tree (the master branch). Patches 4 and 5 are
already merged upstream. I did a couple of minor modifications in
On Wed, 2012-07-04 at 11:05 +0300, Shmulik Ladkani wrote:
> The existing mechanism of reserving PEBs for bad PEB handling has two
> flaws:
> - It is calculated as a percentage of good PEBs instead of total PEBs.
> - There's no limit on the amount of PEBs UBI reserves for future bad
> eraseblock
On Wed, 2012-07-04 at 11:05 +0300, Shmulik Ladkani wrote:
The existing mechanism of reserving PEBs for bad PEB handling has two
flaws:
- It is calculated as a percentage of good PEBs instead of total PEBs.
- There's no limit on the amount of PEBs UBI reserves for future bad
eraseblock
On Tue, 2012-07-17 at 10:23 +0300, Shmulik Ladkani wrote:
> But it would only make sense after accepting Richard's patchset as well,
> which suggests configuring per-ubi-device beb limit via the attach ioctl.
Yes, sure, you are right.
--
Best Regards,
Artem Bityutskiy
signature.asc
On Tue, 2012-07-17 at 10:23 +0300, Shmulik Ladkani wrote:
But it would only make sense after accepting Richard's patchset as well,
which suggests configuring per-ubi-device beb limit via the attach ioctl.
Yes, sure, you are right.
--
Best Regards,
Artem Bityutskiy
signature.asc
Description:
Hi Artem,
On Mon, 16 Jul 2012 18:33:57 +0300 Artem Bityutskiy wrote:
> But one more think is the mtd web-site. I've grepped for '1%' and there
> are plenty of them. I've changed them all to 2% more or less
> mechanically - only cleaned up one section by removing out-of-date
> information. Would
Hi Artem,
On Mon, 16 Jul 2012 18:33:57 +0300 Artem Bityutskiy dedeki...@gmail.com wrote:
But one more think is the mtd web-site. I've grepped for '1%' and there
are plenty of them. I've changed them all to 2% more or less
mechanically - only cleaned up one section by removing out-of-date
On Wed, 2012-07-04 at 11:05 +0300, Shmulik Ladkani wrote:
> The existing mechanism of reserving PEBs for bad PEB handling has two
> flaws:
> - It is calculated as a percentage of good PEBs instead of total PEBs.
> - There's no limit on the amount of PEBs UBI reserves for future bad
> eraseblock
On Wed, 2012-07-04 at 11:05 +0300, Shmulik Ladkani wrote:
The existing mechanism of reserving PEBs for bad PEB handling has two
flaws:
- It is calculated as a percentage of good PEBs instead of total PEBs.
- There's no limit on the amount of PEBs UBI reserves for future bad
eraseblock
2012/7/7 Shmulik Ladkani :
> Many thanks for testing.
>
> Could you please verify the crash only occurs with the patch?
>
> Can you provide the vmlinux matching this oops, so I may analyze the
> exact null dereferencing point?
> It seems to be somewhere in ubi_wl_init, however the patch seems not
2012/7/7 Shmulik Ladkani shmulik.ladk...@gmail.com:
Many thanks for testing.
Could you please verify the crash only occurs with the patch?
Can you provide the vmlinux matching this oops, so I may analyze the
exact null dereferencing point?
It seems to be somewhere in ubi_wl_init, however
2012/7/7 Shmulik Ladkani :
> Many thanks for testing.
>
> Could you please verify the crash only occurs with the patch?
>
> Can you provide the vmlinux matching this oops, so I may analyze the
> exact null dereferencing point?
> It seems to be somewhere in ubi_wl_init, however the patch seems not
Hi Richard,
On Fri, 6 Jul 2012 17:27:59 +0200 Richard Genoud
wrote:
> I've got an oops...
> this is my dev-kernel in 3.5-rc5 + some work to be able to boot on my board
> NB: If I use ubi_format it's ok.
> the mtd1 device has 1984 PEB
> the 4 last are UBI reserved + BBT
>
> I didn't test
Hi Richard,
On Fri, 6 Jul 2012 17:27:59 +0200 Richard Genoud richard.gen...@gmail.com
wrote:
I've got an oops...
this is my dev-kernel in 3.5-rc5 + some work to be able to boot on my board
NB: If I use ubi_format it's ok.
the mtd1 device has 1984 PEB
the 4 last are UBI reserved + BBT
I
2012/7/7 Shmulik Ladkani shmulik.ladk...@gmail.com:
Many thanks for testing.
Could you please verify the crash only occurs with the patch?
Can you provide the vmlinux matching this oops, so I may analyze the
exact null dereferencing point?
It seems to be somewhere in ubi_wl_init, however
I've got an oops...
this is my dev-kernel in 3.5-rc5 + some work to be able to boot on my board
NB: If I use ubi_format it's ok.
the mtd1 device has 1984 PEB
the 4 last are UBI reserved + BBT
I didn't test without your patch, but anyway something is wrong there.
# flash_erase /dev/mtd1 0 1980
I've got an oops...
this is my dev-kernel in 3.5-rc5 + some work to be able to boot on my board
NB: If I use ubi_format it's ok.
the mtd1 device has 1984 PEB
the 4 last are UBI reserved + BBT
I didn't test without your patch, but anyway something is wrong there.
# flash_erase /dev/mtd1 0 1980
18 matches
Mail list logo
