On Tue, Aug 30, 2022 at 10:09:11AM +0200, Johannes Zink wrote:
> From: Ahmad Fatoum
>
> This commit ports the axp20x driver from linux kernel code. Structures
> not applicable for barebox are being dropped.
>
> Signed-off-by: Ahmad Fatoum
> Signed-off-by: Johannes Zink
> ---
> drivers/mfd/Kco
On Tue, Aug 30, 2022 at 10:06:13AM +0200, Johannes Zink wrote:
> The i.MX7 GPMI NAND controller is completely backwards compatible with
> the version of the IP on the i.MX6. There are additional features like
> increased ECC strength, that we may want to use in future, but for now
> treating it the
On Tue, Aug 30, 2022 at 09:50:45AM +0200, Ahmad Fatoum wrote:
> Now with i.MX8M feature controller driver support available, have the
> OCOTP provide feature control on the i.MX8MM to ensure the kernel DT
> does not cause Linux to access the VPU and its power domains,
> when barebox knows them to b
On Tue, Aug 30, 2022 at 09:37:55AM +0200, Enrico Scholz wrote:
> The tftp "windowsize" greatly improves the performance of tftp
> transfers. This patchset adds support for it.
>
> The first two patches are a little bit unrelated and enhance the 'cp
> -v' output by giving information about the tra
On Tue, Aug 30, 2022 at 09:38:57AM +0200, Ahmad Fatoum wrote:
> Hello Sascha,
>
> On 30.08.22 09:32, Sascha Hauer wrote:
> > Hi Ahmad,
> >
> > On Thu, Aug 18, 2022 at 07:19:45AM +0200, Ahmad Fatoum wrote:
> >> The i.MX8MM exists in a Lite variant with no VPUs as well as Solo and
> >> Dual variant
On Tue, Aug 30, 2022 at 10:39:37AM +0200, Johannes Zink wrote:
> We register the i.MX7 clock controller driver at core_initcall level and
> then do some initial clock setup/reparenting at postcore_initcall level.
> This doesn't work as expected when deep probe is enabled, because while
> the driver
Some phys registered from devicetree have pinctrl settings attached to
them. Configure the pinctrl state to make sure they can be accessed
before trying to reset and detect the phy devices.
Signed-off-by: Sascha Hauer
---
drivers/net/phy/mdio_bus.c | 2 ++
1 file changed, 2 insertions(+)
diff -
On Tue, 23 Aug 2022 07:18:02 +0200
Ahmad Fatoum wrote:
Hi Ahmad,
> Hello Antony,
>
> On 22.08.22 13:32, Antony Pavlov wrote:
> > On Tue, 26 Apr 2022 08:46:04 +0200
> > Ahmad Fatoum wrote:
> >
> > Hi Ahmad!
> >
> > I'm trying to run SiFive target with emulate.pl, e.g.
> >
> > barebox$ dd if=
On 30.08.22 10:29, Marco Felsch wrote:
> On 22-08-30, Ahmad Fatoum wrote:
>> Hello Marco,
>>
>> On 30.08.22 10:10, Marco Felsch wrote:
>>> Hi Ahamd,
>>>
>>> On 22-08-30, Ahmad Fatoum wrote:
Now with i.MX8M feature controller driver support available, have the
OCOTP provide feature control
We register the i.MX7 clock controller driver at core_initcall level and
then do some initial clock setup/reparenting at postcore_initcall level.
This doesn't work as expected when deep probe is enabled, because while
the driver is registered at core_initcall level, it's only probed
later on, curre
On 22-08-30, Ahmad Fatoum wrote:
> Hello Marco,
>
> On 30.08.22 10:10, Marco Felsch wrote:
> > Hi Ahamd,
> >
> > On 22-08-30, Ahmad Fatoum wrote:
> >> Now with i.MX8M feature controller driver support available, have the
> >> OCOTP provide feature control on the i.MX8MM to ensure the kernel DT
>
Hello Marco,
On 30.08.22 10:10, Marco Felsch wrote:
> Hi Ahamd,
>
> On 22-08-30, Ahmad Fatoum wrote:
>> Now with i.MX8M feature controller driver support available, have the
>> OCOTP provide feature control on the i.MX8MM to ensure the kernel DT
>> does not cause Linux to access the VPU and its p
Hi Ahamd,
On 22-08-30, Ahmad Fatoum wrote:
> Now with i.MX8M feature controller driver support available, have the
> OCOTP provide feature control on the i.MX8MM to ensure the kernel DT
> does not cause Linux to access the VPU and its power domains,
> when barebox knows them to be unavailable.
>
From: Ahmad Fatoum
This commit ports the axp20x driver from linux kernel code. Structures
not applicable for barebox are being dropped.
Signed-off-by: Ahmad Fatoum
Signed-off-by: Johannes Zink
---
drivers/mfd/Kconfig| 8 +
drivers/mfd/Makefile | 1 +
drivers/mfd/axp20x-i2c.c
The i.MX7 GPMI NAND controller is completely backwards compatible with
the version of the IP on the i.MX6. There are additional features like
increased ECC strength, that we may want to use in future, but for now
treating it the same as an i.MX6 works without problems.
Since the gpmi_nand driver r
Now with i.MX8M feature controller driver support available, have the
OCOTP provide feature control on the i.MX8MM to ensure the kernel DT
does not cause Linux to access the VPU and its power domains,
when barebox knows them to be unavailable.
This is needed because the upstream kernel imx8mm.dtsi
Save some bytes on network traffic by reducing the server response for
lookup requests.
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index 7181c58d1083..16e57daaef37 100644
--- a/fs/tftp.c
+++ b/fs/tftp.c
Opening a tftp is done in two steps: at first `tftp_lookup()` is
called to get the filesize and then it is opened again and data are
read.
The `tftp_lookup()` call sends only a RRQ/WRQ, reads then the "tsize"
from the response and closes the transfer by sending an error datagram.
The tftp server w
Results (with the reorder patch; numbers in bytes/s) on an iMX8MP are:
| windowsize | VPN | 1 Gb/s | 100 Mb/s |
||---|||
| 128| 3.869.284 | 98.643.085 | 11.434.852 |
| 64| 3.863.581 | 98.550.375 | 11.434.852 |
| 48
Unittest for window cache functions.
Signed-off-by: Enrico Scholz
---
fs/tftp-selftest.h | 56
fs/tftp.c | 106 -
test/self/Kconfig | 7 +++
3 files changed, 168 insertions(+), 1 deletion(-)
create mode 100644 fs/
The tftp "windowsize" greatly improves the performance of tftp
transfers. This patchset adds support for it.
The first two patches are a little bit unrelated and enhance the 'cp
-v' output by giving information about the transfer speed. They can
be dropped if they are unwanted.
I tested the fun
Is a noop in normal cases (when compiler sees that the condition can
be evaluated without sideeffects) but allows optimizations based on
the condition.
E.g. in
| void foo(int a)
| {
| debug_assert(a == 23);
|
| if (a == 23)
| return;
|
| bar();
| }
the call to 'bar()' will be
This adds a 'port=' mount option for tftp filesystems.
Useful e.g. when working with a local, non-privileged tftp server
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 7 +--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index c26204ae76e6..913ca1df6e42 1
Use the actual blocksize for allocating buffers instead of assuming an
hardcoded value.
This requires to add an additional 'START' state which is entered
after receiving (RRQ) or sending (WRQ) the OACK. Without it, the next
state would be entered and the (not allocated yet) fifo be used.
For non
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index d186e7983a6d..c26204ae76e6 100644
--- a/fs/tftp.c
+++ b/fs/tftp.c
@@ -94,7 +94,7 @@ static int tftp_truncate(struct device_d *dev, FILE *f,
loff_t size)
Having 11 printf arguments with lot of them being 0, makes it
difficulty to read and extend.
Add some comments and use '\0' for %c.
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index c1a193711793.
Some refactoring; makes next patches cleaner.
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index 361661d2180e..c1a193711793 100644
--- a/fs/tftp.c
+++ b/fs/tftp.c
@@ -254,6 +254,7 @@ static void tf
The filesize is not known for push requests and barebox always sent
'0'. Server might reject data because it will always exceed this
length.
Send this option only for RRQ requests.
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 13 +
1 file changed, 9 insertions(+), 4 deletions(-)
d
Signed-off-by: Enrico Scholz
---
Documentation/filesystems/tftp.rst | 38 ++
1 file changed, 38 insertions(+)
diff --git a/Documentation/filesystems/tftp.rst
b/Documentation/filesystems/tftp.rst
index a292765e2511..8929213d3c4a 100644
--- a/Documentation/filesystems/
Catch bad 'blocksize' or 'windowsize' responses from the server.
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 19 ---
1 file changed, 16 insertions(+), 3 deletions(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index 64a94797cda3..4e31adcd60ed 100644
--- a/fs/tftp.c
+++ b/fs/tftp.c
@@ -2
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index 0b2a420b6639..264841f24407 100644
--- a/fs/tftp.c
+++ b/fs/tftp.c
@@ -65,6 +65,7 @@
#define STATE_DONE 8
#define TFTP_BLOCK_SIZE
it should never happen due to the program logic; but detect a failed
kfifo_put() just in case...
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 11 +--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index 5b0dcb2b75d6..e62ff2a80bb5 100644
--- a/fs/tftp.
With tftp window size support in the next patches, the first received
block might be !=1 (e.g. when it was reordered or dropped). There could
be checked whether it is in the first window, but the corresponding
sanity check can be dropped completely:
- OACK logic verifies that we speak with a tftp
Signed-off-by: Enrico Scholz
---
commands/tftp.c | 22 +++---
1 file changed, 19 insertions(+), 3 deletions(-)
diff --git a/commands/tftp.c b/commands/tftp.c
index 48ff00c6217c..6ac822c9e832 100644
--- a/commands/tftp.c
+++ b/commands/tftp.c
@@ -21,15 +21,24 @@ static int do_tftp
These packets are valid in certain points of the transfer only and
accepting them too early or too late can corrupt internal states.
Reject them when they are unexpected.
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 12
1 file changed, 12 insertions(+)
diff --git a/fs/tftp.c b/fs/
move block handling into dedicated function
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 37 +
1 file changed, 21 insertions(+), 16 deletions(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index 16e57daaef37..5b0dcb2b75d6 100644
--- a/fs/tftp.c
+++ b/fs/tftp.c
@@ -298
attribute is not used outside tftp_recv() for RRQ.
Signed-off-by: Enrico Scholz
---
fs/tftp.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/fs/tftp.c b/fs/tftp.c
index 88767223b052..c6c7a2e3c887 100644
--- a/fs/tftp.c
+++ b/fs/tftp.c
@@ -425,7 +425,7 @@ static void tf
With the tftp "windowsize" option, reordering of udp datagrams becomes
an issue. Depending on the network topology, this reordering occurs
several times with large tftp transfers and will heavily reduce the
transfer speed.
This patch adds a packet cache so that datagrams can be reassembled in
the
Hello Sascha,
On 30.08.22 09:32, Sascha Hauer wrote:
> Hi Ahmad,
>
> On Thu, Aug 18, 2022 at 07:19:45AM +0200, Ahmad Fatoum wrote:
>> The i.MX8MM exists in a Lite variant with no VPUs as well as Solo and
>> Dual variants with one or two cores respectively instead of the default
>> four. For i.MX6
On Fri, Aug 26, 2022 at 08:49:53AM +0200, Steffen Trumtrar wrote:
> Since commit 3e62b38ff641f263df2f6c9e3ebda3c7c62f310b
> we use ENTRY_FUNCTION_WITHSTACK in the bringup barebox.
> The patch should have removed the arm_setup_stack line
> as it is already done in the ENTRY_FUNCTION_WITHSTACK macro.
Hi Ahmad,
On Thu, Aug 18, 2022 at 07:19:45AM +0200, Ahmad Fatoum wrote:
> The i.MX8MM exists in a Lite variant with no VPUs as well as Solo and
> Dual variants with one or two cores respectively instead of the default
> four. For i.MX6, we had a manual fixup taking care of deleting the
> excess CP
On Mon, Aug 29, 2022 at 12:52:16PM +0200, Ahmad Fatoum wrote:
> Hello Enrico,
>
> On 28.08.22 16:02, Enrico Scholz wrote:
> > The "tftp: allocate buffers and fifo dynamically" patch in the last
> > patchset broke interaction with non rfc 2347 servers (e.g. when data
> > transfer starts immediately
42 matches
Mail list logo