| Message ID | 4dee5a89157f3dd0e0a8d17ab6ff42292bb7ce79.1721910373.git.jerome.forissier@linaro.org |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Introduce the lwIP network stack | expand |
Hi Jerome, On Thu, 25 Jul 2024 at 15:59, Jerome Forissier <jerome.forissier@linaro.org> wrote: > > Add support for setting the TFTP block size. The default value (1468) > is fine for Ethernet and allows a better throughput than the TFTP > default (512), if the server supports the blksize option of course. > > I tested this change with qemu_arm64_lwip_defconfig. The throughput is > now 875 KiB/s vs. 313 KiB/s before. That is still a low number, but I > think we can't expect more without implementing the windowsize option. > > Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org> > --- > net/Kconfig | 20 ++++++++++---------- > net/lwip/tftp.c | 2 ++ > 2 files changed, 12 insertions(+), 10 deletions(-) > > diff --git a/net/Kconfig b/net/Kconfig > index 952690d677d..7e5406bb923 100644 > --- a/net/Kconfig > +++ b/net/Kconfig > @@ -69,16 +69,6 @@ config SYS_FAULT_ECHO_LINK_DOWN > this option is active, then CONFIG_SYS_FAULT_MII_ADDR also needs to > be configured. > > -config TFTP_BLOCKSIZE > - int "TFTP block size" > - default 1468 > - help > - Default TFTP block size. > - The MTU is typically 1500 for ethernet, so a TFTP block of > - 1468 (MTU minus eth.hdrs) provides a good throughput with > - almost-MTU block sizes. > - You can also activate CONFIG_IP_DEFRAG to set a larger block. > - Why do you have to move the config option? Cheers /Ilias > config TFTP_PORT > bool "Set TFTP UDP source/destination ports via the environment" > help > @@ -263,4 +253,14 @@ config SYS_RX_ETH_BUFFER > since all buffers can be full shortly after enabling the interface on > high Ethernet traffic. > > +config TFTP_BLOCKSIZE > + int "TFTP block size" > + default 1468 > + help > + Default TFTP block size. > + The MTU is typically 1500 for ethernet, so a TFTP block of > + 1468 (MTU minus eth.hdrs) provides a good throughput with > + almost-MTU block sizes. > + You can also activate CONFIG_IP_DEFRAG to set a larger block. > + > endif # if NET || NET_LWIP > diff --git a/net/lwip/tftp.c b/net/lwip/tftp.c > index 78de0bd0dba..1fe948c84ec 100644 > --- a/net/lwip/tftp.c > +++ b/net/lwip/tftp.c > @@ -136,6 +136,8 @@ static int tftp_loop(struct udevice *udev, ulong addr, char *fname, > if (!(err == ERR_OK || err == ERR_USE)) > log_err("tftp_init_client err: %d\n", err); > > + tftp_client_set_blksize(CONFIG_TFTP_BLOCKSIZE); > + > ctx.start_time = get_timer(0); > err = tftp_get(&ctx, &srvip, srvport, fname, TFTP_MODE_OCTET); > /* might return different errors, like routing problems */ > -- > 2.40.1 >
Hi Ilias, On 7/29/24 14:29, Ilias Apalodimas wrote: > Hi Jerome, > > On Thu, 25 Jul 2024 at 15:59, Jerome Forissier > <jerome.forissier@linaro.org> wrote: >> >> Add support for setting the TFTP block size. The default value (1468) >> is fine for Ethernet and allows a better throughput than the TFTP >> default (512), if the server supports the blksize option of course. >> >> I tested this change with qemu_arm64_lwip_defconfig. The throughput is >> now 875 KiB/s vs. 313 KiB/s before. That is still a low number, but I >> think we can't expect more without implementing the windowsize option. >> >> Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org> >> --- >> net/Kconfig | 20 ++++++++++---------- >> net/lwip/tftp.c | 2 ++ >> 2 files changed, 12 insertions(+), 10 deletions(-) >> >> diff --git a/net/Kconfig b/net/Kconfig >> index 952690d677d..7e5406bb923 100644 >> --- a/net/Kconfig >> +++ b/net/Kconfig >> @@ -69,16 +69,6 @@ config SYS_FAULT_ECHO_LINK_DOWN >> this option is active, then CONFIG_SYS_FAULT_MII_ADDR also needs to >> be configured. >> >> -config TFTP_BLOCKSIZE >> - int "TFTP block size" >> - default 1468 >> - help >> - Default TFTP block size. >> - The MTU is typically 1500 for ethernet, so a TFTP block of >> - 1468 (MTU minus eth.hdrs) provides a good throughput with >> - almost-MTU block sizes. >> - You can also activate CONFIG_IP_DEFRAG to set a larger block. >> - > > Why do you have to move the config option? It is moved from the 'if NET' block to the 'if NET || NET_LWIP' block since it now applies to both. Thanks,
On Mon, 29 Jul 2024 at 18:18, Jerome Forissier <jerome.forissier@linaro.org> wrote: > > Hi Ilias, > > On 7/29/24 14:29, Ilias Apalodimas wrote: > > Hi Jerome, > > > > On Thu, 25 Jul 2024 at 15:59, Jerome Forissier > > <jerome.forissier@linaro.org> wrote: > >> > >> Add support for setting the TFTP block size. The default value (1468) > >> is fine for Ethernet and allows a better throughput than the TFTP > >> default (512), if the server supports the blksize option of course. > >> > >> I tested this change with qemu_arm64_lwip_defconfig. The throughput is > >> now 875 KiB/s vs. 313 KiB/s before. That is still a low number, but I > >> think we can't expect more without implementing the windowsize option. > >> > >> Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org> > >> --- > >> net/Kconfig | 20 ++++++++++---------- > >> net/lwip/tftp.c | 2 ++ > >> 2 files changed, 12 insertions(+), 10 deletions(-) > >> > >> diff --git a/net/Kconfig b/net/Kconfig > >> index 952690d677d..7e5406bb923 100644 > >> --- a/net/Kconfig > >> +++ b/net/Kconfig > >> @@ -69,16 +69,6 @@ config SYS_FAULT_ECHO_LINK_DOWN > >> this option is active, then CONFIG_SYS_FAULT_MII_ADDR also needs to > >> be configured. > >> > >> -config TFTP_BLOCKSIZE > >> - int "TFTP block size" > >> - default 1468 > >> - help > >> - Default TFTP block size. > >> - The MTU is typically 1500 for ethernet, so a TFTP block of > >> - 1468 (MTU minus eth.hdrs) provides a good throughput with > >> - almost-MTU block sizes. > >> - You can also activate CONFIG_IP_DEFRAG to set a larger block. > >> - > > > > Why do you have to move the config option? > > It is moved from the 'if NET' block to the 'if NET || NET_LWIP' block since > it now applies to both. > > Thanks, > -- > Jerome > > > > > Cheers > > /Ilias > >> config TFTP_PORT > >> bool "Set TFTP UDP source/destination ports via the environment" > >> help > >> @@ -263,4 +253,14 @@ config SYS_RX_ETH_BUFFER > >> since all buffers can be full shortly after enabling the interface on > >> high Ethernet traffic. > >> > >> +config TFTP_BLOCKSIZE > >> + int "TFTP block size" > >> + default 1468 > >> + help > >> + Default TFTP block size. > >> + The MTU is typically 1500 for ethernet, so a TFTP block of > >> + 1468 (MTU minus eth.hdrs) provides a good throughput with > >> + almost-MTU block sizes. > >> + You can also activate CONFIG_IP_DEFRAG to set a larger block. > >> + > >> endif # if NET || NET_LWIP > >> diff --git a/net/lwip/tftp.c b/net/lwip/tftp.c > >> index 78de0bd0dba..1fe948c84ec 100644 > >> --- a/net/lwip/tftp.c > >> +++ b/net/lwip/tftp.c > >> @@ -136,6 +136,8 @@ static int tftp_loop(struct udevice *udev, ulong addr, char *fname, > >> if (!(err == ERR_OK || err == ERR_USE)) > >> log_err("tftp_init_client err: %d\n", err); > >> > >> + tftp_client_set_blksize(CONFIG_TFTP_BLOCKSIZE); > >> + > >> ctx.start_time = get_timer(0); > >> err = tftp_get(&ctx, &srvip, srvport, fname, TFTP_MODE_OCTET); > >> /* might return different errors, like routing problems */ > >> -- > >> 2.40.1 > >> Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
diff --git a/net/Kconfig b/net/Kconfig index 952690d677d..7e5406bb923 100644 --- a/net/Kconfig +++ b/net/Kconfig @@ -69,16 +69,6 @@ config SYS_FAULT_ECHO_LINK_DOWN this option is active, then CONFIG_SYS_FAULT_MII_ADDR also needs to be configured. -config TFTP_BLOCKSIZE - int "TFTP block size" - default 1468 - help - Default TFTP block size. - The MTU is typically 1500 for ethernet, so a TFTP block of - 1468 (MTU minus eth.hdrs) provides a good throughput with - almost-MTU block sizes. - You can also activate CONFIG_IP_DEFRAG to set a larger block. - config TFTP_PORT bool "Set TFTP UDP source/destination ports via the environment" help @@ -263,4 +253,14 @@ config SYS_RX_ETH_BUFFER since all buffers can be full shortly after enabling the interface on high Ethernet traffic. +config TFTP_BLOCKSIZE + int "TFTP block size" + default 1468 + help + Default TFTP block size. + The MTU is typically 1500 for ethernet, so a TFTP block of + 1468 (MTU minus eth.hdrs) provides a good throughput with + almost-MTU block sizes. + You can also activate CONFIG_IP_DEFRAG to set a larger block. + endif # if NET || NET_LWIP diff --git a/net/lwip/tftp.c b/net/lwip/tftp.c index 78de0bd0dba..1fe948c84ec 100644 --- a/net/lwip/tftp.c +++ b/net/lwip/tftp.c @@ -136,6 +136,8 @@ static int tftp_loop(struct udevice *udev, ulong addr, char *fname, if (!(err == ERR_OK || err == ERR_USE)) log_err("tftp_init_client err: %d\n", err); + tftp_client_set_blksize(CONFIG_TFTP_BLOCKSIZE); + ctx.start_time = get_timer(0); err = tftp_get(&ctx, &srvip, srvport, fname, TFTP_MODE_OCTET); /* might return different errors, like routing problems */
Add support for setting the TFTP block size. The default value (1468) is fine for Ethernet and allows a better throughput than the TFTP default (512), if the server supports the blksize option of course. I tested this change with qemu_arm64_lwip_defconfig. The throughput is now 875 KiB/s vs. 313 KiB/s before. That is still a low number, but I think we can't expect more without implementing the windowsize option. Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org> --- net/Kconfig | 20 ++++++++++---------- net/lwip/tftp.c | 2 ++ 2 files changed, 12 insertions(+), 10 deletions(-)