Re: [PATCH 1/3] push/pop errno in initrd read file path

On 11/10/2016 07:49 AM, Daniel Kiper wrote: On Fri, Mar 11, 2016 at 11:28:33AM -0500, Josef Bacik wrote: If you try to load an initrd from http and it errors out we will free the initrd context but continue on because net_tcp_socket_close() will reset the grub_errno as will grub_initrd_close

[PATCH] efi term: open input console exlusively

the guy behind the wheel gets to type. With this patch we can no provision these weird machines. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/term/efi/console.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/grub-core/term/efi/console.c b/grub-core/te

Re: [PATCH 1/3] push/pop errno in initrd read file path

On 03/11/2016 02:34 PM, Vladimir 'phcoder' Serbinenko wrote: Le ven. 11 mars 2016 19:13, Josef Bacik <jba...@fb.com <mailto:jba...@fb.com>> a écrit : On 03/11/2016 12:23 PM, Vladimir 'phcoder' Serbinenko wrote: > > > On Friday, March 11, 2016, Josef

Re: [PATCH 1/3] push/pop errno in initrd read file path

On 03/11/2016 12:23 PM, Vladimir 'phcoder' Serbinenko wrote: On Friday, March 11, 2016, Josef Bacik <jba...@fb.com <mailto:jba...@fb.com>> wrote: If you try to load an initrd from http and it errors out we will free the initrd context but continue on because net_tcp_

[PATCH 1/3] push/pop errno in initrd read file path

to the appropriate places so we don't lose our errno. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/loader/linux.c | 2 ++ grub-core/net/http.c | 2 ++ 2 files changed, 4 insertions(+) diff --git a/grub-core/loader/linux.c b/grub-core/loader/linux.c index be6fa0f..bd61ca2

[PATCH 3/3] pxenet: process transmit interrupts when out of resources

a netbuff behind for the next call into recv. With this patch we are now able to properly provision boxes suffering from this problem. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/drivers/i386/pc/pxe.c | 155 +--- 1 file change

[PATCH 2/3] tcp: add a dprintf for opening tcp connections

In debugging strange timeouts and other network problems it has been helpful to see when we're opening new connections to get an idea of where we're having a breakdown. Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/tcp.c | 7 +++ 1 file changed, 7 insertions(+) diff

Re: [PATCH] dns: realloc address buffer after each packet

On 03/01/2016 02:22 PM, Andrei Borzenkov wrote: 01.03.2016 19:38, Josef Bacik пишет: On 02/27/2016 12:39 PM, Andrei Borzenkov wrote: 26.02.2016 16:52, Josef Bacik пишет: On 02/26/2016 05:22 AM, Andrei Borzenkov wrote: On Wed, Feb 24, 2016 at 10:11 PM, Josef Bacik <jba...@fb.com>

Re: [PATCH] dns: realloc address buffer after each packet

On 02/27/2016 12:39 PM, Andrei Borzenkov wrote: 26.02.2016 16:52, Josef Bacik пишет: On 02/26/2016 05:22 AM, Andrei Borzenkov wrote: On Wed, Feb 24, 2016 at 10:11 PM, Josef Bacik <jba...@fb.com> wrote: Sometimes DNS responses come in slower than we poll for them which can lead us to p

Re: [PATCH] bootp: use the ipv* version given by the dhcp server

On 02/27/2016 01:05 PM, Andrei Borzenkov wrote: 24.02.2016 22:10, Josef Bacik пишет: We were setting DNS_OPTION_PREVER_IPV* based on which version of the dhcp request we go back. However this would mean that we would mistakenly fall back on ipv4 in an ipv6 only system, or vice versa

Re: [PATCH] dns: realloc address buffer after each packet

On 02/26/2016 05:22 AM, Andrei Borzenkov wrote: On Wed, Feb 24, 2016 at 10:11 PM, Josef Bacik <jba...@fb.com> wrote: Sometimes DNS responses come in slower than we poll for them which can lead us to process multiple DNS packets which overflows the addresses array. So instead realloc the

[PATCH] dns: realloc address buffer after each packet

the caching of the addresses outside of the recv hook so we can be sure to cache all the responses at once instead of one packet at a time. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/dns.c | 65 - 1 file changed, 34 inse

[PATCH] tcp: add mss option support

awesome. Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/tcp.c | 15 ++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/grub-core/net/tcp.c b/grub-core/net/tcp.c index 6902c10..2d1706e 100644 --- a/grub-core/net/tcp.c +++ b/grub-core/net/tcp.c @@ -111,6

[PATCH] bootp: use the ipv* version given by the dhcp server

response. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/bootp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/grub-core/net/bootp.c b/grub-core/net/bootp.c index 345ad7a..cde818c 100644 --- a/grub-core/net/bootp.c +++ b/grub-core/net/bootp.c @@

Re: [PATCH 10/14] dns: poll card between each dns request

On 02/15/2016 01:45 AM, Andrei Borzenkov wrote: On Thu, Feb 11, 2016 at 12:21 AM, Josef Bacik <jba...@fb.com> wrote: If we have dns servers that we prefer to get records from we'll send a packet and immediately check data.naddresses to see if we got a response. If we didn't

Re: [PATCH V2] net: fix ipv6 routing

On 02/22/2016 03:07 AM, Andrei Borzenkov wrote: 22.01.2016 17:20, Josef Bacik пишет: On 01/22/2016 08:01 AM, Vladimir 'φ-coder/phcoder' Serbinenko wrote: Were Andrey's comments ever adressed? Other than his comments the patch looks good. I haven't addressed them yet, had other issues come up

Re: [PATCH v3] tcp: add window scaling and RTTM support

On 02/13/2016 10:40 AM, Andrei Borzenkov wrote: 01.02.2016 21:30, Josef Bacik пишет: ... @@ -745,6 +817,7 @@ grub_net_recv_tcp_packet (struct grub_net_buff *nb, { struct tcphdr *tcph; grub_net_tcp_socket_t sock; + grub_uint32_t tsecr = 0; grub_err_t err; /* Ignore broadcast

Re: [PATCH 11/14] dns: reset data->naddresses for every packet we receive

On 02/13/2016 11:05 AM, Andrei Borzenkov wrote: 11.02.2016 00:21, Josef Bacik пишет: I noticed when debugging a problem that we'd corrupt memory if our dns server didn't respond fast enough and we ended up asking for both an and A record for a server. The problem is we alloc data

Re: [PATCH 10/14] dns: poll card between each dns request

On 02/15/2016 01:45 AM, Andrei Borzenkov wrote: On Thu, Feb 11, 2016 at 12:21 AM, Josef Bacik <jba...@fb.com> wrote: If we have dns servers that we prefer to get records from we'll send a packet and immediately check data.naddresses to see if we got a response. If we didn't

Re: [PATCH 05/14] net: don't free uninitialized sockets in dns

On 02/13/2016 12:59 PM, Andrei Borzenkov wrote: 11.02.2016 00:20, Josef Bacik пишет: If we cannot open a connection to our dns server we will have NULL sockets in our array, so don't do the cleanup on any sockets that didn't get created. Was not it already fixed by Yes it was, sorry about

[PATCH 00/14] Facebook's netbooting patches

I've recently fixed up a bunch of new issues and rebased onto todays master so I figured I'd send our current backlog. So first the new patches that you guys haven't seen yet [PATCH 10/14] dns: poll card between each dns request [PATCH 11/14] dns: reset data->naddresses for every packet we

[PATCH 10/14] dns: poll card between each dns request

the packet and checking data.naddresses we'll send a request for the A record and then poll the card. Instead we need to make sure we poll after we issue each request to make sure we give the server enough time to respond to our initial request. Signed-off-by: Josef Bacik <jba...@fb.com> ---

[PATCH 04/14] tcp: add window scaling and RTTM support

set to 1mb but can be configured to whatever a user wants, and we will calculate the appropriate window size and scale settings. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/net.c | 1 + grub-core/net/tcp.c | 186 +-

[PATCH 14/14] net: add interfaces when we open a card

ave the two people that call ->open use the helper function instead. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/ethernet.c | 12 +++- grub-core/net/net.c | 43 +++ include/grub/net.h | 3 +++ 3 files

[PATCH 03/14] Use UEFI MAC device as default configured by net_bootp6

From: Michael Chang The grub_efinet_findcards will register cards by checking if it can support EFI Simple Netowork Protocol which create more than one device to a physical NIC device. If without specifying any device to be configured by net_bootp6, it should pick up one from

[PATCH 08/14] efinet: filter multicast traffic based on addresses

. Whenever we add a new address to the card we calculate the solicited node multicast address to the multicast filter. With this patch my broken hardware is still broken but functional. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/drivers/efi/efinet.

[PATCH 11/14] dns: reset data->naddresses for every packet we receive

esses memory. So fix this by noticing if we already have an address and free the old memory and reset naddresses so we don't overflow our new array. Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/dns.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/grub-core/net/dns

[PATCH 05/14] net: don't free uninitialized sockets in dns

If we cannot open a connection to our dns server we will have NULL sockets in our array, so don't do the cleanup on any sockets that didn't get created. Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/dns.c | 5 - 1 file changed, 4 insertions(+), 1 deletion(-) diff --git

[PATCH 02/14] UEFI IPv6 PXE support

From: Michael Chang When system is booted from UEFI IPv6 PXE, the network interface can be configured directly by cached DHCPv6 packet in firmware without doing it all over again by contacting DHCP server. --- grub-core/net/drivers/efi/efinet.c | 24 +++- 1

[PATCH 09/14] efinet: clear the txbuffer before modifying the receive filters

firmware now properly updates the multicast receive filters. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/drivers/efi/efinet.c | 30 ++ 1 file changed, 26 insertions(+), 4 deletions(-) diff --git a/grub-core/net/drivers/efi/efinet.c b/grub-co

[PATCH 06/14] net: fix ipv6 routing

used, we do this by setting the masksize to 0 to mean it encompasses all networks. The routing code will automatically select the best route so if there is a closer match we will use that. With this patch I can now talk to ipv6 addresses outside of my local network. Thanks, Signed-off-by: Josef Bacik

[PATCH 13/14] bootp: don't add multiple interfaces for the same address

and return. Also make these functions void as nobody uses the return value. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/bootp.c | 29 ++--- include/grub/net.h| 4 ++-- 2 files changed, 20 insertions(+), 13 deletions(-) diff --git

[PATCH 12/14] icmp6: use default interface as the route interface

getting RA's, check to see if our default interface is a v6 interface and matches the card we got the RA on, and if so use that as the route interface. This allows us to get traffic from the dhcp interface instead of through the slaac interface. Signed-off-by: Josef Bacik <jba...@fb.com> --- gru

[PATCH 01/14] Added net_bootp6 command

From: Michael Chang The net_bootp6 is used to configure the ipv6 network interface through the DHCPv6 protocol Solict/Advertise/Request/Reply. --- grub-core/net/bootp.c | 885 - grub-core/net/ip.c | 35 ++

[PATCH 07/14] efinet: retransmit if our device is busy

deal with transient transmission problems without flooding the box. This fixes an issue we were seeing in production. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/drivers/efi/efinet.c | 10 ++ 1 file changed, 10 insertions(+) diff --git a/grub-core/net/d

[PATCH] efinet: clear the txbuffer before modifying the receive filters

firmware now properly updates the multicast receive filters. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/drivers/efi/efinet.c | 30 ++ 1 file changed, 26 insertions(+), 4 deletions(-) diff --git a/grub-core/net/drivers/efi/efinet.c b/grub-co

Re: [PATCH V2] tcp: add window scaling and RTTM support

On 02/01/2016 03:43 AM, Andrei Borzenkov wrote: On Sat, Jan 30, 2016 at 12:48 AM, Josef Bacik <jba...@fb.com> wrote: Sometimes we have to provision boxes across regions, such as California to Sweden. The http server has a 10 minute timeout, so if we can't get our 250mb image transferre

Re: [PATCH V2] tcp: add window scaling and RTTM support

On 02/01/2016 03:43 AM, Andrei Borzenkov wrote: On Sat, Jan 30, 2016 at 12:48 AM, Josef Bacik <jba...@fb.com> wrote: Sometimes we have to provision boxes across regions, such as California to Sweden. The http server has a 10 minute timeout, so if we can't get our 250mb image transferre

[PATCH v3] tcp: add window scaling and RTTM support

set to 1mb but can be configured to whatever a user wants, and we will calculate the appropriate window size and scale settings. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- v2->v3: -Addressed all of Andrei's comments. -Fixed the tcp_window_size option to set the global window s

[PATCH V2] tcp: add window scaling and RTTM support

set to 1mb but can be configured to whatever a user wants, and we will calculate the appropriate window size and scale settings. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- V1->V2: -Address Andrei's concerns about making the window size configurable. -Also make the tcp option

Re: [PATCH V2] net: fix ipv6 routing

On 01/22/2016 08:01 AM, Vladimir 'φ-coder/phcoder' Serbinenko wrote: Were Andrey's comments ever adressed? Other than his comments the patch looks good. I haven't addressed them yet, had other issues come up and this is working for us fine in production. I'm getting a test environment set

Re: [PATCH] tcp: ack when we get an OOO/lost packet

On 12/07/2015 12:59 PM, Andrei Borzenkov wrote: 12.08.2015 18:16, Josef Bacik пишет: While adding tcp window scaling support I was finding that I'd get some packet loss or reordering when transferring from large distances and grub would just timeout. This is because we weren't ack'ing when we

[PATCH REBASED] net: reset nb->data per dns record lookup loop

We were resetting nb->data every time we tried a new server, but we need to do it every time we try for a different record, otherwise we don't end up falling back to the A record properly. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- -This is a rebase onto this mornings maste

Re: [PATCH] efinet: filter multicast traffic based on addresses

On 11/20/2015 06:02 AM, Andrei Borzenkov wrote: 17.11.2015 21:35, Josef Bacik пишет: We have some hardware that claims to support PROMISCUOUS_MULTICAST but doesn't actually work. Instead utilize the multicast filters and specifically enable the multicast traffic we care about. In reality we

Re: [PATCH] net: reset nb->data per dns record lookup loop

On 11/20/2015 09:47 AM, Andrei Borzenkov wrote: ping I'll rebase today, in the middle of something else atm. Thanks, Josef ___ Grub-devel mailing list Grub-devel@gnu.org https://lists.gnu.org/mailman/listinfo/grub-devel

[PATCH] efinet: filter multicast traffic based on addresses

. Whenever we add a new address to the card we calculate the solicited node multicast address to the multicast filter. With this patch my broken hardware is still broken but functional. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/drivers/efi/efinet.

[PATCH] net: reset nb->data per dns record lookup loop

We were resetting nb->data every time we tried a new server, but we need to do it every time we try for a different record, otherwise we don't end up falling back to the A record properly. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/dns.c | 2 +- 1 file c

Re: [PATCH] efinet: check for broken firmware

On 11/14/2015 09:13 AM, Vladimir 'phcoder' Serbinenko wrote: Le 14 nov. 2015 5:08 AM, "Josef Bacik" <jba...@fb.com <mailto:jba...@fb.com>> a écrit : > > > > On Nov 13, 2015, at 10:19 PM, Andrei Borzenkov <arvidj...@gmail.com <mailto:arvidj...@gmail.c

Re: [PATCH] efinet: check for broken firmware

On 11/13/2015 09:38 AM, Andrei Borzenkov wrote: On Fri, Nov 13, 2015 at 5:30 PM, Josef Bacik <jba...@fb.com> wrote: @@ -291,6 +286,9 @@ grub_efinet_findcards (void) 0, _handles); if (! handles) return; + + attributes = grub_snp_attr

Re: [PATCH] efinet: check for broken firmware

On 11/13/2015 09:39 AM, Andrei Borzenkov wrote: On Fri, Nov 13, 2015 at 5:38 PM, Vladimir 'phcoder' Serbinenko wrote: Please try the patch currently used in Solaris flavour of GRUB. I think of upstreaming their mnp driver where can we get it? It is in part2 of the

Re: [PATCH] efinet: check for broken firmware

On 11/13/2015 09:38 AM, Vladimir 'phcoder' Serbinenko wrote: Please try the patch currently used in Solaris flavour of GRUB. I think of upstreaming their mnp driver This driver doesn't work at all for me. I may have done a bad job porting it, or the firmware is even crappier with MNP, either

Re: [PATCH] efinet: check for broken firmware

> On Nov 13, 2015, at 10:19 PM, Andrei Borzenkov <arvidj...@gmail.com> wrote: > > 13.11.2015 22:34, Josef Bacik пишет: >>> On 11/13/2015 09:38 AM, Vladimir 'phcoder' Serbinenko wrote: >>> Please try the patch currently used in Solaris flavour of GRUB. I think &g

Re: [PATCH] efinet: check for broken firmware

On 11/13/2015 08:10 AM, Andrei Borzenkov wrote: On Fri, Nov 13, 2015 at 1:07 AM, Josef Bacik <jba...@fb.com> wrote: The firmware bug I've been tracking down has been too extensive to work around effectively. Basically once we switch to EXCLUSIVE everything is completely horked. I c

[PATCH] efinet: check for broken firmware

instead of EXCLUSIVE. This also keeps the original protocol we use when scanning the cards and leaves the initialization stuff for ->open. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/net/drivers/efi/efinet.c | 99 ++ 1 file ch

Re: [PATCH] efinet: add efinet_multicast_filter command

On 11/05/2015 03:28 PM, Vladimir 'phcoder' Serbinenko wrote: I don't have EFI spec under my hand now. Can we get away with making it a default or at least for the case when no interface overrides mac address. Extra config to workaround firmware bugs is usually harmful So I made us start

Re: [PATCH] efinet: add efinet_multicast_filter command

On 11/05/2015 11:15 PM, Andrei Borzenkov wrote: 05.11.2015 22:23, Josef Bacik пишет: We have some hardware that doesn't honor EFI_SIMPLE_NETWORK_RECEIVE_PROMISCUOUS_MULTICAST properly so we aren't getting You mean that driver advertises promiscuous multicast support but does not implement

[PATCH] efinet: add efinet_multicast_filter command

command to allow a user to specify a multicast receive filter. We use it like this efinet_multicast_filter efinet0 33:33:0:0:0:1 to get ipv6 multicasts which allows us to receive the router advertisements. Thanks, Signed-off-by: Josef Bacik <jba...@fb.com> --- grub-core/kern/efi

Re: [PATCH] efinet: add efinet_multicast_filter command

On 11/05/2015 03:28 PM, Vladimir 'phcoder' Serbinenko wrote: I don't have EFI spec under my hand now. Can we get away with making it a default or at least for the case when no interface overrides mac address. Extra config to workaround firmware bugs is usually harmful

Re: [PATCH] net: don't free uninitialized sockets in dns

On 10/10/2015 03:03 AM, Andrei Borzenkov wrote: ping? Oops sorry forgot to tell you this patch worked as well. Thanks, Josef ___ Grub-devel mailing list Grub-devel@gnu.org https://lists.gnu.org/mailman/listinfo/grub-devel

Re: GNU GRUB maintenance

On 10/07/2015 02:14 PM, Vladimir 'φ-coder/phcoder' Serbinenko wrote: Hello, all. I'm sorry for not being available to do enough maintenance for GRUB in last time but I was overbooked. Yet there is a good news. At Google there is a 20% project and GRUB has been approved as 20% project for me. The

Re: [PATCH] efinet: retransmit if our device is busy

On 09/11/2015 12:49 PM, Andrei Borzenkov wrote: 31.08.2015 20:34, Josef Bacik пишет: When I fixed the txbuf handling I ripped out the retransmission code since it was flooding our network when we had the buggy behavior. Turns out this was too heavy handed as we can still have transient tx

[PATCH V2] net: fix ipv6 routing

used, we do this by setting the masksize to 0 to mean it encompasses all networks. The routing code will automatically select the best route so if there is a closer match we will use that. With this patch I can now talk to ipv6 addresses outside of my local network. Thanks, Signed-off-by: Josef Bacik

Re: GRUB release schedule?

On 07/20/2015 11:22 AM, Peter Jones wrote: Hi everyone, Is there a plan for when upcoming GNU GRUB releases will happen? As far as I can tell, the last official release on ftp://ftp.gnu.org/gnu/grub/ was 2.00 on 28-Jun-2012, and the last beta on http://alpha.gnu.org/pub/gnu/grub/ for the next

Re: GRUB release schedule?

On 08/21/2015 10:11 AM, Konrad Rzeszutek Wilk wrote: On Fri, Aug 21, 2015 at 09:56:59AM -0700, Josef Bacik wrote: On 07/20/2015 11:22 AM, Peter Jones wrote: Hi everyone, Is there a plan for when upcoming GNU GRUB releases will happen? As far as I can tell, the last official release on ftp

Re: GRUB release schedule?

On 08/21/2015 10:30 AM, Konrad Rzeszutek Wilk wrote: On Fri, Aug 21, 2015 at 10:18:08AM -0700, Josef Bacik wrote: On 08/21/2015 10:11 AM, Konrad Rzeszutek Wilk wrote: On Fri, Aug 21, 2015 at 09:56:59AM -0700, Josef Bacik wrote: On 07/20/2015 11:22 AM, Peter Jones wrote: Hi everyone

Re: [PATCH] tcp: ack when we get an OOO/lost packet

On 08/17/2015 05:38 AM, Andrei Borzenkov wrote: On Thu, Aug 13, 2015 at 4:59 PM, Josef Bacik jba...@fb.com wrote: On 08/13/2015 04:19 AM, Andrei Borzenkov wrote: On Wed, Aug 12, 2015 at 6:16 PM, Josef Bacik jba...@fb.com wrote: While adding tcp window scaling support I was finding that I'd

[PATCH] net: don't free uninitialized sockets in dns

If we cannot open a connection to our dns server we will have NULL sockets in our array, so don't do the cleanup on any sockets that didn't get created. Signed-off-by: Josef Bacik jba...@fb.com --- grub-core/net/dns.c | 5 - 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/grub

Re: [PATCH] tcp: add window scaling support

On 08/17/2015 09:15 AM, Andrei Borzenkov wrote: 12.08.2015 18:57, Josef Bacik пишет: Sometimes we have to provision boxes across regions, such as California to Sweden. The http server has a 10 minute timeout, so if we can't get our 250mb image transferred fast enough our provisioning fails

Re: [PATCH] tcp: ack when we get an OOO/lost packet

On 08/13/2015 04:19 AM, Andrei Borzenkov wrote: On Wed, Aug 12, 2015 at 6:16 PM, Josef Bacik jba...@fb.com wrote: While adding tcp window scaling support I was finding that I'd get some packet loss or reordering when transferring from large distances and grub would just timeout

Re: [PATCH] tcp: ack when we get an OOO/lost packet

On 08/13/2015 01:13 PM, Andrei Borzenkov wrote: On 13.08.2015 16:59, Josef Bacik wrote: On 08/13/2015 04:19 AM, Andrei Borzenkov wrote: On Wed, Aug 12, 2015 at 6:16 PM, Josef Bacik jba...@fb.com wrote: While adding tcp window scaling support I was finding that I'd get some packet loss

[PATCH] tcp: ack when we get an OOO/lost packet

eventually it would fill the window and stop transmitting, and we'd time out. Fix this by ACK'ing when we don't find our next sequence numbered packet. With this fix I no longer time out. Thanks, Signed-off-by: Josef Bacik jba...@fb.com --- grub-core/net/tcp.c | 5 - 1 file changed, 4 insertions

[PATCH] tcp: add window scaling support

to 1mb. With this change we're able to get higher sustained transfers between regions and can transfer our image in well below 10 minutes. Without this patch we'd time out every time halfway through the transfer. Thanks, Signed-off-by: Josef Bacik jba...@fb.com --- grub-core/net/tcp.c | 42

Re: [PATCH 3/3] net: fix ipv6 routing

On 08/09/2015 10:58 AM, Andrei Borzenkov wrote: В Wed, 5 Aug 2015 14:36:39 -0400 Josef Bacik jba...@fb.com пишет: Currently we cannot talk to ipv6 addresses outside of our local link area because we don't setup our routes properly nor do we talk to the router properly. Currently

[PATCH V2] efinet: handle get_status() on buggy firmware properly

have the SNP stuff opened in exclusive mode just assume any non-NULL txbuf means that our transmit occurred properly. This makes grub able to do its networking stuff properly on our broken firmware. Thanks, cc: Peter Jones pjo...@redhat.com Signed-off-by: Josef Bacik jba...@fb.com --- V1-V2

Re: [PATCH 0/3] fix ipv6 support

On 08/06/2015 12:06 AM, Michael Chang wrote: On Wed, Aug 05, 2015 at 10:50:37AM -0700, Josef Bacik wrote: These patches are on top of Michael Chang's bootp6 code (which we also really need so it would be great if those could go in as well). At Facebook we have ipv6 only clusters that we need

Re: [PATCH 1/3] efinet: handle get_status() properly

On 08/05/2015 11:42 PM, Andrei Borzenkov wrote: В Wed, 5 Aug 2015 22:32:13 +0200 Vladimir 'phcoder' Serbinenko phco...@gmail.com пишет: This patch improperly assumes that GRUB is the only thing in EFI that transmits. Actually since recently we try to ensure that grub *is* the only user of

[PATCH 0/3] fix ipv6 support

These patches are on top of Michael Chang's bootp6 code (which we also really need so it would be great if those could go in as well). At Facebook we have ipv6 only clusters that we need to be able to provision over the network. The current grub2 support for ipv6 is broken in a few ways. The

[PATCH 3/3] net: fix ipv6 routing

now do net_ipv6_autoconf and immediately talk to ipv6 addresses outside of my local network. Thanks, Signed-off-by: Josef Bacik jba...@fb.com --- grub-core/net/bootp.c | 2 +- grub-core/net/drivers/ieee1275/ofnet.c | 2 +- grub-core/net/icmp6.c | 50

[PATCH 1/3] efinet: handle get_status() properly

transmit one buffer at a time, just assume that a non-NULL txbuf means that our transmit happened properly. With this patch I can now do networking on our EFI enabled boxes. Thanks, cc: Peter Jones pjo...@redhat.com Signed-off-by: Josef Bacik jba...@fb.com --- grub-core/net/drivers/efi/efinet.c

[PATCH 2/3] net: add local route when creating link local ipv6 interface

In order to talk to link local ipv6 addresses we need to have a route out of the link local interface for this to work. Thanks, Signed-off-by: Josef Bacik jba...@fb.com --- grub-core/net/net.c | 14 +- 1 file changed, 13 insertions(+), 1 deletion(-) diff --git a/grub-core/net/net.c

[PATCH 2/3] net: add local route when creating link local ipv6 interface

In order to talk to link local ipv6 addresses we need to have a route out of the link local interface for this to work. Thanks, Signed-off-by: Josef Bacik jba...@fb.com --- grub-core/net/net.c | 14 +- 1 file changed, 13 insertions(+), 1 deletion(-) diff --git a/grub-core/net/net.c

[PATCH 1/3] efinet: handle get_status() properly

transmit one buffer at a time, just assume that a non-NULL txbuf means that our transmit happened properly. With this patch I can now do networking on our EFI enabled boxes. Thanks, cc: Peter Jones pjo...@redhat.com Signed-off-by: Josef Bacik jba...@fb.com --- grub-core/net/drivers/efi/efinet.c

Re: [PATCH 1/3] efinet: handle get_status() properly

On 08/05/2015 04:04 PM, Andrei Borzenkov wrote: В Wed, 5 Aug 2015 14:36:37 -0400 Josef Bacik jba...@fb.com пишет: The EFI SNP documentation isn't super clear on the value that is returned in txbuf when calling into GetStatus. The documentation says its the pointer to the recycle buffer

Re: [PATCH 1/3] efinet: handle get_status() properly

On 08/05/2015 04:32 PM, Vladimir 'phcoder' Serbinenko wrote: This patch improperly assumes that GRUB is the only thing in EFI that transmits. Your patch surely fixed your machine but likely breaks some other machines. Could you instead make an explicit check for (void *)1 and add a comment on

Re: [PATCH 1/3] efinet: handle get_status() properly

On 08/05/2015 04:39 PM, Josef Bacik wrote: On 08/05/2015 04:32 PM, Vladimir 'phcoder' Serbinenko wrote: This patch improperly assumes that GRUB is the only thing in EFI that transmits. Your patch surely fixed your machine but likely breaks some other machines. Could you instead make an explicit