From: Jason Wang <jasowang@redhat.com>
To: Leonid Bloch <leonid.bloch@ravellosystems.com>, qemu-devel@nongnu.org
Cc: Dmitry Fleytman <dmitry@daynix.com>,
Leonid Bloch <leonid@daynix.com>,
Shmulik Ladkani <shmulik.ladkani@ravellosystems.com>
Subject: Re: [Qemu-devel] [PATCH 4/6] e1000: Fixing the received/transmitted octets' counters
Date: Tue, 20 Oct 2015 14:16:50 +0800 [thread overview]
Message-ID: <5625DC52.5010006@redhat.com> (raw)
In-Reply-To: <1445154799-31083-5-git-send-email-leonid.bloch@ravellosystems.com>
On 10/18/2015 03:53 PM, Leonid Bloch wrote:
> Previously, the lower parts of these counters (TORL, TOTL) were
> resetting after reaching their maximal values, and since the continuation
> of counting in the higher parts (TORH, TOTH) was triggered by an
> overflow event of the lower parts, the count was not correct.
>
> Additionally, TORH and TOTH were counting the corresponding frames, and
> not the octets, as they supposed to do.
>
> Additionally, these 64-bit registers did not stick at their maximal
> values when (and if) they reached them.
>
> This fix resolves all the issues mentioned above, and makes the octet
> counters behave according to Intel's specs.
>
> Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
> Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
> ---
> hw/net/e1000.c | 34 ++++++++++++++++++++++++++--------
> 1 file changed, 26 insertions(+), 8 deletions(-)
>
> diff --git a/hw/net/e1000.c b/hw/net/e1000.c
> index 5530285..7f977b6 100644
> --- a/hw/net/e1000.c
> +++ b/hw/net/e1000.c
> @@ -583,6 +583,28 @@ inc_reg_if_not_full(E1000State *s, int index)
> }
> }
>
> +static void
> +grow_8reg_if_not_full(E1000State *s, int index, int size)
> +{
> + uint32_t lo = s->mac_reg[index];
> + uint32_t hi = s->mac_reg[index+1];
> +
> + if (lo == 0xffffffff) {
> + if ((hi += size) > s->mac_reg[index+1]) {
> + s->mac_reg[index+1] = hi;
> + } else if (s->mac_reg[index+1] != 0xffffffff) {
> + s->mac_reg[index+1] = 0xffffffff;
> + }
> + } else {
> + if (((lo += size) < s->mac_reg[index])
> + && (s->mac_reg[index] = 0xffffffff)) { /* setting low to full */
> + s->mac_reg[index+1] += ++lo;
> + } else {
> + s->mac_reg[index] = lo;
> + }
> + }
> +}
How about something easier:
uint64_t sum = s->mac_reg[index] | (uint64_t)s->mac_reg[index+1] <<32;
if (sum + size < sum) {
sum = 0xffffffffffffffff;
} else {
sum += size;
}
s->max_reg[index] = sum;
s->max_reg[index+1] = sum >> 32;
> +
> static inline int
> vlan_enabled(E1000State *s)
> {
> @@ -632,7 +654,7 @@ static void
> xmit_seg(E1000State *s)
> {
> uint16_t len, *sp;
> - unsigned int frames = s->tx.tso_frames, css, sofar, n;
> + unsigned int frames = s->tx.tso_frames, css, sofar;
> struct e1000_tx *tp = &s->tx;
>
> if (tp->tse && tp->cptse) {
> @@ -678,10 +700,8 @@ xmit_seg(E1000State *s)
> } else
> e1000_send_packet(s, tp->data, tp->size);
> inc_reg_if_not_full(s, TPT);
> + grow_8reg_if_not_full(s, TOTL, s->tx.size);
> s->mac_reg[GPTC] = s->mac_reg[TPT];
> - n = s->mac_reg[TOTL];
> - if ((s->mac_reg[TOTL] += s->tx.size) < n)
> - s->mac_reg[TOTH]++;
> }
>
> static void
> @@ -1096,11 +1116,9 @@ e1000_receive_iov(NetClientState *nc, const struct iovec *iov, int iovcnt)
> /* TOR - Total Octets Received:
> * This register includes bytes received in a packet from the <Destination
> * Address> field through the <CRC> field, inclusively.
> + * Always include FCS length (4) in size.
> */
> - n = s->mac_reg[TORL] + size + /* Always include FCS length. */ 4;
> - if (n < s->mac_reg[TORL])
> - s->mac_reg[TORH]++;
> - s->mac_reg[TORL] = n;
> + grow_8reg_if_not_full(s, TORL, size+4);
>
> n = E1000_ICS_RXT0;
> if ((rdt = s->mac_reg[RDT]) < s->mac_reg[RDH])
next prev parent reply other threads:[~2015-10-20 6:17 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-18 7:53 [Qemu-devel] [PATCH 0/6] e1000: Various fixes and registers' implementation Leonid Bloch
2015-10-18 7:53 ` [Qemu-devel] [PATCH 1/6] e1000: Cosmetic and alignment fixes Leonid Bloch
2015-10-18 7:53 ` [Qemu-devel] [PATCH 2/6] e1000: Trivial implementation of various MAC registers Leonid Bloch
2015-10-20 5:40 ` Jason Wang
2015-10-21 9:13 ` Leonid Bloch
2015-10-22 7:19 ` Jason Wang
2015-10-22 14:05 ` Leonid Bloch
2015-10-23 3:10 ` Jason Wang
2015-10-25 19:39 ` Leonid Bloch
2015-10-18 7:53 ` [Qemu-devel] [PATCH 3/6] e1000: Fixing the received/transmitted packets' counters Leonid Bloch
2015-10-18 7:53 ` [Qemu-devel] [PATCH 4/6] e1000: Fixing the received/transmitted octets' counters Leonid Bloch
2015-10-20 6:16 ` Jason Wang [this message]
2015-10-21 12:20 ` Leonid Bloch
2015-10-22 7:20 ` Jason Wang
2015-10-18 7:53 ` [Qemu-devel] [PATCH 5/6] e1000: Fixing the packet address filtering procedure Leonid Bloch
2015-10-18 7:53 ` [Qemu-devel] [PATCH 6/6] e1000: Implementing various counters Leonid Bloch
2015-10-20 6:34 ` Jason Wang
2015-10-21 9:20 ` Leonid Bloch
2015-10-20 6:37 ` [Qemu-devel] [PATCH 0/6] e1000: Various fixes and registers' implementation Jason Wang
2015-10-21 13:32 ` Leonid Bloch
2015-10-22 7:22 ` Jason Wang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5625DC52.5010006@redhat.com \
--to=jasowang@redhat.com \
--cc=dmitry@daynix.com \
--cc=leonid.bloch@ravellosystems.com \
--cc=leonid@daynix.com \
--cc=qemu-devel@nongnu.org \
--cc=shmulik.ladkani@ravellosystems.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).