Skip to content

cloud-kernel
cloud-kernel

openanolis / cloud-kernel
大约 1 年 前同步成功

通知 155
Star 36
Fork 7
cloud-kernel
cloud-kernel

体验新版 GitCode，发现更多精彩内容 >>

切换分支/标签
  1. 30 5月, 2005 7 次提交
  3. 24 5月, 2005 4 次提交
  5. 19 5月, 2005 4 次提交
    • D
      [TG3]: In tg3_poll(), resample status_tag after doing work. · f7383c22
      由 David S. Miller 提交于 
      Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      f7383c22
    • D
      [TG3]: Refine DMA boundary setting. · 59e6b434
      由 David S. Miller 提交于 
      Extract DMA boundary bit selection into a seperate
function, tg3_calc_dma_bndry().  Call this from
tg3_test_dma().

Make DMA test more reliable by using no DMA boundry
setting during the test.  If the test passes, then
use the setting we selected before the test.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NMichael Chan <mchan@broadcom.com>
      59e6b434
    • D
      [TG3]: Set minimal hw interrupt mitigation. · 15f9850d
      由 David S. Miller 提交于 
      Even though we do software interrupt mitigation
via NAPI, it still helps to have some minimal
hw assisted mitigation.

This helps, particularly, on systems where register
I/O overhead is much greater than the CPU horsepower.

For example, it helps on NUMA systems.  In such cases
the PIO overhead to disable interrupts for NAPI accounts
for the majority of the packet processing cost.  The
CPU is fast enough such that only a single packet is
processed by each NAPI poll call.

Thanks to Michael Chan for reviewing this patch.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      15f9850d
    • D
      [TG3]: Add tagged status support. · fac9b83e
      由 David S. Miller 提交于 
      When supported, use the TAGGED interrupt processing support
the chip provides.  In this mode, instead of a "on/off" binary
semaphore, an incrementing tag scheme is used to ACK interrupts.

All MSI supporting chips support TAGGED mode, so the tg3_msi()
interrupt handler uses it unconditionally.  This invariant is
verified when MSI support is tested.

Since we can invoke tg3_poll() multiple times per interrupt under
high packet load, we fetch a new copy of the tag value in the
status block right before we actually do the work.

Also, because the tagged status tells the chip exactly which
work we have processed, we can make two optimizations:

1) tg3_restart_ints() need not check tg3_has_work()
2) the tg3_timer() need not poke the chip 10 times per
   second to keep from losing interrupt events

Based upon valuable feedback from Michael Chan <mchan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      fac9b83e
  7. 06 5月, 2005 3 次提交
  9. 29 4月, 2005 1 次提交
  11. 26 4月, 2005 3 次提交
    • M
      [TG3]: Fix tg3_restart_ints() · 04237ddd
      由 Michael Chan 提交于 
      tg3_restart_ints() is called to re-enable interrupts after tg3_poll()
has finished all the work. It calls tg3_cond_int() to force an interrupt
if the status block updated bit is set. The updated bit will be set if
there is a new status block update sometime during tg3_poll() and it can
be very often. The worst part is that even if all the work has been
processed, the updated bit remains set and an interrupt will be forced
unnecessarily.

The fix is to call tg3_has_work() instead to determine if new work is
posted before forcing an interrupt. The way to force an interrupt is
also changed to use "coalesce_now" instead of "SETINT". The former is
generally a safer way to force the interrupt.

Also deleted the first parameter to tg3_has_work() which is unused.
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      04237ddd
    • M
      [TG3]: Refresh hw index in tg3_rx() · 52f6d697
      由 Michael Chan 提交于 
      This patch refreshes the hw rx producer in tg3_rx() so that additional
work posted by the hardware can be processed.
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      52f6d697
    • M
      [TG3]: Fix bug in tg3_rx() · 483ba50b
      由 Michael Chan 提交于 
      This patch fixes a bug that causes tg3_has_work() to always return 1.

rx work is determined by comparing tp->rx_rcb_ptr with the current hw
producer index. The hw producer index is modulo the ring size, but tp-
>rx_rcb_ptr is a free running counter that goes up beyond the ring size.
After the ring wraps around once, tg3_has_work() will always return 1.

The fix is to always do modulo arithmetic on tp->rx_rcb_ptr.
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      483ba50b
  13. 25 4月, 2005 1 次提交
  15. 22 4月, 2005 17 次提交