Page MenuHomeFreeBSD
Differential D4940

mlx5en: Add clarifying comment about CQE zipping
ClosedPublic
Actions

Authored by markb_mellanox.com on Jan 14 2016, 1:52 PM.
Tags
None
Referenced Files
Unknown Object (File)
Mon, Apr 14, 7:24 AM
Unknown Object (File)
Mon, Apr 14, 6:42 AM
Unknown Object (File)
Mon, Apr 14, 2:44 AM
Unknown Object (File)
Mon, Apr 14, 2:21 AM
Unknown Object (File)
Feb 19 2025, 1:07 AM
Unknown Object (File)
Feb 19 2025, 1:07 AM
Unknown Object (File)
Feb 11 2025, 5:39 PM
Unknown Object (File)
Jan 17 2025, 7:28 AM
View All 97 Files
Subscribers
imp

Details

Reviewers
• hselasky
gnn
Commits
rS294919: MFC r294318:
rS294318: Add clarifying comment about CQE zipping.
Summary

Add a comment to clarify how CQE zipping works.

Submitted by: Hans Petter Selasky <hps@selasky.org>
Sponsored by: Mellanox Technologies

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

markb_mellanox.com updated this revision to Diff 12313.Jan 14 2016, 1:52 PM
markb_mellanox.com retitled this revision from to mlx5en: Add clarifying comment about CQE zipping.
markb_mellanox.com updated this object.
markb_mellanox.com edited the test plan for this revision. (Show Details)
markb_mellanox.com added reviewers: gnn, • hselasky.
markb_mellanox.com set the repository for this revision to rS FreeBSD src repository - subversion.
Herald added a subscriber: imp. · View Herald TranscriptJan 14 2016, 1:52 PM
markb_mellanox.com added a parent revision: D4939: mlx5en: Fix declaration of variables.Jan 14 2016, 1:52 PM
markb_mellanox.com added a child revision: D4941: mlx5en: Rework code which checks the HW LRO sysctl.Jan 14 2016, 1:55 PM
gnn accepted this revision.Jan 14 2016, 8:16 PM
gnn edited edge metadata.
This revision is now accepted and ready to land.Jan 14 2016, 8:16 PM
• hselasky accepted this revision.Jan 19 2016, 10:18 AM
• hselasky edited edge metadata.
Closed by commit rS294318: Add clarifying comment about CQE zipping. (authored by • hselasky). · Explain WhyJan 19 2016, 10:20 AM
This revision was automatically updated to reflect the committed changes.

Revision Contents
Changeset List

PathSize
head/
sys/
dev/
mlx5/
mlx5_en/
5 lines

Diff 12460

View Options

head/sys/dev/mlx5/mlx5_en/mlx5_en_rx.c

Show First 20 Lines • Show All 211 Lines • ▼ Show 20 Lines if ((frame->tf_scause == EXCP_FAULT_STORE) ||
ftype = VM_PROT_EXECUTE; ftype = VM_PROT_EXECUTE;
} else { } else {
ftype = VM_PROT_READ; ftype = VM_PROT_READ;
} }
if (pmap_fault_fixup(map->pmap, va, ftype)) if (pmap_fault_fixup(map->pmap, va, ftype))
goto done; goto done;
if (map != kernel_map) { error = vm_fault_trap(map, va, ftype, VM_FAULT_NORMAL, &sig, &ucode);
/*
* Keep swapout from messing with us during this
* critical time.
*/
PROC_LOCK(p);
++p->p_lock;
PROC_UNLOCK(p);
/* Fault in the user page: */
error = vm_fault(map, va, ftype, VM_FAULT_NORMAL);
PROC_LOCK(p);
--p->p_lock;
PROC_UNLOCK(p);
} else {
/*
* Don't have to worry about process locking or stacks in the
* kernel.
*/
error = vm_fault(map, va, ftype, VM_FAULT_NORMAL);
}
alcUnsubmitted
Done
Inline Actions

I'm surprised to see this snippet brought back to life in the riscv code. We removed it from elsewhere in the source tree almost exactly 4 years ago (r287625).

I would suggest committing the removal of this snippet now, and thus separately from the rest.

alc: I'm surprised to see this snippet brought back to life in the riscv code. We removed it from…
if (error != KERN_SUCCESS) { if (error != KERN_SUCCESS) {
if (usermode) { if (usermode) {
sig = SIGSEGV;
if (error == KERN_PROTECTION_FAILURE)
ucode = SEGV_ACCERR;
else
ucode = SEGV_MAPERR;
call_trapsignal(td, sig, ucode, (void *)stval); call_trapsignal(td, sig, ucode, (void *)stval);
} else { } else {
if (pcb->pcb_onfault != 0) { if (pcb->pcb_onfault != 0) {
frame->tf_a[0] = error; frame->tf_a[0] = error;
frame->tf_sepc = pcb->pcb_onfault; frame->tf_sepc = pcb->pcb_onfault;
return; return;
} }
goto fatal; goto fatal;
▲ Show 20 Lines • Show All 139 Lines • Show Last 20 Lines