On 15 May 2021, at 21:56, Mateusz Guzik <mjgu...@gmail.com> wrote:
>
> One of these commits breaks tinderbox:
> i386 MINIMAL kernel failed, check _.i386.MINIMAL for details
> i386 GENERIC kernel failed, check _.i386.GENERIC for details
> i386 GENERIC-NODEBUG kernel failed, check _.i386.GENERIC-NODEBUG for details
> i386 PAE kernel failed, check _.i386.PAE for details
> i386 LINT kernel failed, check _.i386.LINT for details
> i386 LINT-NOINET kernel failed, check _.i386.LINT-NOINET for details
> i386 LINT-NOINET6 kernel failed, check _.i386.LINT-NOINET6 for details
> i386 LINT-NOIP kernel failed, check _.i386.LINT-NOIP for details
>
> /usr/src/sys/dev/cxgbe/tom/t4_ddp.c:1045:15: error: invalid operands
> to binary expression ('void *' and 'int')
> start_pva = trunc_page(sgl->addr);
> ^~~~~~~~~~~~~~~~~~~~~
> ./machine/param.h:150:29: note: expanded from macro 'trunc_page'
> #define trunc_page(x) ((x) & ~PAGE_MASK)
> ~~~ ^ ~~~~~~~~~~
> /usr/src/sys/dev/cxgbe/tom/t4_ddp.c:1300:8: error: invalid operands to
> binary expression ('void *' and 'int')
> pva = trunc_page(sgl->addr);
> ^~~~~~~~~~~~~~~~~~~~~
> ./machine/param.h:150:29: note: expanded from macro 'trunc_page'
> #define trunc_page(x) ((x) & ~PAGE_MASK)
> ~~~ ^ ~~~~~~~~~~
It seems amd64, arm64 and riscv cast to unsigned long first, but arm, i386,
mips and powerpc do not. These macros should probably just become MI in
sys/sys/param.h rather than fixing the MD copies to include casts.
Jess
> On 5/14/21, John Baldwin <j...@freebsd.org> wrote:
>> The branch main has been updated by jhb:
>>
>> URL:
>> https://cgit.FreeBSD.org/src/commit/?id=e73e2ee0acf5a0e0f47b9c2bcd73c835c4922fab
>>
>> commit e73e2ee0acf5a0e0f47b9c2bcd73c835c4922fab
>> Author: John Baldwin <j...@freebsd.org>
>> AuthorDate: 2021-05-14 19:20:57 +0000
>> Commit: John Baldwin <j...@freebsd.org>
>> CommitDate: 2021-05-14 19:21:34 +0000
>>
>> cxgbei: Handle target transfers with excess unsolicited data.
>>
>> The CTL frontend might have provided a buffer that is smaller than the
>> FirstBurstLength and thus smaller than the amount of unsolicited data
>> included in the request PDU. Treat these transfers as an empty
>> transfer.
>>
>> Reported by: Jithesh Arakkan @ Chelsio
>> Sponsored by: Chelsio Communications
>>
>> Differential Revision: https://reviews.freebsd.org/D29940
>> ---
>> sys/dev/cxgbe/cxgbei/icl_cxgbei.c | 9 +++++++--
>> 1 file changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/sys/dev/cxgbe/cxgbei/icl_cxgbei.c
>> b/sys/dev/cxgbe/cxgbei/icl_cxgbei.c
>> index 7f638c96483a..c4eb4a35ad31 100644
>> --- a/sys/dev/cxgbe/cxgbei/icl_cxgbei.c
>> +++ b/sys/dev/cxgbe/cxgbei/icl_cxgbei.c
>> @@ -1064,10 +1064,15 @@ icl_cxgbei_conn_transfer_setup(struct icl_conn *ic,
>> union ctl_io *io,
>> /*
>> * Note that ICL calls conn_transfer_setup even if the first
>> * burst had everything and there's nothing left to transfer.
>> + *
>> + * NB: The CTL frontend might have provided a buffer
>> + * whose length (kern_data_len) is smaller than the
>> + * FirstBurstLength of unsolicited data. Treat those
>> + * as an empty transfer.
>> */
>> - MPASS(ctsio->kern_data_len >= first_burst);
>> xferlen = ctsio->kern_data_len;
>> - if (xferlen - first_burst < ci->ddp_threshold) {
>> + if (xferlen < first_burst ||
>> + xferlen - first_burst < ci->ddp_threshold) {
>> no_ddp:
>> /*
>> * No DDP for this transfer. Allocate a TTT (based on
>>
>
>
> --
> Mateusz Guzik <mjguzik gmail.com>
_______________________________________________
dev-commits-src-main@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/dev-commits-src-main
To unsubscribe, send any mail to "dev-commits-src-main-unsubscr...@freebsd.org"
